THE ROLE OF THE MATRIX IN SHAPING EXTINCTION RISK AND
CONSERVATION OPPORTUNITIES FOR TERRESTRIAL MAMMALS
by
Juan Pablo Ramírez-Delgado
M.Sc., El Colegio de la Frontera Sur (ECOSUR), 2012
B.Sc., New Granada Military University, 2008

DISSERTATION SUBMITTED IN PARTIAL FULFILLMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
IN
NATURAL RESOURCES AND ENVIRONMENTAL STUDIES

UNIVERSITY OF NORTHERN BRITISH COLUMBIA
August 2025
© Juan Ramírez-Delgado, 2025

Abstract
The global biodiversity crisis, driven primarily by habitat loss and fragmentation, has
traditionally led conservation efforts to focus almost exclusively on primary habitats. However,
the ecological importance of the matrix, defined as areas surrounding primary habitat, has
remained largely overlooked.
In this dissertation, I used global-scale spatial analyses coupled with statistical modeling
to (i) quantify how matrix condition influences the effects of habitat fragmentation on extinction
risk for terrestrial mammals, (ii) compare the predictive performance of alternative habitat
intactness models (patch-matrix, continuum, and hybrid models) for assessing extinction risk in
terrestrial mammals, and (iii) map global patterns of terrestrial mammal species richness within
the matrix to identify conservation opportunities beyond primary habitats.
My findings demonstrate that matrix condition plays a key mediating role in the
relationship between habitat fragmentation and extinction risk, with greater predictive power
than habitat loss or habitat amount alone. Moreover, I found that the predictive importance of
fragmentation increases as matrix condition deteriorates, suggesting that managing or restoring
the matrix represents a strategic conservation action to mitigate the negative effects of
fragmentation on biodiversity. Additionally, the hybrid habitat intactness model—which
integrates discrete habitat patches with continuous gradients of habitat quality—consistently
outperforms traditional patch-matrix and continuum models, regardless of species’ habitat
specialization. Notably, the magnitude of the relationship between habitat intactness and
extinction risk was greater when using the hybrid model, highlighting that integrating discrete
and continuous habitat representations can improve extinction risk analyses and provide valuable
insights for conservation. My results further reveal that hotspots of species richness within the
ii

matrix occupy only about 1% of Earth's terrestrial surface, yet could support more than half of all
terrestrial mammal species. Matrix areas identified as having high conservation potential—based
on overlapping richness hotspots—are primarily concentrated in tropical strongholds such as the
Amazon Basin, Colombian Tropical Andes, Brazilian Atlantic Forest, and Albertine Rift.
Importantly, many of these matrix areas face intense human pressures and remain inadequately
represented within existing protected areas and other area-based conservation measures,
underscoring their value as strategic opportunities for biodiversity conservation.
Collectively, my results highlight an urgent need for a paradigm shift in conservation
strategies that explicitly recognize, manage, and restore matrix areas as integral components of
global biodiversity conservation. Integrating the matrix into conservation planning closely aligns
with international biodiversity frameworks, particularly Target 2 of the Kunming-Montreal
Global Biodiversity Framework, which calls for restoring at least 30% of degraded terrestrial
ecosystems to enhance ecological integrity and connectivity. Such integration could substantially
improve biodiversity outcomes, ecosystem resilience, and landscape connectivity, ultimately
making critical contributions toward reversing global biodiversity declines.

iii

Table of contents
Abstract .......................................................................................................................................... ii
List of tables.................................................................................................................................. vi
List of figures ............................................................................................................................... vii
Acknowledgments ....................................................................................................................... xii
Chapter 1: Introduction ............................................................................................................. 13
Terrestrial mammals ................................................................................................................... 13
Habitat loss and fragmentation ................................................................................................... 15
The matrix and landscape complexity ........................................................................................ 17
Research objectives .................................................................................................................... 20
Why focus on terrestrial mammals? ........................................................................................... 21
Conceptual definition of habitat ................................................................................................. 22
Conceptual definition of the matrix ............................................................................................ 23
Dissertation structure .................................................................................................................. 23
Chapter 2: Matrix condition mediates the effects of habitat fragmentation on species
extinction risk .............................................................................................................................. 25
Abstract....................................................................................................................................... 25
Introduction ................................................................................................................................ 25
Methods ...................................................................................................................................... 29
Habitat suitability models ..................................................................................................... 29
The degree of habitat fragmentation and the degree of patch isolation as predictors of
extinction risk transitions ...................................................................................................... 31
The matrix condition as a predictor of extinction risk transitions ........................................ 31
Changes in species extinction risk ........................................................................................ 33
Predicting extinction risk transitions .................................................................................... 34
Assessing the influence of the matrix on the importance of fragmentation for predicting
extinction risk transitions ...................................................................................................... 38
Sensitivity analysis................................................................................................................ 39
Results ........................................................................................................................................ 40
Changes in species extinction risk ........................................................................................ 40
Predicting transitions in species extinction risk .................................................................... 41
The influence of the matrix condition on the importance of habitat fragmentation for
predicting extinction risk transitions ..................................................................................... 43
Discussion................................................................................................................................... 47
Chapter 3: Global comparison of habitat intactness models for predicting extinction risk in
terrestrial mammals.................................................................................................................... 55
Abstract....................................................................................................................................... 55
Introduction ................................................................................................................................ 55
Methods ...................................................................................................................................... 57
Species-level data.................................................................................................................. 57
Human footprint .................................................................................................................... 58
Patch-matrix model as a predictor of extinction risk ............................................................ 60
Continuum model as a predictor of extinction risk ............................................................... 60
Hybrid model as a predictor of extinction risk ..................................................................... 62
iv

Statistical analysis ................................................................................................................. 62
Assessing model performance .............................................................................................. 65
Sensitivity analysis................................................................................................................ 65
Results ........................................................................................................................................ 66
Discussion................................................................................................................................... 70
Chapter 4: Mapping the conservation potential of the matrix for terrestrial mammals ..... 74
Abstract....................................................................................................................................... 74
Introduction ................................................................................................................................ 75
Methods ...................................................................................................................................... 77
Species area of habitat .......................................................................................................... 77
Species richness maps ........................................................................................................... 79
Species-rich matrix areas ...................................................................................................... 79
Hotspots of species richness ................................................................................................. 80
Levels of conservation potential within matrix areas ........................................................... 80
Species-level hotspot representation ..................................................................................... 81
Human pressure in matrix hotspots....................................................................................... 81
Conservation coverage of matrix hotspots ............................................................................ 82
Results ........................................................................................................................................ 83
Global patterns of terrestrial mammal concentration in the matrix ...................................... 83
Matrix areas with conservation potential .............................................................................. 88
Human footprint and conservation coverage in matrix areas with conservation potential ... 91
Discussion................................................................................................................................... 94
Chapter 5: Conclusions ............................................................................................................ 103
Conservation implications ........................................................................................................ 104
Limitations and future directions.............................................................................................. 105
References .................................................................................................................................. 110
Appendix A: Supplementary material for “Chapter 2: Matrix condition mediates the
effects of habitat fragmentation on species extinction risk” ................................................. 135
Appendix B: Supplementary material for “Chapter 3: Global comparison of habitat
intactness models for predicting extinction risk for terrestrial mammals” ........................ 142
Appendix C: Supplementary material for “Chapter 4: Mapping the conservation potential
of matrix habitats for terrestrial mammals” .......................................................................... 150

v

List of tables
Table 2.1: Description of the selected variables to predict extinction risk transitions in
terrestrial mammals.................................................................................................................... 35
Table 3.1: Description of variables used to predict species extinction risk. We selected the
most explanatory percentile from our continuum and hybrid models of habitat intactness to
compare how patch-matrix, continuum, and hybrid models predict species extinction risk. ....... 63
Table 4.1: Geographic coverage and species representation of hotspot types. The table
shows the total area of each hotspot type and the number of species whose matrix area
overlapped with each hotspot type. Numbers in parentheses indicate the percentage of each
hotspot type relative to the global extent of species-rich matrix areas (15,922,965 km²), or the
percentage of overlapping species relative to the total number of extant terrestrial mammals
(5,808 species worldwide). Slight differences in area across hotspot types reflect clustering
requirements and the threshold optimization process (see Methods). .......................................... 87
Table A.1: Cross-validation results of the Random Forest models for the prediction of
extinction risk transitions in terrestrial mammals. *Cross-validation results when the extent
of suitable habitat is represented by high and medium habitat suitability combined, and the extent
of the matrix by unsuitable habitat alone. †Cross-validation results when the extent of suitable
habitat is represented by high habitat suitability, and the extent of the matrix by medium habitat
suitability and unsuitable habitat combined. The cross-validation was performed in terms of
proportion of correctly classified species (accuracy), proportion of correctly classified high-risk
species (sensitivity), proportion of correctly classified low-risk species (specificity), and the true
skill statistic (TSS = sensitivity + specificity – 1). ..................................................................... 137
Table C.1: Geographic coverage and species representation of hotspot types under
different hotspot definitions. The table includes the total area and number of species whose
matrix area overlap with each hotspot type, based on two different thresholds: the top 2.5% and
the top 10% of pixels with the highest richness values. Numbers in parentheses indicate the
percentage of species-rich matrix areas (out of 15,922,965 km² globally) or the percentage of
extant terrestrial mammal species (out of 5,808 species worldwide). ........................................ 151

vi

List of figures
Figure 2.1: Classification of species extinction risk transitions based on past and present
IUCN Red List categories. Low-risk transitions included species that retained a category of
least concern, together with those species that moved from any higher category of threat to a
lower category between 1996 and 2020 (a). High-risk transitions included all species that
retained a category of threatened or near threatened, together with those species that moved from
any lower category of threat to a higher category between 1996 and 2020 (b). Acronyms refer to
the IUCN Red List categories, including Least Concern (LC), Near Threatened (NT), Vulnerable
(VU), Endangered (EN), and Critically Endangered (CR). .......................................................... 40
Figure 2.2: Relative importance of selected variables for the prediction of extinction risk
transitions in terrestrial mammals. Variables are colour-coded according to their broad class
(human pressure, environment, and life-history). The description of each variable can be found in
Table 2.1. High levels of the human footprint (HFP) included values of 3 or above. .................. 41
Figure 2.3: Partial dependence plots to show the effect of the degree of habitat
fragmentation, the degree of patch isolation, and the matrix condition on extinction risk
transitions in terrestrial mammals. The plots show the probability of high-risk transitions as a
function of the degree of fragmentation (a), the degree of patch isolation (b), and the extent of
high human footprint values within the matrix (c). Solid red lines and shading represent fitted
LOESS curves and 95% credible intervals for the relationships between the probability of highrisk transitions and each explanatory variable. As partial dependence plots for Boolean response
variables mirror each other, the probability of low-risk transitions as a function of these variables
are not depicted in the figure. Values of the degree of fragmentation and the degree of patch
isolation were ln-transformed for visual purposes. The degree of fragmentation was inversecoded so high values represent high degrees of fragmentation. High values of the degree of patch
isolation represent high degrees of isolation between patches of suitable habitat. The description
of each variable is given in Table 2.1. High levels of the human footprint (HFP) included values
of 3 or above. ................................................................................................................................ 42
Figure 2.4: Influence of the matrix condition on the relative importance of selected
variables for the prediction of extinction risk transitions in terrestrial mammals. Relative
importance of each predictor for species with a low-quality matrix (a), which included
proportions > 84.2% of the extent of their matrix overlapping with high human footprint values
(n = 1,815 low-risk species and 1,027 high-risk species). Relative importance of each predictor
for species with a high-quality matrix (b), which included proportions < 15.8% of the extent of
their matrix overlapping with high human footprint values (n = 60 low-risk species and 29 highrisk species). Variables are colour-coded according to their broad class (human pressure,
environment, and life-history). The description of each variable is given in Table 2.1. High levels
of the human footprint (HFP) included values of 3 or above. ...................................................... 45
Figure 3.1: An example of habitat intactness models used to predict species extinction risk.
(a) This example is based on the geographic range size (polygon delimited in black) of the
western gorilla (Gorilla gorilla). (b) The patch-matrix model illustrates discrete patches of intact
habitat surrounded by an ‘inhospitable’ matrix. (c) The continuum model depicts a gradient of
habitat intactness, ranging from intact habitats (in blue) to degraded habitats (in red). (d) The
hybrid model combines characteristics of patch-matrix and continuum models, representing
vii

discrete habitat patches and a gradient of habitat intactness within these patches, which is
influenced by the matrix condition. .............................................................................................. 59
Figure 3.2: Effects of three habitat intactness models (a, b) and their temporal reduction (c,
d) on extinction risk for terrestrial mammals classified as habitat generalists and habitat
specialists. Each variable is colour-coded according to the habitat model (patch-matrix,
continuum, or hybrid model) it represents. Effect sizes were estimated while holding other
variables known to influence extinction risk (see Table 3.1) at their mean. All variables were
standardized (z-transformed) before analysis to ensure comparable results. Error bars represent
95% confidence intervals. The statistical significance of the coefficients is indicated by the
absence of overlap between their confidence intervals and zero (dashed green line). Acronyms:
patch-matrix model (PMM), continuum model (CM), hybrid model (HM), reduction in the
patch-matrix model (RPMM), reduction in the continuum model (RCM), and reduction in the
hybrid model (RHM). ................................................................................................................... 68
Figure 3.3: Predicted probabilities of a species being threatened as a function of three
habitat intactness models for both habitat generalists (a) and habitat specialists (b).
Predicted probabilities were estimated while holding other variables known to influence
extinction risk (see Table 3.1) at their mean. Variables representing patch-matrix, continuum,
and hybrid models were back-scaled to improve results interpretability. Shaded portions
represent 95% confidence intervals. ............................................................................................. 70
Figure 4.1: Global distribution of species richness within species-rich matrix areas for
terrestrial mammals. Species richness is shown for all species (a), species with declining
population trends (b), and species with threatened status (c). ...................................................... 84
Figure 4.2: Hotspots of species richness within species-rich matrix areas for terrestrial
mammals. Hotspots are shown for all mammals (a), mammals with declining population trends
(b), and mammals with threatened status (c). Dark red represents the top 2.5% of grid cells with
the highest richness values for each species group, red (including dark red) represents the top
5%, and pink (including both red and dark red) represents the top 10%. All hotspots were
aggregated into contiguous clusters with a minimum area of 10 km². ......................................... 86
Figure 4.3: Overlap of hotspot types and conservation potential in matrix areas. The map
shows the global distribution of species-rich matrix areas, categorized into levels of conservation
potential based on spatial overlap among three hotspot types: all mammals, mammals with
declining population trends, and mammals with threatened status (a). Venn diagrams show the
spatial overlap among hotspot types (b) and the number of species whose matrix habitat
overlapped one, two, or all three types (c), respectively. Each species was counted only once,
based on the unique combination of hotspot types overlapping its matrix habitat (e.g., one, two,
or all three types), and was assigned to the corresponding category. Areas of high conservation
potential corresponded to overlaps among all three hotspot types, moderate conservation
potential to overlaps between any two types, and low conservation potential to areas identified
by a single hotspot type................................................................................................................. 89
Figure 4.4: Human footprint within matrix areas with conservation potential. The map
shows the global distribution of matrix areas with high, moderate, and low conservation
potential, categorized by human pressure levels: intact, low-disturbance, and highly modified.
Insets show examples from Brazil’s Amazon (a) and Atlantic Forest (b), the Western Congolian
forest-savanna mosaic (c), and Sumatra’s forests (d). Human pressure levels were derived from
viii

the 2015 global HFP map developed by Gassert et al. (2023) at a 100-m spatial resolution. For
details on thresholds used to define each human pressure level, see Methods. ............................ 92
Figure A.1: Transition matrix of extinction risk categories for terrestrial mammals
between 1996 and 2020. The plot shows the transition matrix of the first and last Red List
category reported between 1996 and 2020. The colour scheme represents the relative frequency
of individual species in each transition. Acronyms refer to the IUCN Red List categories,
including Least Concern (LC), Near Threatened (NT), Vulnerable (VU), Endangered (EN), and
Critically Endangered (CR). The relative frequency was arcsine square-root-transformed for
visual purposes and to avoid variance instability when handling proportions close to zero. ..... 135
Figure A.2: Sensitivity analysis of selected variables for the prediction of extinction risk
transitions in terrestrial mammals. Relative importance of each predictor when the extent of
suitable habitat is represented by high and medium habitat suitability combined, and the extent of
the matrix by unsuitable habitat alone (a). Relative importance of each predictor when the extent
of suitable habitat is represented by high habitat suitability, and the extent of the matrix by
medium habitat suitability and unsuitable habitat combined (b). Variables are colour-coded
according to their broad class (human pressure, environment, and life-history). The description
of each variable is given in Table 2.1 of the main manuscript. High levels of the human footprint
(HFP) included values of 3 or above. ......................................................................................... 136
Figure A.3: Distribution of the matrix condition in low-risk and high-risk species at a
global scale. The plot represents the frequency distribution of the extent of high human footprint
values within the matrix of 4,329 species (68% low-risk species and 32% high-risk species).
Species with ranges that did not overlap with the human footprint (2% of 4,426 terrestrial
mammals in our sample) were excluded from the calculations. Blue and red lines refer to lowrisk and high-risk species, as reported in legend. Dashed lines denote mean of distributions. High
levels of the human footprint (HFP) included values of 3 or above. .......................................... 138
Figure A.4: Distribution of the matrix condition in low-risk and high-risk species at the
scale of individual biogeographic realms. The plots represent the frequency distribution of the
extent of high human footprint values within the matrix of 1.109 species in the Neotropical realm
(68% low-risk species and 32% high-risk species), 990 species in the Afrotropical realm (65%
low-risk species and 35% high-risk species), 596 species in the Indo-Malay realm (52% low-risk
species and 48% high-risk species), 552 species in the Australasia realm (58% low-risk species
and 42% high-risk species), 498 species in the Palearctic realm (79% low-risk species and 21%
high-risk species), and 286 species in the Nearctic realm (83% low-risk species and 17% highrisk species). The Oceania realm included a total of 4 species only (17% low-risk species and
83% high-risk species), which are not depicted on the map. Species with ranges that overlapped
with two or more realms (9% of 4,426 terrestrial mammals in our sample) were excluded from
the calculations. Blue and red lines refer to low-risk and high-risk species, as reported in legend.
Dashed lines denote mean of distributions. High levels of the human footprint (HFP) included
values of 3 or above. ................................................................................................................... 139
Figure A.5: Effect size (Cohen’s d statistic) of the degree of habitat fragmentation and the
degree of patch isolation between low-risk and high-risk species with a low-quality matrix
and a high-quality matrix. Effect size of the degree of fragmentation and the degree of patch
isolation between low-risk and high-risk species with a matrix of low-quality habitat (a). Effect
size of the degree of fragmentation and the degree of patch isolation between low-risk and highix

risk species with a matrix of high-quality habitat (b). Green and yellow points represent the
effect sizes or the standardized difference of mean values (Cohen’s d statistic) of the degree of
fragmentation and the degree of patch isolation between low-risk and high-risk species,
respectively. Error bars represent 95% confidence intervals for each effect size calculated. Lowquality matrices included species with proportions > 84.2% of the extent of their matrix
overlapping with high human footprint values (n = 1,815 low-risk species and 1,027 high-risk
species). High-quality matrices included species with proportions < 15.8% of the extent of their
matrix overlapping with high human footprint values (n = 60 low-risk species and 29 high-risk
species). High levels of the human footprint included values of 3 or above. ............................. 140
Figure B.1: Effects of the 5th, 10th, 50th, 90th, and 95th percentiles of the continuum and
hybrid habitat intactness models on extinction risk for terrestrial mammals classified as
habitat generalists and specialists. Each variable is colour-coded according to the habitat
model (continuum or hybrid model) it represents. Effect sizes were estimated while holding other
variables known to influence extinction risk (see Table 3.1) at their mean. All variables were
standardized (z-transformed) before analysis to ensure comparable results. Error bars represent
95% confidence intervals. Confidence intervals not containing zero (dashed green line)
indicate statistical significance. .................................................................................................. 143
Figure B.2: Correlation between variables used to predict extinction risk for terrestrial
mammals classified as habitat generalists and habitat specialists. Pearson’s correlation
coefficient (r) was used to quantify relationships between variables. Acronyms: patch-matrix
model (PMM), reduction in the patch-matrix model (RPMM), continuum model (CM), reduction
in the continuum model (RCM), hybrid model (HM), reduction in the hybrid model (RHM),
range size (RS), gestation length (GL), and weaning age (WA). Detailed descriptions of these
variables are provided in Table 3.1. ............................................................................................ 144
Figure B.3: Predictive performance of extinction risk models for both habitat generalists
and habitat specialists as a function of three habitat intactness models and how they
decrease over time. AUC values were estimated from mixed-effects logistic regression models
using a spatially blocked cross-validation and a ten-fold cross-validation scheme (see Methods).
AUCs above 0.5 (dashed green line) indicate that model performance is better than random
chance. All extinction risk models accounted for the effects of other variables known to
influence extinction risk (see Table 1). Error bars represent 95% confidence intervals. Acronyms:
extinction risk model (ER), patch-matrix model (PMM), continuum model (CM), hybrid model
(HM), reduction in the patch-matrix model (RPMM), reduction in the continuum model (RCM),
and reduction in the hybrid model (RHM). ................................................................................ 145
Figure B.4: Predictive performance of extinction risk models for both habitat generalists
and habitat specialists as a function of three habitat intactness models and their temporal
decline, before and after excluding threatened species assessed solely under IUCN Red List
Criterion B. AUC values were estimated from mixed-effects logistic regression models using a
spatially blocked cross-validation scheme (see Methods). AUCs above 0.5 (dashed green line)
indicate that model performance is better than random chance. All extinction risk models
accounted for the effects of other variables known to influence extinction risk (see Table 1).
Error bars represent 95% confidence intervals. Acronyms: extinction risk model (ER), patchmatrix model (PMM), continuum model (CM), hybrid model (HM), reduction in the patchmatrix model (RPMM), reduction in the continuum model (RCM), and reduction in the hybrid
model (RHM). ............................................................................................................................. 146
x

Figure B.5: Predictive performance of extinction risk models for both habitat generalists
and habitat specialists as a function of three habitat intactness models and their decline
over time, before and after excluding threatened species assessed solely under IUCN Red
List Criterion B. AUC values were estimated from mixed-effects logistic regression models
using a ten-fold cross-validation scheme (see Methods). AUCs above 0.5 (dashed green line)
indicate that model performance is better than random chance. All extinction risk models
accounted for the effects of other variables known to influence extinction risk (see Table 1).
Error bars represent 95% confidence intervals. Acronyms: extinction risk model (ER), patchmatrix model (PMM), continuum model (CM), hybrid model (HM), reduction in the patchmatrix model (RPMM), reduction in the continuum model (RCM), and reduction in the hybrid
model (RHM). ............................................................................................................................. 147
Figure B.6: Effects of three habitat intactness models and their decline over time on
extinction risk for terrestrial mammals classified as habitat generalists and habitat
specialists, before and after excluding threatened species assessed solely under IUCN Red
List Criterion B. Effect sizes were estimated while holding other variables known to influence
extinction risk (see Table 3.1) at their mean. All variables were standardized (z-transformed)
before analysis to ensure comparable results. Error bars represent 95% confidence intervals. The
statistical significance of the coefficients is indicated by the absence of overlap between their
confidence intervals and zero (dashed green line). Acronyms: patch-matrix model (PMM),
continuum model (CM), hybrid model (HM), reduction in the patch-matrix model (RPMM),
reduction in the continuum model (RCM), and reduction in the hybrid model (RHM). ............ 148
Figure C.1: Species richness in matrix areas for terrestrial mammals. Maps show species
richness within matrix areas for all species (a), species with declining population trends (b), and
species with threatened status (c). The dataset includes 5,439 species in the all-mammal group,
1,765 in the declining population trend group, and 1,212 in the threatened status group. ......... 150
Figure C.2: Species richness in suitable habitats for terrestrial mammals. The map shows
species richness within suitable habitats for the all-mammal group........................................... 151
Figure C.3: Overlap of hotspot types and conservation potential in matrix areas across
different hotspot definitions. Venn diagrams illustrate the spatial overlap of hotspot types (a, c)
and species richness overlap across these hotspot types (b, d) under two different thresholds: the
top 2.5% and the top 10% of pixels with the highest richness values for each species group (all
mammals, mammals with declining population trends, and mammals with threatened status). 152
Figure C.4: Conservation status and human footprint within matrix areas with
conservation potential. The figure shows the proportion of matrix hotspots classified as high-,
moderate-, and low-priority that fall within protected areas (PAs) (a), other effective area-based
conservation measures (OECMs) (b), and the combined coverage of PAs and OECMs (c). Bars
are categorized by human footprint level: intact, low-disturbance, and highly modified. ......... 153

xi

Acknowledgments
I am deeply grateful to my dissertation committee—Oscar Venter, Chris Johnson,
Moreno Di Marco, and James Watson—for their invaluable guidance, thoughtful feedback, and
encouragement throughout this journey. Oscar, thank you for your mentorship, patience, and the
countless opportunities you provided for me to grow as a scientist. Chris, your perspective and
insights have challenged me to think critically and refine my ideas. Moreno, I am thankful for
your expertise and for broadening the scope of my research. James, I am especially grateful for
the ideas you contributed, your thoughtful feedback, and your generosity in sharing your deep
knowledge of conservation science. I would also like to thank my external examiner, Rob
Serrouya, for his thorough review and insightful suggestions, which helped strengthen the final
version of this dissertation.
My heartfelt thanks go to my partner, Paula Ospina, whose unwavering support, patience,
and understanding have sustained me through the most challenging moments. Her
encouragement, love, and belief in me have been constant sources of strength.
I am also thankful to my friends, colleagues, and collaborators who have supported me
along the way—whether by sharing data and expertise, engaging in stimulating discussions, or
simply being there during long days of work.
Last but not least, to my family—thank you for your love, patience, and belief in me,
even from afar. This work is as much yours as it is mine.

xii

Chapter 1: Introduction
Biodiversity—the variety of life forms across genes, species, and ecosystems—underpins
ecological functions and services upon which human societies depend (Cardinale et al. 2012).
These ecosystem services include nutrient cycling, climate regulation, pollination, and the
provision of goods such as food, medicine, and freshwater (Cardinale et al. 2012, Díaz et al.
2019). Over the past century, however, biodiversity has declined globally at unprecedented rates,
raising concerns about ecosystem stability and human well-being (IPBES 2019). Current
extinction rates are estimated to be tens to hundreds of times greater than natural background
levels, providing strong evidence that the Earth is undergoing its sixth mass extinction event
(Barnosky et al. 2011, Pimm et al. 2014, De Vos et al. 2015, Ceballos et al. 2015).
Recent global assessments underscore the alarming magnitude of biodiversity loss. For
example, vertebrate populations have declined by almost 73% since 1970 (WWF 2024), and
41% of amphibians, 27% of mammals, 21% of reptiles, and 12% of birds are currently classified
as threatened on the International Union for Conservation of Nature (IUCN) Red List (IUCN
2025). These declines reflect profound ecological transformations driven primarily by humaninduced pressures, including habitat destruction, overexploitation, climate change, pollution, and
invasive species (Newbold et al. 2015, Maxwell et al. 2016, Jaureguiberry et al. 2022).

Terrestrial mammals
Among vertebrates, terrestrial mammals have been widely prioritized for monitoring
broad biodiversity trends, given their ecological versatility, cultural significance, charisma, and
functional roles within ecosystems (Ceballos and Ehrlich 2002, Cardillo et al. 2005, Hoffmann et
al. 2011, Ripple et al. 2014). Terrestrial mammals contribute to ecosystem functioning by

13

serving as top predators, keystone herbivores, seed dispersers, and ecosystem engineers (Estes et
al. 2011, Dirzo et al. 2014). Their presence helps regulate vegetation structure, trophic dynamics,
and ecological connectivity (Ripple et al. 2017, Galetti et al. 2018). For instance, the loss of large
mammalian herbivores such as elephants (Loxodonta africana) in Africa and tapirs (Tapirus
terrestris) in South America, as well as apex carnivores like wolves (Canis lupus) in the
Northern Hemisphere and jaguars (Panthera onca) in the Neotropics, can lead to dramatic
ecological changes through altered browsing, grazing, and predation dynamics (Estes et al. 2011,
Ripple et al. 2017, Galetti et al. 2018). These losses can also disrupt trophic cascades by
removing key predators that control prey populations, and reduce ecological connectivity by
limiting seed dispersal and animal movement across landscapes (Terborgh et al. 2001, Dirzo et
al. 2014).
However, despite their ecological importance, terrestrial mammals are experiencing
widespread and accelerating declines. Recent global assessments estimate that up to 20% of
terrestrial mammal species are at risk of rapid extinction (Cardillo et al. 2023), and numerous
others not formally listed as threatened are already undergoing population declines and range
contractions (Ceballos et al. 2017, 2020), reflecting broad vulnerability and reduced resilience to
ongoing environmental disturbances.
In addition to these alarming trends, much of their remaining high-suitability habitat
occurs outside the boundaries of protected areas (PAs)—geographical areas designated and
managed primarily for biodiversity conservation (Dudley 2008)—further elevating their risk of
extinction (Crooks et al. 2017, Cardillo et al. 2023). Anthropogenic pressures—such as
agricultural expansion and hunting—have been identified as primary drivers of these trends, with
the most severe impacts observed in the Indomalayan and Australasian biogeographic realms

14

(Hoffmann et al. 2011). Large-bodied species, characterized by slow reproductive rates, wideranging movements, and specialized habitat needs, have exhibited the most pronounced declines,
underscoring their susceptibility to multiple forms of human pressure (Ripple et al. 2014, Dirzo
et al. 2014, Benítez-López et al. 2017).

Habitat loss and fragmentation
Central to understanding the current biodiversity crisis is recognizing the extensive scale
of human-induced habitat modification. Human activities have transformed between 75% and
95% of Earth’s ice-free land surface, primarily through agricultural expansion, urbanization,
resource extraction, and infrastructure development (Venter et al. 2016, Ellis et al. 2021, Ellis
2021). Habitat fragmentation—the breaking apart of continuous habitat into smaller and isolated
patches—further exacerbates biodiversity loss by reducing habitat area, isolating populations,
disrupting species interactions, and limiting organism movement and dispersal (Haddad et al.
2015). Consequently, fragmented landscapes can affect genetic diversity, population viability,
species richness, and ecological resilience, making habitat fragmentation a central challenge for
global biodiversity conservation (Haddad et al. 2015, Wilson et al. 2016, Pfeifer et al. 2017).
Terrestrial mammals are particularly vulnerable to habitat fragmentation due to their
large spatial requirements, ecological specialization, and restricted dispersal abilities (Fahrig
2003, Crooks et al. 2017, Tucker et al. 2018). Fragmentation disrupts mammalian ecology by
increasing habitat isolation, which restricts dispersal, limits access to habitat resources, and
reduces gene flow, potentially leading to genetic bottlenecks and inbreeding depression
(Cushman et al. 2006, Thatte et al. 2020). Empirical studies have shown that fragmentation
intensifies edge effects—such as increased predation risk, altered species interactions, and

15

microclimatic changes—that disproportionately affect mammalian populations and accelerate
their declines (Pfeifer et al. 2017). Moreover, global analyses consistently indicate that terrestrial
mammals experiencing higher levels of fragmentation within their geographic ranges face an
elevated risk of extinction (Crooks et al. 2017, Ramírez-Delgado et al. 2022).
Historically, global conservation responses have predominantly focused on establishing
PAs to mitigate biodiversity declines (Bruner et al. 2001, Rodrigues et al. 2004, Watson et al.
2014). Although the global PA network has significantly expanded over recent decades (Watson
et al. 2014), it currently covers only around 16% of Earth’s terrestrial ecosystems (UNEPWCMC and IUCN 2025)—far short of the coverage necessary to effectively protect biodiversity
(Allan et al. 2022). Furthermore, many PAs lack ecological connectivity and often fail to
represent critical habitats adequately, particularly in regions experiencing intensive agricultural
and urban development pressures (Allan et al. 2017, Geldmann et al. 2019). Thus, reliance solely
on PAs is increasingly recognized as insufficient to halt biodiversity loss (Watson et al. 2014,
Maxwell et al. 2020, Allan et al. 2022).
For terrestrial mammals, evidence suggests that existing PAs and targeted conservation
actions—while critically important for many species (Pacifici et al. 2020)—remain insufficient
to halt or reverse declining trends (Craigie et al. 2010, Di Marco et al. 2014a, Cardillo et al.
2023). For instance, previous research has shown that for every mammal species whose
conservation status improved due to targeted interventions, approximately seven species
deteriorated, primarily because conservation measures failed to adequately address major threats
such as habitat loss, habitat fragmentation, habitat degradation, and hunting pressures (Hoffmann
et al. 2011).

16

Recognizing these critical shortcomings, recent international policy initiatives, such as
the Kunming-Montreal Global Biodiversity Framework (GBF), aim to conserve at least 30% of
terrestrial and restore at least 30% of degraded ecosystems by 2030 (Obura 2023). Achieving
these ambitious global conservation goals requires transformative strategies extending beyond
traditional PA models (Watson et al. 2018b, Maxwell et al. 2020, Allan et al. 2022). Specifically,
conservation efforts must incorporate biodiversity within human-modified landscapes,
recognizing habitats previously considered marginal or unsuitable for conservation (Kremen and
Merenlender 2018, Ellis 2019, Arroyo‐Rodríguez et al. 2020, Britnell et al. 2023).

The matrix and landscape complexity
Habitat loss and fragmentation, long recognized as major drivers of biodiversity decline
(Newbold et al. 2015, Maxwell et al. 2016, Jaureguiberry et al. 2022), have played a central role
in shaping conservation theory and practice over recent decades (Ricketts 2001, Watson et al.
2005, Laurance 2008). Early conservation frameworks, informed by classic island biogeography
theory, viewed fragmented terrestrial habitats as isolated ‘islands’ embedded within uniformly
hostile ‘matrix’ environments (MacArthur and Wilson 1967, Haila 2002). Under this model,
species richness and extinction risk were primarily determined by patch size and isolation,
emphasizing straightforward colonization-extinction dynamics dependent on habitat area
(MacArthur and Wilson 1967). As a result, traditional conservation strategies predominantly
emphasized protecting large, intact habitat patches through networks of PAs (Margules and
Pressey 2000, Martin et al. 2012), often regarding the surrounding matrix as either ecologically
irrelevant or uniformly inhospitable.

17

However, a growing body of evidence has challenged this simplistic habitat-matrix
dichotomy, revealing that biodiversity responses to habitat loss and fragmentation are more
nuanced and context-dependent than previously recognized (Ricketts 2001, Laurance 2008,
Prugh et al. 2008, Prevedello and Vieira 2010, Watling et al. 2011, Driscoll et al. 2013). These
insights have inspired theoretical advances in landscape ecology, leading to frameworks such as
countryside biogeography (Daily et al. 2001, 2003) and the landscape continuum model (Fischer
and Lindenmayer 2006), which represent habitats within a gradient of ecological suitability
rather than binary classifications of ‘habitat’ versus ‘non-habitat.’
Empirical studies increasingly support this continuum perspective, demonstrating that the
environmental characteristics of the matrix—defined as areas surrounding primary habitat
patches—play a critical role in shaping species persistence, dispersal, and extinction risk (Prugh
et al. 2008, Gardner et al. 2009, Prevedello and Vieira 2010, Watling et al. 2011, Driscoll et al.
2013, Boesing et al. 2018, Ferreira et al. 2018, Ramírez-Delgado et al. 2022). The matrix
typically comprises landscapes modified or managed by human activities, such as agriculture,
forestry, settlements, or infrastructure (Fahrig 2001, Franklin and Lindenmayer 2009). Beyond
simply isolating habitat patches, the matrix provides supplementary resources, facilitates species
movement and dispersal, and can buffer populations against disturbances depending on its
composition and quality (Prevedello and Vieira 2010, Driscoll et al. 2013, Fletcher et al. 2024).
For instance, human-modified habitats, such as secondary forests, agroforestry systems, and
heterogeneous agricultural mosaics, can sustain substantial biodiversity by providing critical
supplementary resources and acting as conduits for dispersal (Daily et al. 2003, Prevedello and
Vieira 2010, Watling et al. 2011, Ferreira et al. 2018). Indeed, such permeable, high-quality
matrices often reduce edge effects, mitigate extinction thresholds, and sustain viable populations

18

even when habitat patches fall below sizes typically required in landscapes dominated by
intensive agriculture or urbanization (Perfecto and Vandermeer 2008, Mendenhall et al. 2014,
Boesing et al. 2018).
Despite these theoretical and empirical advances, existing conservation frameworks—
such as the Key Biodiversity Area (KBA) approach (IUCN 2016) and the Aichi Biodiversity
Targets (CBD 2010)—continue to overlook the matrix, often underestimating its potential
contributions to biodiversity conservation (Prugh et al. 2008, Franklin and Lindenmayer 2009,
Driscoll et al. 2013, Watson et al. 2018a, Kennedy et al. 2019, Galán-Acedo et al. 2019, Arroyo‐
Rodríguez et al. 2020, Fletcher et al. 2024). This persistent oversight highlights an urgent need
for comprehensive global-scale research explicitly evaluating the ecological roles and
conservation potential of matrix areas.
Overall, my findings contribute to our understanding of the conservation potential of the
matrix for terrestrial mammals globally. Specifically, my research provides evidence that the
condition of the matrix can mitigate the impacts of habitat fragmentation on extinction risk.
Additionally, by comparing alternative conceptual models of habitat intactness, my work
provides practical guidance on selecting models that more accurately characterize the
relationship between habitat condition and extinction risk. Finally, by identifying areas within
the matrix harboring high levels of species richness, my research underscores conservation
opportunities that exist beyond primary habitats. Collectively, these findings contribute to a
growing body of evidence highlighting the need to explicitly recognize and integrate the matrix
into global biodiversity frameworks and management strategies. Such integration could
strengthen conservation efforts by better supporting the persistence of terrestrial mammal species
across human-modified landscapes.

19

To fully situate these objectives within the broader context of this research, the following
subsections clarify the rationale for focusing on terrestrial mammals and outline the conceptual
framework used to define and operationalize both habitat and the matrix throughout this
dissertation.

Research objectives
In this dissertation, I evaluate the ecological role and conservation value of the matrix for
terrestrial mammals on a global scale. Specifically, my research aims to:
1. Quantify how the matrix condition influences the effects of habitat fragmentation on
extinction risk in terrestrial mammals.
2. Compare the predictive performance of alternative conceptual models of habitat
intactness (patch-matrix, continuum, and hybrid models) for assessing extinction risk in
terrestrial mammals.
3. Map global patterns of terrestrial mammal species richness within the matrix to identify
conservation opportunities beyond primary habitats.
Overall, my findings contribute to our understanding of the conservation potential of the
matrix for terrestrial mammals globally. Specifically, my research provides evidence that the
condition of the matrix can mitigate the impacts of habitat fragmentation on extinction risk.
Additionally, by comparing alternative conceptual models of habitat intactness, my work
provides practical guidance on selecting models that more accurately characterize the
relationship between habitat condition and extinction risk. Finally, by identifying areas within
the matrix harboring high levels of species richness, my research underscores conservation
opportunities that exist beyond primary habitats. Collectively, these findings contribute to a
20

growing body of evidence highlighting the need to explicitly recognize and integrate the matrix
into global biodiversity frameworks and management strategies. Such integration could
strengthen conservation efforts by better supporting the persistence of terrestrial mammal species
across human-modified landscapes.
To fully situate these objectives within the broader context of this research, the following
subsections clarify the rationale for focusing on terrestrial mammals and outline the conceptual
framework used to define and operationalize both habitat and the matrix throughout this
dissertation.

Why focus on terrestrial mammals?
In light of their ecological roles, conservation concerns, and ongoing declines, terrestrial
mammals represent a compelling focal group for global biodiversity assessments (Ceballos and
Ehrlich 2002, Cardillo et al. 2005, Ripple et al. 2014). Beyond their ecological and conservation
significance, the decision to center this research on terrestrial mammals was also shaped by
practical considerations. At the time this work was initiated, they were the only vertebrate group
for which globally consistent, high-resolution habitat data were available (Rondinini et al. 2011).
These data offered the spatial resolution, taxonomic coverage, and global consistency required to
develop the large-scale analyses presented in Chapter 2 of this dissertation. Although comparable
high-resolution datasets are now available for birds (Lumbierres et al. 2022b), this study was
designed to take advantage of the best available data at the time.

21

Conceptual definition of habitat
In this dissertation, habitat refers to the set of resources and conditions that support
occupancy and, arguably, demographic performance (survival and reproduction) of a given
species (Hall et al. 1997, Krausman and Morrison 2016). Habitat quality refers to the capacity of
an environment to support persistence and should ideally be inferred from demographic evidence
rather than species presence alone. Although habitat is often conceptualized along a continuum
of suitability (Fischer and Lindenmayer 2006), the global-scale analyses in this work required
simplifying this gradient into binary classes of “suitable” and “unsuitable” habitat, as defined by
globally consistent habitat suitability models or Area of Habitat (AOH) maps (Rondinini et al.
2011, Brooks et al. 2019, Lumbierres et al. 2022b). This binary representation enabled tractable,
comparable assessments across thousands of species while recognizing that areas classified as
“unsuitable” may still provide important ecological functions.
In addition, habitat, as defined here in a species-specific sense, emerges from both abiotic
conditions and biotic interactions (Morris 2003, Soberón 2007). Predation, competition, disease
dynamics, and mutualisms can strongly shape realized habitat and demographic performance,
including via trophic cascades (Paine 1969, Schmitz et al. 2000) and interaction-mediated edge
effects (Ries et al. 2004). While these processes are central to many local and regional systems,
they cannot be mapped consistently and comparably for thousands of species at a global scale.
Consequently, the AOH maps applied in Chapters 2 and 4, together with the habitat intactness
models developed in Chapter 3, provide standardized and defensible proxies for habitat condition
but do not explicitly encode interspecific interactions—processes that have been shown to
profoundly influence habitat structure and ecosystem function (e.g., Terborgh et al. 2001, Estes
et al. 2011, Ripple et al. 2017).

22

Conceptual definition of the matrix
In this dissertation, the matrix refers to the land surrounding suitable habitat patches,
regardless of ecological quality or degree of human modification. While much of the
fragmentation literature conceptualizes the matrix as human-modified land that may hinder
movement or reduce suitability (e.g., Fahrig 2003, Haddad et al. 2015, Wilson et al. 2016),
matrix areas can also be entirely natural and undisturbed and, for some species, may provide
suitable habitat (Prugh et al. 2008, Prevedello and Vieira 2010, Galán-Acedo et al. 2019). For
example, in the Brazilian Cerrado, forested gallery patches occur within a natural savanna matrix
(Oliveira-Filho and Ratter 2002). Although the savanna matrix differs ecologically from the
forest patches, it can serve as primary habitat for savanna-adapted species while simultaneously
influencing the persistence and movement of forest-dependent species. In this study, the
operational focus on terrestrial mammal distributions naturally emphasizes land-based matrices,
many of which are human-modified. However, this framing does not exclude the broader
ecological definition in which matrix encompasses any non-focal land-cover type that influences
species persistence and movement.

Dissertation structure
The chapters that follow this introduction (Chapters 2–4) are structured as standalone
manuscripts intended for publication in peer-reviewed journals. Accordingly, these chapters are
written in the first-person plural to acknowledge the contributions of my collaborators. A version
of Chapter 2 has been published in Nature Communications, with M. Di Marco, J.E.M. Watson,
C.J. Johnson, C. Rondinini, X. Corredor Llano, M. Arias, and O. Venter as co-authors (RamírezDelgado et al. 2022). A version of chapter 3 has been accepted for publication in Ecography,

23

with M. Di Marco, C.J. Johnson, J.E.M. Watson, H.L. Beyer, L. de Assis Barros, R. Pillay, and
O. Venter as co-authors (Ramírez-Delgado et al. in press). A version of chapter 4 has been
prepared for submission, with M. Di Marco, J.E.M. Watson, C.J. Johnson, and O. Venter as coauthors (Ramírez-Delgado et al. in prep.). In the concluding chapter, I synthesize the key insights
from these chapters and discuss their broader implications for conservation and future research.

24

2. Chapter 2: Matrix condition mediates the effects of habitat fragmentation on species
extinction risk
Abstract
Habitat loss is the leading cause of the global decline in biodiversity, but the influence of human
pressure within the matrix surrounding habitat fragments remains poorly understood. Here, we
measure the relationship between fragmentation (the degree of fragmentation and the degree of
patch isolation), matrix condition (measured as the extent of high human footprint levels), and
the change in extinction risk of 4,426 terrestrial mammals. We find that the degree of
fragmentation is strongly associated with changes in extinction risk, with higher predictive
importance than life-history traits and human pressure variables. Importantly, we discover that
fragmentation and the matrix condition are stronger predictors of risk than habitat loss and
habitat amount. Moreover, the importance of fragmentation increases with an increasing
deterioration of the matrix condition. These findings suggest that restoration of the habitat matrix
may be an important conservation action for mitigating the negative effects of fragmentation on
biodiversity.

Introduction
Although habitat loss is the leading cause of the ongoing biodiversity crisis (Newbold et
al. 2015, 2016, Maxwell et al. 2016, Betts et al. 2017), the degree to which habitat fragmentation,
defined as the spatial arrangement of remaining habitat for a given amount of habitat loss,
influences the loss of biodiversity has remained the focus of considerable debate (Haddad et al.
2015, Fahrig 2017, 2019, Fletcher et al. 2018, Fahrig et al. 2019, Miller-Rushing et al. 2019).
Central to the debate has been a persistent uncertainty in disentangling the effects of habitat loss

25

on biodiversity from the effects of fragmentation per se, especially relative to the reduction in
patch size and the increase in patch isolation (Fahrig 2003, 2019). While some studies have
challenged the assumption of the impacts of fragmentation (Fahrig 2013, 2017, 2019, Fahrig et
al. 2019), others have demonstrated that the effects of fragmentation are negative and stronger
for local species (Hanski 2015, Haddad et al. 2015, Pfeifer et al. 2017, Fletcher et al. 2018),
particularly in the tropics (Betts et al. 2019) and at intermediate (30–60%) levels of habitat
amount (Pardini et al. 2010, Villard and Metzger 2014). Resolving this debate is critical to not
just informing efforts to prioritize the protection and management of intact and fragmented
landscapes with the same total amount of habitat, but also to better understand the role of the
areas surrounding patches of habitat, commonly referred to as ‘the matrix’, in maintaining
biodiversity (Prugh et al. 2008, Franklin and Lindenmayer 2009, Miller-Rushing et al. 2019).
The traditional characterization of landscapes, which views patches of habitat as islands
embedded in a matrix of ‘non-habitat’, as assumed in classical theoretical models (MacArthur
and Wilson 1967, Haila 2002), has been strongly criticized (Watson 2002, Watson et al. 2005,
Prugh et al. 2008, Mendenhall et al. 2014). This characterization has progressively been relaxed
with approaches based on the premise that the matrix should be treated as a heterogeneous
mosaic of different land covers (e.g., ‘countryside biogeography’ (Daily et al. 2003), and the
‘land-sharing’ and ‘land-sparing’ approaches (Green et al. 2005, Perfecto and Vandermeer 2008,
Law and Wilson 2015, Phalan 2018, Balmford et al. 2019)), as it is recognized that species use
different matrices for foraging, dispersing, and reproduction purposes (Prevedello and Vieira
2010, Ferreira et al. 2018). While high-contrast matrices (e.g., intensive agricultural or built
environments) act as movement barriers or ecological traps with an elevated risk of mortality for
many species (Battin 2004), low-contrast matrices (e.g., secondary forests or shade-grown low-

26

intensive agriculture in forested regions) may act as permeable barriers with a reduced risk of
mortality for many others, even for those typically considered as habitat specialists (Daily et al.
2003, Perfecto and Vandermeer 2008). To date, however, conservation and management
assessments have focused mainly on species’ primary habitat (Martin et al. 2012, Di Marco et al.
2019), limiting our understanding of their response to the habitat matrix, which may have direct
implications for the design of functional landscapes (Fahrig et al. 2011) and the prioritization of
conservation actions in fragmented landscapes (Arroyo‐Rodríguez et al. 2020).
Comparative extinction risk modelling is an approach for assessing the drivers of
extinction risk and the change in risk over time. These models are based on the relationship
between species’ life histories, the habitat pressures within species’ geographic ranges, and their
threat status (Purvis et al. 2000, Fisher et al. 2003, Cardillo et al. 2005, Davidson et al. 2009, Di
Marco et al. 2015, 2018b). Built with readily available data, this approach allows for the
prediction of the risk of extinction of a larger number of species compared with that provided by
expert-based assessments. This more rapid approach can substantially reduce resource
requirements, as well as proactively inform conservation and management strategies (Rondinini
et al. 2014, Bland et al. 2015a). However, despite the fact that the loss and fragmentation of
habitat are among the main determinants of species extinction risk (Maxwell et al. 2016, Crooks
et al. 2017, Di Marco et al. 2018b, Lucas et al. 2019), the influence of the matrix condition on
the effects of fragmentation and its relationship with the risk of extinction has not been well
evaluated for any animal taxon at a global scale.
Here, we quantify the relationship between changes in the extinction risk of 4,426
terrestrial mammals over a 24-year period (1996-2020), the fragmentation of their suitable
habitat (in terms of the degree of fragmentation and the degree of patch isolation), and the levels

27

of human pressure within the associated habitat matrix. Our goal is to test the influence of human
pressure within the matrix on the effects of fragmentation for determining changes in species
extinction risk globally. We focus on terrestrial mammals as they have been used as a focal taxon
in previous extinction risk analyses (Arregoitia 2016), they are known to be sensitive to
fragmentation (Crooks et al. 2017), and data are available to delineate levels of suitable habitat
(i.e., high and medium habitat suitability) and unsuitable habitat (i.e., the matrix) within their
ranges (Rondinini et al. 2011). For each species, we quantify the degree of fragmentation as the
average Euclidean distance into ‘core’ suitable habitat from the nearest patch edge, the degree of
patch isolation as the average Euclidean distance between patches of suitable habitat through the
surrounding matrix, and the matrix condition as the extent of high human pressure levels
overlapping with areas of unsuitable habitat. Spatial data representing the condition of the matrix
were obtained from the recently updated human footprint maps (Venter et al. 2016, Williams et
al. 2020), which provide a single metric of the combined area and intensity of human activities,
all of which are driving the current biodiversity crisis (Maxwell et al. 2016). We define a human
footprint threshold of ≥ 3 out of 50 to represent the extent of human-modified habitat within the
matrix. This threshold was used as it has been shown to be the strongest predictor of transitions
in extinction risk for terrestrial mammals (Di Marco et al. 2018b). Furthermore, this human
footprint threshold is associated with the highest declines in mammalian movements (Tucker et
al. 2018). Following previous studies (Hoffmann et al. 2010, Di Marco et al. 2015, 2018b), we
classify species into two groups of extinction risk, ‘low-risk’ transitions and ‘high-risk’
transitions, based on the first and last Red List category registered between 1996 and 2020. In
combination with other predictors of extinction risk (see Methods, Table 2.1), we quantify the

28

relative importance of the degree of fragmentation, the degree of patch isolation, and the
condition of the matrix for determining extinction risk transitions in terrestrial mammals.
Our analyses reveal that the condition of the matrix plays a major role on the effects of
fragmentation for predicting extinction risk transitions in terrestrial mammals. Our results
suggest that the negative effects of fragmentation may be somewhat mitigated when the matrix is
associated with lower levels of human pressure.

Methods
Habitat suitability models
We used habitat suitability models developed by Rondinini et al. (2011) to represent the
extent of suitable habitat patches and the extent of the matrix of 4,426 out of 5,709 extant
terrestrial mammals, corresponding to ~78% of all species in the group (IUCN 2021). The
models were limited to occur within the known geographic range of each species (i.e., the current
“limits of distribution of a species, accounting for all known, inferred or projected sites of
occurrence”, as defined by the IUCN Red List of Threatened Species (IUCN 2016)), and built
for the year 2000 at a spatial resolution of 300 m based on species’ elevation range and other
habitat affinities, including preferred land cover types, tolerance to human impact, and
relationship to water bodies. Species’ elevation range was incorporated into the habitat suitability
models when known and recorded in the IUCN Red List. Textual descriptions of other habitat
affinities for each species, derived from the input of thousands of mammal experts belonging to
more than 30 specialist groups of the IUCN Species Survival Commission (IUCN/SSC)
(Schipper et al. 2008), were also extracted from the IUCN database and input as quantitative data
into the habitat suitability models. The models ranked areas with three levels of habitat

29

suitability: (i) high, representing primary habitat or preferred habitat where the species can
persist; (ii) medium, representing secondary habitat where the species can occur but not persist
without nearby high suitable habitat; and (iii) unsuitable, representing locations where the
species is expected to occasionally or never be found. A subset of the models and their
associated levels of habitat suitability were validated against available points of known species
occurrences. Full details on the development of the models are available elsewhere (Rondinini et
al. 2011).
When delineating the levels of habitat suitability for each species, small contiguous
groups of pixels ( < 4 adjacent pixels of the same level of habitat suitability) were removed and
replaced with the pixel value of the largest and nearest group of pixels, based on eight
neighboring pixels of the same class. This reduced the influence of isolated groups of pixels of
the same level of habitat suitability, and improved the computational efficiency of the analysis,
as also reported in other studies (Crooks et al. 2011, 2017).
For our analysis, we combined high and medium habitat suitability to represent the extent
of suitable habitat patches, and used the level of unsuitable habitat to represent the extent of the
matrix of each species. We also applied a different combination of the levels of habitat suitability
when representing the extent of suitable habitat patches (high suitability instead of high and
medium suitability combined) and the extent of the matrix (medium suitability and unsuitable
combined instead of unsuitable habitat alone) of each species as a sensitivity analysis (see
Sensitivity analysis section).

30

The degree of habitat fragmentation and the degree of patch isolation as predictors of
extinction risk transitions
For each species, we measured the degree of habitat fragmentation by calculating the
average Euclidean distance of all the pixels of suitable habitat from the nearest edge (Ripple et
al. 1991), edges demarcated by the boundary between suitable and unsuitable habitat. Large
values of the average Euclidean distance represented low degrees of habitat fragmentation,
whereas small values represented high degrees of habitat fragmentation. Additionally, we
calculated the average Euclidean distance between patches of suitable habitat through the
surrounding matrix (i.e., the average Euclidean distance of all the pixels of unsuitable habitat
from the nearest edge) to account for patch isolation (after Crooks et al. 2017). Here, large values
of the average Euclidean distance represented high degrees of patch isolation, and small values
represented low degrees of patch isolation. The average Euclidean distance was considered
because this metric does not require a predetermined distance threshold of what constitutes an
edge, accounts for different shapes of fragments and landscapes patterns and arrangements,
accounts for the distribution of habitat area (Li and Archer 1997), is comparable across
landscapes of different extents, and provides stable and readily interpretable information (Crooks
et al. 2011). Moreover, average Euclidean distance has been shown to be singularly valuable in
quantifying the relationship between habitat fragmentation and extinction risk of the world’s
terrestrial mammals (Crooks et al. 2017), which made it highly suitable for our analyses.

The matrix condition as a predictor of extinction risk transitions
Spatially explicit data on the condition of the matrix, as represented by the extent and
change over time of high human pressure levels overlapping with the area of unsuitable habitat

31

surrounding patches of suitable habitat (after Di Marco et al. 2018), was obtained from the
recently updated global human footprint maps (Williams et al. 2020). These maps represent the
most comprehensive global distribution of changing human pressure on the environment at 1 km
resolution between 2000 and 2013, based on eight pressure layers (Venter et al. 2016): (i) built
environments; (ii) intensive agriculture; (iii) pasture land; (iv) human population density; (v)
night-time lights; (vi) roads; (vii) railways; and (viii) navigable waterways, all of which are
driving the current extinction crisis (Maxwell et al. 2016). Each human footprint map provides a
single pressure metric ranging from 0 to 50, where a value of 0 represents areas free of any
human influence (e.g., terrestrial remaining wilderness), values of 4 or below represent areas of
low human pressure (e.g., pasture lands), and values above 20 represent areas with very high
pressure levels (e.g., densely populated semi-urban and urban environments).
In this analysis, we measured the extent of high human footprint values and the change of
this extent over time (between 2000 and 2013) within areas of unsuitable habitat, using a defined
human footprint threshold of 3 or above. This threshold was used as it has shown to be the
strongest predictor of extinction risk transitions in terrestrial mammals (Di Marco et al. 2018b).
Moreover, this human footprint threshold is associated with the highest declines in mammalian
movements (Tucker et al. 2018). Based on previous studies (Di Marco et al. 2013, 2018b), we
used the extent of high human footprint values within the matrix as the extent of high pressure
levels within species’ ranges has been shown to be more sensitive to predict extinction risk than
using mean values of human pressure within species’ ranges. We also considered the change in
the extent of high human footprint values after discarding areas where the human footprint was
lower in 2013 than in 2000 (assuming no change in these particular areas), as decreases in human

32

pressure levels are likely to take time before having a measurable effect on species threat status,
particularly for species with a long generation time period (Di Marco et al. 2018b).

Changes in species extinction risk
We used the IUCN Red List of Threatened Species (IUCN 2012, 2021), the retrospective
Red List assessments published in Hoffmann et al. (2010), and the IUCN list of genuine changes
in the conservation status of mammal species (https://www.iucnredlist.org/resources/summarystatistics) to represent trends in extinction risk of terrestrial mammals. Following the
classification of extinction risk transitions developed by Di Marco et al. (2015, 2018), we
classified the species into two main groups, ‘low-risk’ transitions and ‘high-risk’ transitions (Fig.
2.1). The low-risk group included species that retained a category of least concern, together with
those species that moved from any higher category of threat to a lower category assessment
period. The high-risk group included all species that retained a category of threatened or near
threatened, together with those species that moved from any lower category of threat to a higher
category over time.
For our analysis, we classified species into the two extinction risk groups (low-risk
transitions and high-risk transitions) based on the first and last Red List category registered
between 1996 and 2020. In order to test the sensitivity of this classification, we also classified all
species into the two extinction risk groups based on the last two Red List assessments registered
between 1996 and 2020 (i.e., the second to last and last Red List categories registered during this
time period). With this classification, however, only two species (0.05% of 4,426 species in our
sample) changed their extinction risk transition (from a high-risk transition to a low-risk
transition) compared to that based on the first and last Red List category registered during the

33

study period. We thus only reported the main results using the first and last Red List category
registered between 1996 and 2020.
We excluded species without a defined level of habitat suitability, those not evaluated in
the Red List, and those categorized as Data Deficient, Extinct and Extinct in the Wild in the last
Red List assessment reported during the study period, as long as they have not shown a defined
transition of extinction risk (see Fig. 2.1) along the study period.

Predicting extinction risk transitions
We used a multivariate Random Forest model to predict extinction risk transitions in
terrestrial mammals (Fig. 2.1). Random Forest is a non-parametric, tree-based, machine-learning
technique that produces multiple decision trees using a randomly selected subset of training
samples and variables to make a prediction (Breiman 2001, Cutler et al. 2007). Due to its limited
assumptions on data distributions, its high classification stability and performance, and its ability
to cope well with a large number of potentially correlated predictors and non-linear responses,
Random Forest is a highly suitable technique for species threat status classification (Murray et al.
2014). Furthermore, Random Forest modelling has demonstrated superior performance among
several machine learning techniques tested for the prediction of global extinction risk of
terrestrial mammals (Bland et al. 2015b), making it ideal for this study.
In this analysis, we optimized the number of trees to grow and the number of predictors
sampled for splitting at each node from 3 repeats of 10-fold cross-validation, using 75% of the
data as training data and 25% as test data. Predictors included: (i) the extent of high human
footprint values in the matrix; (ii) the extent of high human footprint values in patches of suitable
habitat; (iii) the change over time of high human footprint values in the matrix; (iv) the change

34

over time of high human footprint values in patches of suitable habitat; (v) the degree of
fragmentation of suitable habitat; (vi) the degree of isolation between patches of suitable habitat;
(vii) the proportion of suitable habitat; and (viii) the biogeographic realm in which the species
can be encountered (see Table 2.1 for a description). Because mammals of greater body size
usually move farther (Jetz et al. 2004), and diet may influence their movements as a result of
differences in availability of resource types and foraging cost (McNab 1986, Tucker et al. 2014),
we decided to include body size and dietary breadth as life-history predictors. We also included
the reproductive traits, weaning age, and gestation length. Other life-history traits were broadly
captured by including taxonomic orders. Because the levels of habitat suitability are limited by
the size of species’ geographic ranges, we did not include species’ range size as a predictor in
order to avoid potential circularity in the estimation of extinction risk (Purvis et al. 2000).

Table 2.1: Description of the selected variables to predict extinction risk transitions in
terrestrial mammals.
Class

Variable

Description

Source

Pressure

High human footprint

Proportion of unsuitable

Rondinini et al.

extent in the matrix

habitat overlapping with high

2011, Williams et al.

human footprint values in

2020

2000.
High human footprint

Proportion of suitable habitat

Rondinini et al.

extent in patches of

overlapping with high human

2011, Williams et al.

suitable habitat

footprint values in 2000.

2020

35

High human footprint

Difference in the proportion

Rondinini et al.

change in the matrix

overlap between the area of

2011, Williams et al.

unsuitable habitat and high

2020

human footprint values during
2000 and 2013.
High human footprint

Difference in the proportion

Rondinini et al.

change in patches of

overlap between the area of

2011, Williams et al.

suitable habitat

suitable habitat and high

2020

human footprint values during
2000 and 2013.
Environment Degree of habitat
fragmentation

Average of the Euclidean

Ripple et al. 1991,

distance from the edge to the

Rondinini et al. 2011

‘core’ (i.e., the interior) of each
patch of suitable habitat.
Degree of patch

Average of the Euclidean

isolation

distance between patches of

Crooks et al. 2017

suitable habitat from the edge
to the ‘core’ (i.e., the interior)
of each area of unsuitable
habitat.
Proportion of suitable

Proportion of suitable habitat

habitat

within the range of each

Rondinini et al. 2011

species.

36

Realm

Biogeographic realm in which

IUCN 2012, 2021

the species can be encountered.
Life-history

Body mass

A generic proxy of species life

Smith et al. 2003,

history and energetic

Jones et al. 2009,

requirements.

Tacutu et al. 2013,
Verde Arregoitia et
al. 2013, Faurby et
al. 2018

Diet

Dietary categories: vertebrate

Kissling et al. 2014,

carnivore ( > 90% vertebrate

Wilman et al. 2014,

matter ingested), invertebrate

Faurby et al. 2018

carnivore ( > 90% invertebrate
matter ingested), omnivore
(10-90% animal matter
ingested or 10-90% plant
matter ingested), herbivore ( >
90% plant matter ingested).
Weaning age

Gestation length

Order

A proxy of species

Jones et al. 2009,

reproductive timing.

Tacutu et al. 2013

A proxy of species

Jones et al. 2009,

reproductive output.

Tacutu et al. 2013

Species taxonomic order.

IUCN 2012

37

We measured the predictive importance of each variable using the mean decrease in
classification accuracy (MDA) metric (Breiman 2001), which reports the model’s ability to
correctly classify data if the values of a predictor variable are randomly permuted. Based on this
metric, we then calculated the relative importance of each variable using the model improvement
ratio (MIR) metric (Murphy et al. 2010), which scales raw importance scores from 0 to 1. Unlike
the raw importance scores, the MIR metric is not influenced by the total number of variables and
is comparable among models. MIR is calculated as [In/Imax], where In is the importance of a given
variable, and Imax is the maximum model improvement score. We also reported the overall
performance of the Random Forest model through cross-validation in terms of proportion of
correctly classified species (accuracy), proportion of correctly classified high-risk species
(sensitivity), proportion of correctly classified low-risk species (specificity), and the true skill
statistic (TSS = sensitivity + specificity – 1) (Allouche et al. 2006).

Assessing the influence of the matrix on the importance of fragmentation for predicting
extinction risk transitions
To measure the influence of the matrix condition on the importance of habitat
fragmentation (in terms of the degree of fragmentation and the degree of patch isolation) for the
prediction of extinction risk transitions, we first defined two broad levels of quality of the matrix,
‘low-quality’ matrices and ‘high-quality’ matrices, based on the proportion of high human
footprint values within the matrix of each species. When delimiting the two levels of quality of
the matrix, the extent of high human footprint values in the matrix of each species was
discretized into two intervals based on the positive and negative effect that the matrix condition
had on the probability of high-risk transitions (see Fig. 2.3c). We then built separate Random

38

Forest models for species restricted to such levels of quality of the matrix. Using the MDA
metric (Breiman 2001) and the MIR metric (Murphy et al. 2010), we measured the relative
importance of the degree of habitat fragmentation and the degree of patch isolation, including the
other selected predictors of extinction risk (Table 2.1), from the built Random Forest models. We
used cross-validated measures of accuracy, sensitivity, specificity, and the true skill statistic to
evaluate the overall performance of the models (Allouche et al. 2006). We also used Wilcoxon
rank sum tests to test for statistical differences in the degree of fragmentation and patch isolation
between low-risk and high-risk species restricted to the defined levels of quality of the matrix. In
order to determine the effect size of the degree of habitat fragmentation and patch isolation
between low-risk and high-risk species for each of the levels of quality of the matrix, we used
Cohen’s d statistic (Cohen 1988).

Sensitivity analysis
To test the sensitivity of our model, we built additional Random Forest models based on a
different combination of the levels of habitat suitability to represent the extent of suitable habitat
patches (high suitability instead of high and medium suitability combined) and the extent of the
matrix (medium suitability and unsuitable combined instead of unsuitable habitat alone). From
these models, the relative importance of each variable was quantified using the MDA metric
(Breiman 2001) and the MIR metric (Murphy et al. 2010). The overall performance of these
models was reported through cross-validation in terms of accuracy, sensitivity, specificity, and
the true skill statistic (Allouche et al. 2006).
All spatial analyses were performed in python using the ArcPy processing module from
ArcGIS Pro 2.8.2 (ESRI 2021). Statistical analyses were performed in R (R Core Team 2021),

39

using the packages ‘randomforest’ (Liaw and Wiener 2002), ‘caret’ (Kuhn 2008), ‘iml’ (Molnar
and Schratz 2020), and ‘effsize’ (Torchiano 2017).

Results
Changes in species extinction risk
We found that 2,984 (67.4%) terrestrial mammals faced a low-risk transition and 1,442
(32.6%) a high-risk transition between 1996 and 2020 (Fig. 2.1). A total of 4,124 species
(93.2%) retained the same Red List category, while 302 (6.82%) changed their category through
time (Appendix A, Fig. A.1).

Figure 2.1: Classification of species extinction risk transitions based on past and present
IUCN Red List categories. Low-risk transitions included species that retained a category of
least concern, together with those species that moved from any higher category of threat to a
lower category between 1996 and 2020 (a). High-risk transitions included all species that
retained a category of threatened or near threatened, together with those species that moved from
any lower category of threat to a higher category between 1996 and 2020 (b). Acronyms refer to
the IUCN Red List categories, including Least Concern (LC), Near Threatened (NT), Vulnerable
(VU), Endangered (EN), and Critically Endangered (CR).
40

Predicting transitions in species extinction risk
We used a Random Forest model for classification (Breiman 2001) to measure the
performance of an array of pressure, environmental and life-history variables (Table 2.1) for the
prediction of extinction risk transitions in terrestrial mammals. We found that the degree of
fragmentation of suitable habitat had higher predictive performance than species life-history
traits, human pressure variables, and other environmental conditions (Fig. 2.2).

Figure 2.2: Relative importance of selected variables for the prediction of extinction risk
transitions in terrestrial mammals. Variables are colour-coded according to their broad class
(human pressure, environment, and life-history). The description of each variable can be found in
Table 2.1. High levels of the human footprint (HFP) included values of 3 or above.

Interestingly, our results show that the degree of fragmentation, the extent of high human
footprint values in the matrix, and the degree of patch isolation had higher predictive
performance than the change in high human footprint values (as defined by increases in high
human footprint values through time) within suitable habitat and the proportion of suitable
habitat (Fig. 2.2). This result was supported by a sensitivity analysis where a different
41

combination of the levels of habitat suitability was applied (Appendix A, Fig. A.2a,b). This
suggests that habitat fragmentation and the matrix condition better predict changes in species
extinction risk than habitat loss and habitat amount at a global scale.
Partial dependence plots show that the probability of high-risk transitions is higher with
an increasing degree of fragmentation (Fig. 2.3a), a decreasing degree of patch isolation (Fig.
2.3b), and an extent of high human footprint values within the matrix of 100% (Fig. 2.3c).

Figure 2.3: Partial dependence plots to show the effect of the degree of habitat
fragmentation, the degree of patch isolation, and the matrix condition on extinction risk
transitions in terrestrial mammals. The plots show the probability of high-risk transitions as a
function of the degree of fragmentation (a), the degree of patch isolation (b), and the extent of
high human footprint values within the matrix (c). Solid red lines and shading represent fitted
LOESS curves and 95% credible intervals for the relationships between the probability of highrisk transitions and each explanatory variable. As partial dependence plots for Boolean response
variables mirror each other, the probability of low-risk transitions as a function of these variables
are not depicted in the figure. Values of the degree of fragmentation and the degree of patch

42

isolation were ln-transformed for visual purposes. The degree of fragmentation was inversecoded so high values represent high degrees of fragmentation. High values of the degree of patch
isolation represent high degrees of isolation between patches of suitable habitat. The description
of each variable is given in Table 2.1. High levels of the human footprint (HFP) included values
of 3 or above.

Our model showed good overall classification ability during cross-validation, with an
accuracy of 81.2%. The proportion of correctly classified high-risk transitions (sensitivity =
60.5%) was lower than the proportion of correctly classified low-risk transitions (specificity =
90.9%), with a true skill statistic of 0.51.
The predictive performance of our model did not markedly change compared to the
model built with a different combination of the levels of habitat suitability (Appendix A, Table
A.1). Thus, the model is robust to changes in the levels of habitat suitability.

The influence of the matrix condition on the importance of habitat fragmentation for
predicting extinction risk transitions
In order to measure the influence of the matrix condition on the importance of
fragmentation (i.e., the degree of fragmentation and the degree of patch isolation) for the
prediction of extinction risk transitions, we first discretized the extent of high human footprint
values within the matrix of each species into two broad levels as a proxy for matrix quality:
‘low-quality’ matrices and ‘high-quality’ matrices. As the global distribution of the matrix
condition showed to be uneven in both low-risk and high-risk species (Appendix A, Fig. A.3),
even at the scale of individual biogeographic realms (Appendix A, Fig. A.4), we defined cutoff

43

values for each of the levels of quality of the matrix based on the positive and negative effect that
the matrix condition had on the probability of high-risk transitions (Fig. 2.3c). Low-quality
matrices were therefore represented by species with extents > 84.2% of their matrix overlapping
with high human footprint values (n = 1,815 low-risk species and 1,027 high-risk species), while
high-quality matrices by those species with extents < 15.8% of their matrix overlapping with
high human footprint values (n = 60 low-risk species and 29 high-risk species). We then built
separate Random Forest models for each level of quality of the matrix in order to compare the
relative importance of the degree of fragmentation of suitable habitat and the degree of isolation
of patches of suitable habitat between species with a matrix of low-quality habitat and species
with a matrix of high-quality habitat.
We found that the degree of fragmentation and the degree of patch isolation had higher
relative importance for species with a low-quality matrix (Fig. 2.4a) than that observed for
species with a high-quality matrix (Fig. 2.4b), with a decrease of 33.3% and 62.5%, respectively.
Notably, the relative importance of the extent of high human footprint values in the matrix was
markedly higher for those species with a low-quality matrix than for those with a high-quality
matrix, with a decrease of 116.4%, suggesting that the lower the quality of the matrix, the higher
the predictive importance of the matrix for predicting extinction risk transition in terrestrial
mammals.

44

Figure 2.4: Influence of the matrix condition on the relative importance of selected
variables for the prediction of extinction risk transitions in terrestrial mammals. Relative
importance of each predictor for species with a low-quality matrix (a), which included
proportions > 84.2% of the extent of their matrix overlapping with high human footprint values
(n = 1,815 low-risk species and 1,027 high-risk species). Relative importance of each predictor
for species with a high-quality matrix (b), which included proportions < 15.8% of the extent of
their matrix overlapping with high human footprint values (n = 60 low-risk species and 29 high-

45

risk species). Variables are colour-coded according to their broad class (human pressure,
environment, and life-history). The description of each variable is given in Table 2.1. High levels
of the human footprint (HFP) included values of 3 or above.

When looking at the difference in the degree of fragmentation and the degree of patch
isolation between low-risk and high-risk species with a low-quality matrix, we found that both
variables were significantly higher for those species classified as high-risk (p-values < 0.001;
Wilcoxon rank sum test, one-sided). We also found that the difference in the degree of
fragmentation and the degree of patch isolation between low-risk and high-risk species with a
high-quality matrix was not statistically significant (p-values > 0.05; Wilcoxon rank sum test,
one-sided). The degree of fragmentation and the degree of patch isolation showed a greater effect
size between low-risk and high-risk species with a matrix of low-quality habitat, with an
estimated Cohen’s d of 0.23 and 0.32, respectively (Appendix A, Fig. A.5a,b). This indicates a
greater effect of both the degree of fragmentation and the degree of isolation between patches of
suitable habitat on the risk of extinction of those terrestrial mammals with a matrix of lowquality habitat.
Our results show that the model for species with a matrix of low-quality habitat had
higher predictive performance (true skill statistic = 0.57) than the model for species with a matrix
of high-quality habitat (true skill statistic = 0.36). Our results also show that the classification
ability was higher in the model for species with a low-quality matrix (accuracy = 0.81) compared
to that shown in the model for species with a high-quality matrix (accuracy = 0.76). Although the
proportion of correctly classified high-risk transitions was higher in the model for species with a
low-quality matrix (sensitivity = 68.4%), relative to that shown in the model for species with a

46

high-quality matrix (sensitivity = 42.9%), the model for species with a low-quality matrix had a
lower proportion of correctly classified low-risk transitions (specificity = 88.7%) than the model
for species with a high-quality matrix (specificity = 92.9%). This indicates a higher imbalance
between the proportion of low-risk species and high-risk species correctly classified in the model
for species with a high-quality matrix.

Discussion
Understanding the external conditions under which a species is likely to face an increased
risk of extinction are necessary to inform conservation policies and management strategies (Di
Marco et al. 2015). We found that the condition of the matrix, as defined by the extent of high
human footprint values between patches of suitable habitat, strongly influenced the effects of
fragmentation on extinction risk transitions of terrestrial mammals. Specifically, we found that
the degree of fragmentation and the degree of patch isolation had a higher relative importance for
species with a matrix of low-quality habitat compared to those with a matrix of high-quality
habitat when determining extinction risk transitions in terrestrial mammals. To the best of our
knowledge, these findings are the first to demonstrate the extent to which human pressure within
the matrix alters the importance of fragmentation metrics as predictors of extinction risk
transitions in terrestrial mammals. These findings are in line with previous studies showing that
the use of the matrix is among the main determinants of the vulnerability of mammalian
populations to local extinction in fragmented landscapes (e.g., Laurance 1991, Daily et al. 2003,
Viveiros de Castro and Fernandez 2004), and support recent findings showing that species-area
relationships are steeper (i.e., more extinction driven) in forested landscapes with a low-quality
matrix, and shallower (i.e., less extinction driven) in those forested landscapes with a higher

47

quality matrix (Reider et al. 2018). This suggests that the magnitude of the effects of
fragmentation depends on the structural similarity between suitable habitat patches and the
matrix, as also suggested by a growing body of evidence across multiple taxa on a local scale
(Ewers and Didham 2006, Prevedello and Vieira 2010).
Our results showed that species with a greater degree of fragmentation, a lower degree of
patch isolation, and a lower quality matrix within their ranges tended to be at greater risk of
extinction. This indicates that the persistence of terrestrial mammals depends not only on the
proportion of suitable habitat and its spatial configuration, but also on the quality of the matrix.
This pattern may reflect the fact that species occurring in regions with low rates of historical
disturbance are more likely to be sensitive to fragmentation (Betts et al. 2019), and thus more
likely to face an increased risk of extinction. That would suggest that those species within the
high-risk group are mainly concentrated in the tropics, particularly in forested landscapes where
deforestation continues at a rapid rate (Schipper et al. 2008, Tracewski et al. 2016). In our
sample, the majority of these species (68.4%) were restricted to the Neotropical (32.0%),
Afrotropical (34.9%), and Indo-Malay (48.0%) biogeographic realms, which is consistent with
these findings.
In our global analysis, the degree of patch isolation refers to the average Euclidean
distance between patches of suitable habitat across a species’ entire range, rather than local
dispersal distances within a single landscape. High isolation values can indicate widely separated
patches—sometimes occurring on different continents—rather than fragmentation of a
continuous habitat. The loss of any single patch does not necessarily elevate extinction risk if
other distant patches remain intact. While local-scale studies consistently show that greater
isolation between patches reduces colonization, limits gene flow, and increases extinction

48

probability (Fahrig 2003, Haddad et al. 2015), larger-scale research reveals more complex, scaledependent patterns. For example, variation in the arrangement of habitat patches across broad
areas can enhance regional diversity (Perrin et al. 2025), and networks of small but widely
distributed patches may support higher species richness even when total habitat area remains the
same (Riva and Fahrig 2023). Our finding—that high-risk transitions decreased with increasing
patch isolation—should therefore be seen as a global-scale effect, where broad geographic
separation between patches can buffer species against synchronized threats, rather than as
evidence that isolation is beneficial at local scales.
In our extinction risk model, some variables had higher predictive performance than
others. For example, the degree of fragmentation of suitable habitat was shown to be the most
important predictor of changes in species extinction risk when compared with species life-history
traits, measures of human pressure, and other environmental conditions. This finding is in line
with previous extinction risk modelling, showing that the inclusion of the degree of
fragmentation as a predictor increases the explanatory power of the models (Crooks et al. 2017,
Lucas et al. 2019). In particular, this result supports recent findings showing that terrestrial
mammals with a higher degree of fragmentation have smaller ranges, lower proportions of
suitable habitat, and are at greater risk of extinction (Crooks et al. 2017).
The second most important predictor of extinction risk transitions was the extent of high
human footprint values within the matrix (i.e., the condition of the matrix). This result contrasts
with the findings from previous extinction risk modelling exercises for mammals, where the
predictive importance of human pressure was found to be lower than life-history traits or
environmental conditions different from fragmentation (Cardillo et al. 2008, Murray et al. 2014,
Arregoitia 2016). However, it complements the findings of one recent extinction risk modelling

49

exercise for mammals (Di Marco et al. 2018b), where the extent of high human footprint values
within species’ ranges had higher predictive importance than species life-history traits,
environmental conditions (without consideration of habitat fragmentation), and other pressure
variables. Given that species are not homogeneously distributed throughout their ranges
(Rondinini et al. 2006, 2011), this result specifically suggests that the condition of the matrix
surrounding patches of suitable habitat is strongly correlated with extinction risk transitions in
terrestrial mammals. This may in part be explained by the fact that habitat loss and fragmentation
have opened up the path to a series of other threat mechanisms through the matrix, such as
hunting, disease spread, and invasive species (Cardillo et al. 2005, Haddad et al. 2015, Betts et
al. 2017). It may also be related to the fact that species are increasingly obligated to inhabit
human-modified landscapes (Watling et al. 2011, Galán-Acedo et al. 2019), many of which have
a matrix that likely prevents their movement (Tucker et al. 2018) and elevates their mortality
(Battin 2004) (e.g., by roadkill (Fahrig and Rytwinski 2009) or increasing predation (May and
Norton 1996, Woinarski et al. 2015)).
Conflicting results on the effects of fragmentation on biodiversity have arisen from
studies attempting to separate ‘independent’ effects of habitat loss from those of habitat
fragmentation (Fahrig 2003). Some studies have argued that the effects of habitat loss are greater
and more negative (e.g., Fahrig 2013, 2017, 2019, Fahrig et al. 2019), while others have
demonstrated that the effects due to fragmentation (such as declining patch size, increasing
habitat isolation, and increasing edge effects) are essentially negative and lasting (e.g., Hanski
2015, Haddad et al. 2015, Pfeifer et al. 2017, Fletcher et al. 2018). However, in real landscapes,
habitat loss inevitably causes habitat fragmentation, and both act in synergy with other threats to
biodiversity (Peres 2001, Laurance and Useche 2009, Côté et al. 2016). Thus, there is little

50

practical value in attempting to separate the effects of habitat loss and fragmentation (Didham et
al. 2012, Ruffell et al. 2016, Morante‐Filho et al. 2018, Fletcher et al. 2018). Our study does not
attempt to resolve the current debate as to whether and how habitat fragmentation per se (i.e., the
spatial arrangement of remaining habitat for a given amount of habitat loss) influences
biodiversity (Haddad et al. 2015, Fahrig 2017, 2019, Fletcher et al. 2018, Fahrig et al. 2019,
Miller-Rushing et al. 2019), but our results showed that the degree of fragmentation of suitable
habitat, the extent of high human footprint values within the matrix, and the degree of isolation
between patches of suitable habitat were more important predictors of extinction risk transitions
than the change in high human footprint values (as represented by increases in high human
pressure levels over time) within patches of suitable habitat and the proportion of suitable
habitat. This suggests that changes in species extinction risk are primarily determined by the
fragmentation of habitat and the matrix condition, and secondarily by the loss and the amount of
habitat within species’ ranges. However, there is also the possibility that the loss of most suitable
habitat patches had already occurred before the beginning of the study period, resulting in the
degree of fragmentation, the matrix condition, and the degree of patch isolation being more
important predictors of extinction risk than habitat loss and habitat amount.
Although biogeography was not a key parameter for determining extinction risk
transitions in our models, we found some differences in the way the matrix condition was
distributed in low-risk and high-risk species among biogeographic realms. The Indo-Malay realm
represented a particular case, with a highly left-skewed distribution (i.e., towards a higher extent
of high human footprint values within the matrix) in species classified as low-risk, very similar
to that shown in high-risk species. With ~87% of terrestrial mammals showing a low-quality
matrix in the Indo-Malayan realm, this might indicate that species living in the Indo-Malayan

51

realm are relatively more resilient to those human activities included in the human footprint, but
more vulnerable to other threats (such as overexploitation, relevant in Southeast Asia (Sodhi et
al. 2004)), as also suggested by others (Di Marco et al. 2018b). The Australasia realm also
represented a particular case, with an approximately bimodal distribution of the matrix condition
in low-risk species and a less right-skewed distribution in those classified as high-risk relative to
that shown in other realms. Interestingly, the Australasia realm showed a lower proportion of
species with a low-quality matrix (37.7%) when compared to other realms, suggesting that
species restricted to this realm have relatively lower levels of human activity within their matrix.
This is perhaps unsurprising given the fact that most recent declines of Australian terrestrial
mammals have occurred in areas with low human population pressures, where native vegetation
has not been significantly removed, particularly in the interior deserts and tropical savannas
(Woinarski et al. 2015). In fact, the decline of most Australian species has been directly related to
predation by introduced species (such as the feral cat, Felis catus, and the red fox, Vulpes vulpes)
and changes in fire regimes (May and Norton 1996, Woinarski et al. 2015), which are not
included in the human footprint.
Our models were better at correctly classifying low-risk transitions than high-risk
transitions. This suggests that the external conditions leading to a high-risk transition might be
more difficult to identify than those leading to a low-risk transition, as also indicated in previous
studies (Davidson et al. 2009, Murray et al. 2014, Bland et al. 2015b, Di Marco et al. 2015,
2018b). However, it is important to acknowledge that the exclusion of other variables associated
with pressure (such as overhunting, disease, invasive species, and climate change) and lifehistory traits (such as rarity, dispersal mode, and ranging behavior) could have increased the

52

uncertainty of our predictions, and thus influenced the ability of our models to correctly classify
high-risk transitions.
An important next step will be to create a global map by weighting the extent of the
matrix of the world’s terrestrial mammals with the human footprint in order to highlight those
matrix habitats with a high number of species and low human pressure levels, and those with a
high number of species and high human pressure levels. If species threat statuses are considered,
such an analysis could have the potential to identify where conservation actions need to be
improved. For example, in those locations where species with an increased risk of extinction
show a low-quality habitat within their matrix, a land-sparing approach could be effective as it
maximizes conservation actions on the remaining patches of suitable habitat while concentrating
agricultural production elsewhere (Green et al. 2005, Law and Wilson 2015, Phalan 2018,
Balmford et al. 2019). Alternatively, in those locations where species with an increased risk of
extinction show a high-quality habitat within their matrix, a land-sharing approach would work
better as it minimizes the impact of agricultural production by maintaining or restoring the
conservation value of the land already farmed (Perfecto and Vandermeer 2008, Law and Wilson
2015, Phalan 2018, Balmford et al. 2019).
Our results indicate that species suffering from greater pressure in their matrix require
particular conservation attention. Among these species, those with smaller ranges require careful
management of the areas surrounding their suitable habitat, especially in light of the current and
future effects of climate and land-use change (Segan et al. 2016). Our results also highlight the
potential of high-quality matrices to mitigate the negative effects of fragmentation on species
extinction risk, thus suggesting that in addition to efforts to maintain remaining suitable hábitat

53

(Ferreira et al. 2018, Maron et al. 2018, Di Marco et al. 2019), there is a need for restoration of
habitats in the matrix.

54

3. Chapter 3: Global comparison of habitat intactness models for predicting extinction
risk in terrestrial mammals
Abstract
The effects of habitat condition on biodiversity are primarily investigated using discrete (patchmatrix) habitat models that consider habitat fragments as islands embedded in an inhospitable
matrix. Recently, continuum habitat models, which focus on ecological gradients without
defining habitat or matrix, have emerged. However, no formal comparison between patchmatrix, continuum, and hybrid habitat models (which combine characteristics of both) has been
undertaken globally. Here, we compare the ability of patch-matrix, continuum, and hybrid
models of habitat intactness to explain the risk of extinction for terrestrial mammals on a global
scale. We discover that hybrid models outperform both patch-matrix and continuum models of
habitat intactness in predicting extinction risk, regardless of a species’ habitat specialization.
Moreover, we find that the magnitude of the relationship between habitat intactness and the
predicted probability of a species being threatened is stronger when employing hybrid habitat
models. Our results suggest that combining a patch-matrix designation with a continuous
representation of habitat condition can improve extinction risk analyses and provide valuable
insights for conservation efforts.

Introduction
The traditional characterization of habitat, which views habitat patches as islands
embedded in a matrix of non-habitat, has shaped how habitat condition has been assessed over
the past five decades (Ricketts 2001, Watson et al. 2005, Laurance 2008). This binary
representation of habitat, commonly referred to as the patch-matrix model, has progressively

55

been relaxed with the emergence of new approaches (e.g., Daily et al. 2001, 2003, Green et al.
2005) that recognize the potential of the matrix to provide resources, refugia, and microclimatic
conditions for sustaining biodiversity (Fischer and Lindenmayer 2007). Rather than defining
habitat as isolated patches surrounded by an inhospitable matrix, the continuum model represents
habitat as a gradient of ecological conditions with varying levels of suitability and connectivity
(Fischer and Lindenmayer 2007, Brudvig et al. 2017). However, despite the argument that
combining characteristics of patch-matrix and continuum models to represent habitat—often
referred to as hybrid models—can better explain species diversity and abundance (Price et al.
2009), formal comparisons between patch-matrix, continuum, and hybrid habitat models remain
largely unexplored (Brudvig et al. 2017).
Species vary in their sensitivity to habitat degradation, and understanding how different
representations of habitat intactness capture extinction risk is critical for effective conservation
planning (Henle et al. 2004, Ewers and Didham 2006). Habitat condition—including both its
current state and recent degradation—is increasingly recognized as a key determinant of species
persistence (e.g., Fahrig 2001, Betts et al. 2017). Yet no study has systematically compared
patch-matrix, continuum, and hybrid representations of habitat intactness in their ability to
explain extinction risk for any animal taxon at a global scale.
Here, we compare the ability of patch-matrix, continuum, and hybrid models of habitat
intactness, defined as the degree or extent to which a species’ range remains free of human
activities or significant human alterations, to predict the risk of extinction for terrestrial mammal
species globally. We focus our analysis on terrestrial mammals as they are known to be sensitive
to habitat conditions (Crooks et al. 2017, Ramírez-Delgado et al. 2022) and have often served as
a focal taxon in extinction risk analyses (Arregoitia 2016). We classified species as ‘habitat

56

generalists’ or ‘habitat specialists’ to recognize that specialized species are particularly
vulnerable to intactness given their discontinuous distributions, reduced local abundance, and
sensitivities to anthropogenic disturbances (McKinney and Lockwood 1999, Henle et al. 2004,
Charrette et al. 2006, Devictor et al. 2008). We expected patch-matrix models to outperform
continuum models in predicting extinction risk for habitat specialists because their food and
other resources are often patchy in space and time (Devictor et al. 2008, Arroyo‐Rodríguez et al.
2020). Conversely, we expected that continuum models would be better than patch-matrix
models at predicting extinction risk for habitat generalists because of their propensity to use a
broader range of habitat types (McKinney and Lockwood 1999, Henle et al. 2004, Devictor et al.
2008). Given that hybrid models standardize data inputs across the two model structures, we
expected hybrid models to provide the most accurate and reliable predictions of extinction risk
for both habitat generalists and habitat specialists.

Methods
Species-level data
We used the International Union for Conservation of Nature (IUCN) Red List (IUCN
2021) to classify species as threatened or non-threatened. Threatened species included species
categorized as ‘Vulnerable,’ ‘Endangered’ or ‘Critically Endangered,’ whereas non-threatened
species included those categorized as ‘Least Concern’ or ‘Near Threatened.’ Species categorized
as ‘Data Deficient,’ ‘Extinct in the wild,’ and ‘Extinct’ on the Red List were excluded from our
analyses. Additionally, we classified species as ‘habitat generalist’ or ‘habitat specialist’ based
on the information available from the Level 1 IUCN Red List habitat classification scheme
(IUCN 2021). Habitat generalists included species that used more than one habitat type as

57

suitable habitat, whereas habitat specialists included all species that used only one habitat type as
suitable habitat. Consistent with previous extinction risk analyses (Purvis et al. 2000, Cardillo et
al. 2008), we used species’ geographic ranges from the IUCN Red List (IUCN 2021) and
selected range polygons where the presence of each species was classified as ‘Extant’ or
‘Probably Extant’. Our final dataset retained 4,774 out of 5,787 extant terrestrial mammals,
representing 82.5% of all species in the group (IUCN 2021).

Human footprint
We used the recent release of the global human footprint (HFP) maps for the years 2000
and 2013 (Williams et al. 2020) to represent the extent and trends in habitat intactness for each
species, using a patch-matrix (Fig. 3.1b), continuum (Fig. 3.1c), and hybrid (Fig. 3.1d) model.
The HFP maps provide a 1-km spatial resolution measure of cumulative human pressure ranging
from 0 to 50 based on eight pressure layers (Venter et al. 2016): (i) built environments; (ii)
intensive agriculture; (iii) pastureland; (iv) human population density; (v) night-time lights; (vi)
roads; (vii) railways; and (viii) navigable waterways. Areas free of any human influence (e.g.,
terrestrial remaining wilderness) are represented by an HFP value of 0, those with low human
pressure (e.g., pasture lands) by HFP values of 4 or below, and those with very high-pressure
levels (e.g., densely populated semi-urban and urban environments) by HFP values above 20.

58

Figure 3.1: An example of habitat intactness models used to predict species extinction risk.
(a) This example is based on the geographic range size (polygon delimited in black) of the
western gorilla (Gorilla gorilla). (b) The patch-matrix model illustrates discrete patches of intact
habitat surrounded by an ‘inhospitable’ matrix. (c) The continuum model depicts a gradient of
habitat intactness, ranging from intact habitats (in blue) to degraded habitats (in red). (d) The
hybrid model combines characteristics of patch-matrix and continuum models, representing
discrete habitat patches and a gradient of habitat intactness within these patches, which is
influenced by the matrix condition.

59

Patch-matrix model as a predictor of extinction risk
We represented habitat intactness as patch-matrix models (Fig. 3.1b) by quantifying the
proportion of each species’ range overlapping with HFP levels lower than 3. This threshold was
used under the assumption that HFP values < 3 corresponded to intact habitat patches and HFP
values ≥ 3 to areas of unsuitable or matrix habitat. We used this HFP threshold because the
extent of low human footprint levels (i.e., HFP values < 3 out of 50) within a species’ range is
among the strongest predictors of extinction risk in terrestrial mammals (Di Marco et al. 2018b).
Furthermore, this HFP threshold is strongly associated with the greatest declines in mammalian
movements (Tucker et al. 2018) and has been used to define intact areas in previous work
(Watson et al. 2016, Jones et al. 2018, Beyer et al. 2020, Mokany et al. 2020, Ward et al. 2020,
Williams et al. 2020, O’Bryan et al. 2021). Additionally, we quantified the decrease in the
proportional overlap between each species’ range and low HFP levels from 2000 to 2013 to
account for the reduction in habitat intactness. We discarded areas where the extent of low HFP
values was lower in 2000 than in 2013, assuming no change in these areas. We only accounted
for increases in HFP values as decreases in pressure levels are likely to take time before having a
measurable effect on species conservation status, particularly for species with long generation
times (Di Marco et al. 2018b).

Continuum model as a predictor of extinction risk
We represented habitat intactness as a continuum model (Fig. 3.1c) for each species using
the metric of habitat intactness proposed by Beyer et al. (2020). This metric integrates habitat
area, habitat quality, and habitat fragmentation into a single continuous variable, all of which
influence the ability of a given habitat to support biodiversity (Ovaskainen and Hanski 2003,

60

Fischer and Lindenmayer 2007, Saura and Rubio 2010, Hanski et al. 2013, Betts et al. 2017).
This metric provides information about the contribution of each cell to the overall intactness of a
spatial feature (e.g., a species’ geographic range) based on an exponential function (𝑤𝑖 =
𝑒𝑥𝑝(−𝛾𝐻𝐹𝑃𝑖 )) that transforms the HFP to quality (Appendix B, Supplementary methods B.1).
The habitat intactness metric ranges from 0 to 1, where higher values indicate intact habitats and
lower values indicate increasing levels of habitat degradation. The contribution of each cell to
intactness (𝑄𝑖′ ) is given by:

𝑄𝑖′ =

∑𝑗∈𝑀 (𝑤𝑖 𝑤𝑗 )𝑧 𝑒𝑥𝑝(−𝛽𝑑𝑖𝑗 )
𝑖

∑𝑗∈𝑀 𝑒𝑥𝑝(−𝛽𝑑𝑖𝑗 )

(1)

𝑖

where 𝑑𝑖𝑗 is the distance between cell 𝑖 and 𝑗 (in km), 𝑀𝑖 is the set of cells falling with a radius
of cell 𝑖 that corresponds to the 99.5% quantile of the exponential distribution (here, 26.5 km), 𝑧
is an exponent that scales the product to two quantiles, 𝑤 is a measure of the quality of the cell in
the range 0–1, and 𝛽 is a parameter that determines how the combined value of pairs of cells
diminishes as a function of the distance between them. Following Beyer et al. (2020), we used a
value of 𝑧 = 0.5 to ensure that the combined weight (𝑤𝑖 𝑤𝑗 ) 𝑧 was directly proportional to the
weight 𝑤𝑖 when 𝑖 = 𝑗 (i.e., 𝑤𝑖 = (𝑤𝑖 𝑤𝑗 ) 𝑧 ), and a value of 𝛽 = 0.2, which corresponded to a 50%
penalty at a separation of 5 km and a 95% reduction of 15 km.
We measured the percentile distribution (i.e., 5th, 10th, 50th, 90th, and 95th percentiles)
of values from the habitat intactness metric within each species’ range to represent habitat
intactness as continuum models. We selected the most explanatory percentile of habitat
intactness for both habitat generalists and habitat specialists to compare the continuum model
with the patch-matrix and hybrid models in predicting extinction risk (see Statistical analysis;
61

Appendix B, Fig. B.1). Additionally, we measured the change in the selected percentiles of
habitat intactness between 2000 and 2013. As with the patch-matrix model, we discarded areas
where habitat intactness values were lower in 2000 than in 2013, assuming no change in these
areas.

Hybrid model as a predictor of extinction risk
We represented species’ habitat intactness as hybrid models by calculating the percentile
distribution of values from the habitat intactness metric within relatively intact habitat patches
(i.e., HFP values < 3) across each species’ range (Fig. 3.1d). We selected the most explanatory
percentile for both habitat generalists and specialists to compare the hybrid model’s performance
with the patch-matrix and continuum habitat models in predicting extinction risk (see Statistical
Analysis; Appendix B, Fig. B.1). Additionally, we measured the change in the selected
percentiles of the hybrid model between 2000 and 2013. As with the continuum model, we
discarded areas where habitat intactness values were lower in 2000 than in 2013, assuming no
change in these areas.
The patch-matrix, continuum, and hybrid models of habitat intactness were obtained
through geospatial analyses conducted in Python, using the ArcPy processing module from
ArcGIS Pro (ESRI 2022).

Statistical analysis
We used mixed-effects logistic regression to model extinction risk (threatened or not
threatened species) as a function of habitat intactness derived from the patch-matrix, continuum,
and hybrid models. Measures of habitat intactness were highly correlated (r > |0.7|; Appendix B,

62

Fig. B.2); thus, each regression model tested one model of habitat intactness while controlling
for the effects of other variables (Table 3.1) known to influence extinction risk (Cardillo et al.
2005, 2008, Di Marco et al. 2014b, Arregoitia 2016). Consistent with other extinction risk
analyses (Betts et al. 2017, Pillay et al. 2022), we fit separate regression models to assess
differences in habitat specialization (i.e., habitat generalists versus habitat specialists). Following
previous work (Betts et al. 2017, Ripple et al. 2017, Lucas et al. 2019), we included taxonomic
orders as random effects to broadly control for phylogenetic non-independence of the
observations. All extinction risk models were fit using the generalized linear mixed-effects
model (glmer) function from the lme4 package in R (Bates et al. 2015, R Core Team 2023).
These models included the same number of predictors (see Table 3.1) to ensure a consistent basis
for comparison.

Table 3.1: Description of variables used to predict species extinction risk. We selected the
most explanatory percentile from our continuum and hybrid models of habitat intactness to
compare how patch-matrix, continuum, and hybrid models predict species extinction risk.
(see Appendix B, Fig. B.1).
Variable

Description

Source

Patch-matrix model

Proportion of a species’ range overlapping with

Di Marco et al. 2018,

areas of low human footprint (HFP < 3) in

Williams et al. 2020

2000.
Reduction in the

Reduction in the proportional overlap between a Di Marco et al. 2018,

patch-matrix model

species’ range and areas with low human

Williams et al. 2020

footprint (HFP < 3) from 2000 to 2013.

63

Continuum model

Percentile distribution (5th, 10th, 50th, 90th,

Beyer et al. 2020,

and 95th percentiles) of habitat intactness

Williams et al. 2020

values within a species’ range in 2000.
Reduction in the

Reduction in the percentile distribution (5th,

Beyer et al. 2020,

continuum model

10th, 50th, 90th, and 95th percentiles) of habitat Williams et al. 2020
intactness values within a species’ range
between 2000 and 2013.

Hybrid model

Percentile distribution (5th, 10th, 50th, 90th,

Beyer et al. 2020,

and 95th percentiles) of habitat intactness

Williams et al. 2020

values within relatively intact habitat patches
(HFP < 3) across a species’ range in 2000.
Reduction in the

Reduction in the percentile distribution (i.e.,

Beyer et al. 2020,

hybrid model

5th, 10th, 50th, 90th, and 95th percentiles) of

Williams et al. 2020

habitat intactness values within relatively intact
habitat patches (HFP < 3) across a species'
range from 2000 to 2013.
Order

A species’ taxonomic order.

IUCN 2021

Range size

Area in square kilometers of a species’

IUCN 2021

distribution.
Gestation length

A proxy for a species’ reproductive output.

Soria et al. 2021

Weaning age

A proxy for a species’ reproductive onset.

Soria et al. 2021

64

Assessing model performance
We used the area under the receiver operating characteristic curve (AUC) metric to assess
the predictive performance of our extinction risk models. This metric reflects true positive rates
as a function of false positive rates for a binary classifier (Swets 1988, 2014). We used a
spatially blocked cross-validation scheme to calculate AUC values for models of extinction risk
for both habitat generalists and habitat specialists. We used this method because it accounts for
entire regions as hold-out datasets, reducing the positive effects of dependency on model
performance metrics (Bahn and McGill 2013, Cazalis et al. 2022). Additionally, we utilized a
ten-fold cross-validation scheme for comparison. For the spatial blocking approach, the
probability of a species being threatened was predicted for each of six biogeographic realms
(Afrotropic, Australasia, Indomalayan, Nearctic, Neotropic, and Palearctic), with the model fit
from all other realms. We only retained those species with the majority (>50%) of their
distributions within a realm based on the biogeographic classification of the world proposed by
(Dinerstein et al. 2017). We pooled species from the Oceania realm into the Australasia realm
because the Oceania realm contained only eight species in our dataset. Following Swets (1988),
we considered models with AUC values of 0.5–0.7 as being ‘low accuracy,’ those ranging
between 0.7 and 0.9 as ‘useful applications’, and those with values greater than 0.9 as being
‘high accuracy.’

Sensitivity analysis
Species designated as threatened due to small geographic range (i.e., those assessed
solely under IUCN Red List Criterion B) can introduce circularity in extinction risk analyses
(Cardillo et al. 2005). Thus, we conducted a sensitivity analysis in which we excluded only

65

species classified as threatened under Criterion B alone. This resulted in the exclusion of 242
habitat generalists and 281 habitat specialists. We then compared results from this reduced
dataset to those obtained using the full dataset, which included 744 habitat generalists and 486
habitat specialists classified as threatened, and 2,667 habitat generalists and 877 habitat
specialists classified as non-threatened.

Results
Overall, our extinction risk models performed well for both habitat generalists and habitat
specialists (Appendix B, Fig. B.3). Under a spatially blocked cross-validation scheme, models
that included hybrid representations of habitat intactness—along with their temporal decline—
demonstrated the highest predictive performance, with an average AUC of 0.83 (95% CI: 0.78–
0.88) for habitat generalists and 0.82 (95% CI: 0.75–0.89) for habitat specialists. Ten-fold crossvalidation yielded similar results, with hybrid-based extinction risk models achieving an average
AUC of 0.85 (95% CI: 0.82–0.87) for habitat generalists and 0.83 (95% CI: 0.81–0.86) for
habitat specialists.
Our results were not substantially altered after excluding threatened species assessed
solely under Criterion B (Figs. B4, B5), indicating the robustness of extinction risk models to
this sensitivity analysis. However, for habitat generalists, models using patch-matrix and
continuum representations of habitat intactness, as well as their temporal decline, showed
increased AUC values with non-overlapping confidence intervals. This suggests that the
exclusion of threatened species assessed solely under Criterion B may partially influence model
performance when habitat condition is represented using patch-matrix or continuum approaches.

66

In contrast, extinction risk models based on hybrid representations remained consistent across all
analyses, highlighting their value for extinction risk assessments.
We found a statistically significant negative association between the risk of extinction
and the patch-matrix, continuum, and hybrid models of habitat intactness for both habitat
generalists (Fig. 3.2a) and habitat specialists (Fig. 3.2b). This association was stronger when
habitat intactness was represented by hybrid models, followed by continuum models.
Additionally, we found that, regardless of the habitat model, the effect of intactness on the risk of
extinction was stronger for habitat specialists. These results remained consistent when we
excluded threatened species that met IUCN Red List criterion B alone (Appendix B, Fig. B.6),
demonstrating that a simple circular relationship between predictor and response variable did not
influence our findings.

67

Figure 3.2: Effects of three habitat intactness models (a, b) and their temporal reduction (c,
d) on extinction risk for terrestrial mammals classified as habitat generalists and habitat
specialists. Each variable is colour-coded according to the habitat model (patch-matrix,
continuum, or hybrid model) it represents. Effect sizes were estimated while holding other
variables known to influence extinction risk (see Table 3.1) at their mean. All variables were
standardized (z-transformed) before analysis to ensure comparable results. Error bars represent
95% confidence intervals. The statistical significance of the coefficients is indicated by the
absence of overlap between their confidence intervals and zero (dashed green line). Acronyms:
patch-matrix model (PMM), continuum model (CM), hybrid model (HM), reduction in the

68

patch-matrix model (RPMM), reduction in the continuum model (RCM), and reduction in the
hybrid model (RHM).

Our results reveal a statistically significant association between extinction risk and the
reduction in habitat intactness over time, exclusively when using continuum models for habitat
generalists (Fig. 3.2c) and patch-matrix models for habitat specialists (Fig. 3.2d). This
association lost significance for habitat generalists but remained statistically significant for
habitat specialists after excluding threatened species meeting IUCN Red List criterion B alone
(Appendix B, Fig. B.6). These findings support our initial expectations, indicating that habitat
specialists are particularly at risk when the size of intact habitat patches decreases over time.
Notably, our results show that the decline in the predicted probability of being threatened
was steepest under hybrid models (Fig. 3.3a,b). This suggests that hybrid models of habitat
intactness more accurately reflect how small reductions in intactness can lead to rapid increases
in extinction risk. This result was consistent for both habitat generalists (Fig. 3.3a) and
specialists (Fig. 3.3b), although specialists showed a comparatively greater probability of being
threatened when their ranges contained patches with the lowest intactness.

69

Figure 3.3: Predicted probabilities of a species being threatened as a function of three
habitat intactness models for both habitat generalists (a) and habitat specialists (b).
Predicted probabilities were estimated while holding other variables known to influence
extinction risk (see Table 3.1) at their mean. Variables representing patch-matrix, continuum,
and hybrid models were back-scaled to improve results interpretability. Shaded portions
represent 95% confidence intervals.

Discussion
Habitat models are fundamental for understanding the impact of human activities on
biodiversity, and for informing conservation policies and management strategies (Fischer and
Lindenmayer 2006, Brudvig et al. 2017). We found that, regardless of a species’ habitat
specialization, hybrid models most accurately predicted extinction risk for terrestrial mammals

70

when comparing patch-matrix, continuum, and hybrid representations of habitat intactness on a
global scale. These results underscore the untapped potential of conceptual habitat models—
particularly hybrid approaches—for improving extinction risk assessments at broad spatial
scales. Our findings are consistent with a previous study showing broad support for hybrid
approaches (Brudvig et al. 2017), which reinforces the suitability of hybrid models for
conceptualizing and interpreting the effects of habitat intactness on biodiversity.
Although these frameworks have often been described as conceptual landscape models
(e.g., Price et al. 2009, Brudvig et al. 2017), we apply them in a species-specific context and
refer to them as models of habitat intactness. This distinction reflects our focus on quantifying
the condition of habitat within each species’ range, rather than characterizing landscape structure
more broadly. Our interpretation aligns with Fischer and Lindenmayer (2006, 2007), who
conceptualized these models as tools to understand how species use habitat across modified
environments. Given that species differ widely in their sensitivity to habitat degradation (Henle
et al. 2004, Ewers and Didham 2006), understanding how different representations of intactness
influence extinction risk is critical for anticipating biodiversity loss.
Contrary to our expectations, we found that continuum models were better than patchmatrix models at predicting extinction risk for habitat specialists. One plausible explanation for
this finding is that some human-modified areas can provide suitable habitats and act as a buffer
mechanism against extinction not only for habitat generalists but also for habitat specialists,
particularly when these areas exhibit structural similarities to their preferred habitat. This notion
is supported by previous studies demonstrating the conservation potential of well-managed
production landscapes, where many habitat specialists can persist (e.g., Daily et al. 2001, 2003,
Benton et al. 2003, Hole et al. 2005).

71

In our study, using the HFP and its associated metrics of habitat intactness provided a
unique opportunity to test the relationship between conceptual habitat models and extinction risk.
However, the exclusion of variables associated with pressure not captured by the HFP—such as
biological invasions, overexploitation, and climate change—could have limited the predictive
success of our models. Considering the inherent challenge of accurately mapping these
influential variables (Belote et al. 2020), we acknowledge this limitation.
Our results suggest that the choice of habitat model can influence the empirical
relationship between habitat intactness and extinction risk. For example, we found that the
probability of a species being threatened was notably greater under conditions of lower habitat
intactness when represented by the hybrid model rather than the patch-matrix model, regardless
of a species’ habitat specialization. This finding suggests that conservation priorities should be
focused not only on minimizing human disturbances within relatively intact habitat patches but
also on enhancing the structural similarity between intact habitat and the surrounding matrix.
This approach to managing the matrix habitats aligns with a recent study showing that high-value
biodiversity habitats in both intact and highly modified regions require focused conservation
commitments and actions to halt ongoing extinctions (Mokany et al. 2020).
Our results indicate that habitat specialists were especially vulnerable to the reduction in
intactness when modeled using the patch-matrix approach, even after excluding threatened
species assessed solely under criterion B. This suggests that including the temporal reduction of
habitat intactness as a patch-matrix model has the potential to enhance model performance when
predicting extinction risk, particularly for habitat specialists.
Our findings reveal a statistically significant negative relationship between habitat
intactness and extinction risk across our patch-matrix, continuum, and hybrid models, with non-

72

threatened species likely to become threatened if current rates of the HFP persist into the future
(Williams et al. 2020, Mu et al. 2022). These results point toward the possibility of future work
to use our habitat models to predict extinction risk under alternative global scenarios of
socioeconomic growth (Watson and Venter 2019). However, undertaking such an endeavor
would require generating future projections of the HFP and predicting shifts in species
distribution in response to climate and land-use changes (Visconti et al. 2016, Beyer and Manica
2020).
Regardless of a species’ habitat specialization, we found that terrestrial mammals are
more likely to be threatened if their ranges contain habitat patches with decreasing structural
intactness. These results may be attributed to the combined impact of habitat loss and
degradation, which indirectly give rise to additional threats such as hunting, disease spread,
selective logging, wildfires, and invasive species (Barlow et al. 2016, Betts et al. 2017). This
indicates that the ability of intact habitat patches to support biodiversity gradually decreases with
the increasing intensity of human pressure around them, highlighting that in addition to efforts to
minimize human disturbance, there is a need for targeted management strategies in the
surrounding matrix.

73

4. Chapter 4: Mapping the conservation potential of the matrix for terrestrial
mammals
Abstract
Conservation strategies have long prioritized primary habitats, often overlooking the matrix—the
surrounding areas of lower habitat suitability—where many species may still persist. This
oversight can compromise species persistence and ecological connectivity, potentially
undermining long-term biodiversity conservation. Here, we present the first global assessment of
species richness within the matrix for terrestrial mammals. Using high-resolution habitat maps
for 5,439 species, we identify species-rich matrix areas, where species richness is greater than in
suitable habitats, and delineate matrix-based richness hotspots—defined as the top 5% of grid
cells with the highest richness values within these areas—for all mammals, declining mammals,
and threatened mammals. We found that species-rich matrix areas cover 12% of Earth’s
terrestrial surface, yet matrix hotspots occupy just 1%. Importantly, over half of terrestrial
mammals potentially occur in these matrix hotspots. Notably, hotspots defined solely by
threatened mammals contain nearly as many species as all hotspot types combined. Matrix areas
identified as having high conservation potential—based on overlapping richness hotspots—are
primarily concentrated in tropical strongholds such as the Amazon Basin, Colombian Tropical
Andes, Brazilian Atlantic Forest, and Albertine Rift. Nearly half of these areas face intense
human pressure and remain outside protected or conserved area networks, suggesting they
represent important opportunities for biodiversity conservation. Our results highlight the critical
importance of integrating the matrix into global biodiversity conservation strategies.

74

Introduction
Industrial-scale human activities have transformed between 75% and 95% of Earth’s icefree land surface (Venter et al. 2016, Ellis et al. 2021, Ellis 2021), leading to increasingly
fragmented landscapes where natural ecosystems are embedded within human-modified
environments (Perfecto and Vandermeer 2008). As a result, many species are now confined to—
or must frequently move through—landscapes dominated by agriculture, urbanization, and other
forms of human land use (Tucker et al. 2018, Kremen and Merenlender 2018, Galán-Acedo et al.
2019, Torres-Romero et al. 2020, Britnell et al. 2023, Zhang et al. 2024). These large-scale
transformations have been widely recognized as the primary drivers of an accelerating
biodiversity crisis (Newbold et al. 2015, Maxwell et al. 2016, Jaureguiberry et al. 2022), with
extinction rates estimated to be tens to thousands of times higher than historical background
levels (Barnosky et al. 2011, Pimm et al. 2014, De Vos et al. 2015, Ceballos et al. 2015). As
habitat loss and fragmentation continue to intensify, the capacity of species to survive beyond
primary habitats—natural and semi-natural areas characterized by minimal human disturbance
and high ecological integrity—is emerging as a key determinant of biodiversity persistence in the
Anthropocene (Gardner et al. 2009, Prevedello and Vieira 2010, Watling et al. 2011).
The areas surrounding primary habitats—often referred to as the matrix—have
traditionally been treated as unsuitable or ecologically irrelevant (MacArthur and Wilson 1967,
Haila 2002). However, a growing body of research challenges this view, demonstrating that the
matrix can either buffer or exacerbate the impacts of habitat loss and fragmentation, depending
on their quality, structure, and spatial configuration (Ricketts 2001, Prugh et al. 2008, Prevedello
and Vieira 2010, Watling et al. 2011, Driscoll et al. 2013, Ramírez-Delgado et al. 2022). Highcontrast matrices, such as intensive agriculture or urban development, can increase isolation and

75

mortality risk (Battin 2004), whereas low-contrast matrices, like agroforestry systems or
secondary forests, can facilitate movement, reduce edge effects, and supplement habitat
resources (Daily et al. 2003, Perfecto and Vandermeer 2008). Despite this growing recognition,
global conservation strategies remain largely focused on primary habitats (Margules and Pressey
2000, Brooks et al. 2006, Gibson et al. 2011, Martin et al. 2012, Watson et al. 2018a, Di Marco
et al. 2019), often overlooking the ecological value and conservation potential of the surrounding
matrix (Prugh et al. 2008, Franklin and Lindenmayer 2009, Prevedello and Vieira 2010, Kremen
and Merenlender 2018, Arroyo‐Rodríguez et al. 2020). This narrow focus limits our ability to
identify conservation opportunities, particularly in human-modified landscapes where many
species now persist (Tucker et al. 2018, Kremen and Merenlender 2018, Galán-Acedo et al.
2019, Torres-Romero et al. 2020, Britnell et al. 2023, Zhang et al. 2024).
Here, we present a global assessment of terrestrial mammal species richness within the
matrix, defined as areas within a species' geographic range but outside its suitable habitat. We
focused on terrestrial mammals because of the availability of consistent, species-level habitat
data (Lumbierres et al. 2022b). Using Area of Habitat (AOH) maps at 100-m spatial resolution
for 5,439 species (Lumbierres et al. 2022b) and geographic range maps from the International
Union for Conservation of Nature (IUCN 2024), we mapped species richness across suitable
habitats and the matrix. AOH maps delineate suitable habitat by combining species-specific
ecological requirements, such as elevation range, land-cover type, and known habitat
associations (Rondinini et al. 2011, Brooks et al. 2019, Lumbierres et al. 2022b). However, AOH
maps assume that species distributions strictly align with areas identified as suitable, potentially
overlooking occurrences in suboptimal or transitional habitats, microhabitat variations, or
species’ ecological plasticity (Rondinini et al. 2011, Lumbierres et al. 2022b). This limitation

76

may underestimate the actual extent of species’ habitat use and their ability to persist in or utilize
areas classified as unsuitable or marginal. We therefore identified species-rich matrix areas,
defined here as locations with disproportionately greater richness of terrestrial mammals
compared to suitable habitats. Based on IUCN Red List assessments (IUCN 2024), we delineated
richness hotspots within these species-rich matrix areas for three groups: all mammals, mammals
with declining population trends, and mammals with threatened status. By overlapping these
hotspots, we classified matrix areas as having high, moderate, or low conservation potential and
further assessed them in terms of human pressure (Gassert et al. 2023) and conservation
coverage (UNEP-WCMC and IUCN 2023). Through this mapping, we identified matrix areas as
critical yet underrecognized components of global conservation strategies, where habitat
restoration or requalification efforts could deliver significant benefits to terrestrial mammal
conservation.

Methods
Species area of habitat
The extent of suitable habitat and matrix areas for each terrestrial mammal was derived
from AOH maps developed by Lumbierres et al. (2022b). These maps, created at a 100-m spatial
resolution for the year 2015, delineate suitable habitat areas within each species’ geographical
range by integrating species-specific habitat preferences and elevational limits from IUCN Red
List data, as well as high-resolution land cover data (Lumbierres et al. 2022a, b). Validation
against known occurrence points was possible for only a subset of terrestrial mammal species
(8% of those mapped), owing to the limited availability of high-accuracy point locality data
(Lumbierres et al. 2022b). Within this subset, ~95% of AOH maps performed better than random

77

within species’ ranges, indicating broad consistency between mapped suitable habitat and
observed occurrences. While this performance threshold represents a minimum benchmark rather
than a high predictive accuracy metric, the maps provide a consistent and spatially explicit basis
for global-scale biodiversity assessments, making them well suited for the aims of this study.
Given that AOH maps only capture suitable habitat for each species, the extent of the
matrix was defined using species range maps from the IUCN Red List (IUCN 2024). Following
established criteria (IUCN 2020, Lumbierres et al. 2022b), we selected polygons classified as
‘Extant’ or ‘Probably Extant’ and labeled as ‘resident,’ ‘Reintroduced (resident),’ or ‘Assisted
Colonisation (resident)’. Pixels outside suitable habitat (as defined by AOH maps) but within the
selected range polygons were classified as unsuitable or matrix areas.
To ensure consistency between the AOH maps and the range data, we excluded 41 AOH
maps from the initial dataset of 5,481. These exclusions resulted from taxonomic updates, spatial
mismatches between AOH maps and range maps, changes in species range classifications, and
cases where AOH maps lacked corresponding range data. As a result, our final dataset included
5,439 species, representing about 94% of the world’s extant terrestrial mammals (IUCN 2024).
From this final dataset, we then delineated suitable habitat and matrix areas within each
species’ range. To minimize the influence of potential inconsistencies in the AOH data, small
contiguous groups of pixels (<4 adjacent pixels of the same class) that may have resulted from
inconsistencies in the underlying datasets of the AOH maps were removed. These small groups
were replaced with the value of the largest and nearest contiguous group, determined using the
eight-neighbor rule. In line with previous work (Crooks et al. 2011, 2017, Ramírez-Delgado et
al. 2022), this process minimized the influence of isolated pixel groups and enhanced the
computational efficiency of our analyses.

78

Species richness maps
We generated species richness maps for matrix areas by aggregating pixels classified as
‘unsuitable’ for each species of terrestrial mammal. Richness maps were created for three
groups: all mammals (Appendix C, Fig. C.1a), mammals with declining population trends
(Appendix C, Fig. C.1b), and mammals with threatened status (Appendix C, Fig. C.1c).
Conservation status and population trend data were sourced from the IUCN Red List (IUCN
2024). Species listed as ‘Vulnerable,’ ‘Endangered,’ or ‘Critically Endangered’ were categorized
as threatened, while those with a population trend labeled ‘Decreasing’ were categorized as
declining.
In addition to the matrix richness maps, we generated a suitable habitat richness map for
the all-mammal group (Appendix C, Fig. C.4). This map served as a broad reference for
identifying species-rich matrix areas, as detailed in the following section.

Species-rich matrix areas
Using the all-mammal group, we calculated pixel-level species richness separately for
matrix areas (Appendix C, Fig. C.1a) and suitable habitats (Appendix C, Fig. C.2), generating
two global richness maps. For each pixel, we computed a richness ratio by dividing the matrix
richness value by the suitable habitat richness value at the same spatial location. Pixels with
richness ratios greater than 1, indicating greater richness in the matrix relative to suitable
habitats, were classified as species-rich matrix areas. These areas represented locations in the
matrix with disproportionately greater species richness compared to suitable habitats. We then
extracted species richness values within these species-rich matrix areas from matrix-based

79

richness maps for each species group (i.e., all mammals, mammals with declining population
trends, and mammals with threatened status).

Hotspots of species richness
Following previous studies (e.g., Orme et al. 2005, Ceballos and Ehrlich 2006, Jenkins et
al. 2013), we identified hotspots of species richness within species-rich matrix areas by selecting
the top 5% of pixels with the highest richness values for each species group. Pixels were
classified as hotspots only if they formed contiguous clusters with a minimum area of 10 km2, a
conservative threshold broadly aligned with generalized minimum-area guidelines suggested to
support mammal assemblages in global biodiversity assessments (Verboom et al. 2014). We used
an iterative threshold optimization process to systematically refine the richness threshold,
ensuring that the final hotspot coverage closely approximated 5% for each species group while
meeting clustering requirements. To test the sensitivity of our results, we also evaluated
alternative thresholds using the top 2.5% and top 10% of pixels.

Levels of conservation potential within matrix areas
Levels of conservation potential within matrix areas were identified based on spatial
overlap among hotspot types. Areas with high conservation potential corresponded to locations
where all three hotspot types (all mammals, mammals with declining population trends, and
mammals with threatened status) overlapped, areas with moderate conservation potential where
any two overlap, and areas with low conservation potential where only a single hotspot type was
present with no overlap.

80

Species-level hotspot representation
To assess species representation across hotspot types, we identified species whose matrix
area spatially overlapped with any of the three hotspot types. We considered a species to occur in
a hotspot if its matrix (defined as areas within the species’ range but outside its suitable habitat)
intersected or overlapped hotspot boundaries. For each species, we determined whether its matrix
area intersected one, two, or all three hotspot types, based on spatial overlays. Each species was
assigned to a single overlap category corresponding to the unique combination of hotspot types
its matrix intersected. These species-level assignments were then used to quantify representation
across individual hotspot types and their combinations.

Human pressure in matrix hotspots
We used the human footprint (HFP) dataset (Gassert et al. 2023) to quantify the influence
of human pressures within matrix areas identified as having conservation potential. Built on
previous HFP mapping efforts (Venter et al. 2016, Williams et al. 2020), this dataset provides a
standardized measure of human pressure on terrestrial environments by integrating data on eight
anthropogenic pressures: (i) built environments, (ii) cropland, (iii) pastureland, (iv) human
population density, (v) night-time lights, (vi) roads, (vii) railways, and (viii) navigable
waterways. Developed at a 100-m spatial resolution for the years 2015–2019 and 2020, the maps
in this dataset assign HFP scores ranging from 0 (minimal human influence) to 50 (maximum
human influence).
For our analysis, we used the 2015 HFP map to ensure temporal consistency with the
AOH maps. Following previous studies (Di Marco et al. 2018b, Mokany et al. 2020, RamírezDelgado et al. 2022, Gassert et al. 2023), HFP scores from this map were classified into three

81

categories to represent varying levels of human impact: intact (HFP ≤ 1), low disturbance (1 <
HFP < 3), and highly modified (HFP ≥ 3). We specifically applied a threshold of HFP ≥ 3 to
identify highly modified areas, as this threshold has been shown to be the strongest predictor of
extinction risk transitions in terrestrial mammals (Di Marco et al. 2018b). To quantify human
influence within matrix areas with conservation potential, the HFP categories were spatially
overlaid on these zones.

Conservation coverage of matrix hotspots
We used spatial data from the World Database on Protected Areas (WDPA) and the
World Database on Other Effective area-based Conservation Measures (WD-OECM) (UNEPWCMC and IUCN 2023) to assess conservation coverage within matrix hotspots.
Following established guidelines (Butchart et al. 2015, Runge et al. 2015, Protected
Planet 2025), we excluded protected areas (PAs) and other effective area-based conservation
measures (OECMs) entries that did not meet international definitions of PAs or OECMs.
Specifically, we removed sites with unknown status, proposed designations, marine-only sites,
the marine portions of coastal sites, and sites lacking defined boundaries. Sites represented only
by point data were converted into circular polygons based on reported area size. Each dataset
(PAs and OECMs) was processed independently and then combined to create a third, integrated
layer representing all conserved areas (PAs + OECMs).
All three datasets were rasterized at 100-m spatial resolution to match the resolution of
the AOH and HFP datasets. We then generated binary raster layers classifying land as either
protected, OECM-designated, or conserved (PAs + OECMs), and overlaid these layers with the
matrix conservation potential map stratified by human footprint category (intact, low-

82

disturbance, and highly modified). For each level of conservation potential—defined by spatial
overlap among hotspot types, where high corresponds to overlap of all three types, moderate to
overlap of any two types, and low to only one—we calculated the proportion of area under
protection, OECM designation, or both combined, overall and within each footprint category.
All spatial data were initially processed in the WGS 1984 coordinate system and
reprojected to the Mollweide equal-area projection for consistent area calculations. Geospatial
analyses were performed using ArcPy (ESRI 2024) and the wdpar package in R (Hanson 2022,
R Core Team 2023).

Results
Global patterns of terrestrial mammal concentration in the matrix
We found that species-rich matrix areas covered approximately 12% of the global
terrestrial area (excluding Antarctica) and were predominantly located in tropical and subtropical
regions (Fig. 4.1a–c). This pattern was evident across all mammals (Fig. 4.1a), mammals with
declining population trends (Fig. 4.1b), and those with threatened status (Fig. 4.1c), with the
highest concentrations occurring in parts of Central and South America, West and Central
Africa, and Southeast Asia.

83

Figure 4.1: Global distribution of species richness within species-rich matrix areas for
terrestrial mammals. Species richness is shown for all species (a), species with declining
population trends (b), and species with threatened status (c).

84

Hotspots of species richness, defined as the top 5% of grid cells with the highest richness
values within species-rich matrix areas, revealed even more pronounced patterns of species
concentration (Fig. 4.2a–c). For all mammals, hotspots were primarily located in tropical regions
of South America and Africa, with particularly high concentrations in the Tropical Andes, the
Colombian Caribbean, and the Llanos (Fig. 4.2a). In contrast, hotspots for mammals with
declining population trends were more scattered and broadly distributed across South America,
sub-Saharan Africa, and Southeast Asia, extending into the Eastern Himalayas and Indo-Burma
(Fig. 4.2b). Hotspots for threatened mammals overlapped in part with those for mammals with
declining population trends and showed a wider distribution across tropical and subtropical
regions, including parts of Central America, Madagascar, and eastern Australia (Fig. 4.2c).

85

Figure 4.2: Hotspots of species richness within species-rich matrix areas for terrestrial
mammals. Hotspots are shown for all mammals (a), mammals with declining population trends
(b), and mammals with threatened status (c). Dark red represents the top 2.5% of grid cells with
the highest richness values for each species group, red (including dark red) represents the top

86

5%, and pink (including both red and dark red) represents the top 10%. All hotspots were
aggregated into contiguous clusters with a minimum area of 10 km².

A total of 3,419 species—approximately 59% of all extant terrestrial mammals—had
matrix areas that overlap with at least one of the three hotspot types (Table 4.1). Collectively, the
three hotspot types spanned 1.65 million km², representing 10% of the global extent of speciesrich matrix areas and just 1% of the Earth’s total terrestrial surface (excluding Antarctica).
Notably, hotspots based solely on threatened species captured nearly as many species as the
combined total across all three hotspot types, differing by only 3%. This pattern held true across
alternative hotspot thresholds (Appendix C, Table C.1), suggesting that threatened species
hotspots could serve as effective surrogates for broader conservation efforts within species-rich
matrix areas.

Table 4.1: Geographic coverage and species representation of hotspot types. The table
shows the total area of each hotspot type and the number of species whose matrix area
overlapped with each hotspot type. Numbers in parentheses indicate the percentage of each
hotspot type relative to the global extent of species-rich matrix areas (15,922,965 km²), or the
percentage of overlapping species relative to the total number of extant terrestrial mammals
(5,808 species worldwide). Slight differences in area across hotspot types reflect clustering
requirements and the threshold optimization process (see Methods).
Hotspot type

Area (km²)

Number of species

All species

799,277 (5.0)

1,918 (33.0)

Declining trend

818,964 (5.1)

2,836 (48.8)

87

Threatened status

702,029 (4.4)

3,255 (56.0)

Total in all hotspots

1,655,469 (10.4)

3,419 (58.9)

Matrix areas with conservation potential
The spatial overlap among the three hotspot types was low overall (Fig. 4.3a,b). Areas
with high conservation potential—where all three hotspot types overlapped—accounted for only
7% of the total hotspot extent. Moderate conservation potential areas, defined by overlap
between two hotspot types, accounted for 26%, while low conservation potential areas, identified
by only a single hotspot type, accounted for the majority (67.1%).

88

Figure 4.3: Overlap of hotspot types and conservation potential in matrix areas. The map
shows the global distribution of species-rich matrix areas, categorized into levels of conservation
potential based on spatial overlap among three hotspot types: all mammals, mammals with
declining population trends, and mammals with threatened status (a). Venn diagrams show the
spatial overlap among hotspot types (b) and the number of species whose matrix habitat
overlapped one, two, or all three types (c), respectively. Each species was counted only once,
based on the unique combination of hotspot types overlapping its matrix habitat (e.g., one, two,
89

or all three types), and was assigned to the corresponding category. Areas of high conservation
potential corresponded to overlaps among all three hotspot types, moderate conservation
potential to overlaps between any two types, and low conservation potential to areas identified
by a single hotspot type.

Areas with high conservation potential harboured 53% of all mammals whose matrix
habitats overlapped hotspot areas, while areas with moderate and low conservation potential
accounted for 29% and 18%, respectively (Fig. 4.3c). Notably, 18% of species with matrix
habitat in hotspots occurred exclusively within a single hotspot type, with the majority (14.8%)
found only in hotspots for species with threatened status. A total of 932 species occurred in both
hotspots for species with declining populations and those for threatened species, suggesting that
these categories frequently overlap and that many threatened species are already experiencing
ongoing population declines. These patterns were consistent across different hotspot thresholds
(Appendix C, Fig. C.3), further highlighting the strong association between population declines
and threat status.
Our map of conservation potential in matrix areas showed that areas of high conservation
potential were primarily concentrated within a few tropical strongholds, including parts of the
Amazon Basin, the Colombian Tropical Andes, the Brazilian Atlantic Forest, and the Albertine
Rift (Fig. 4.3a). Areas of moderate conservation potential were more widely distributed but still
predominantly located in tropical regions across South America, Central Africa, and Southeast
Asia. In contrast, areas of low conservation potential were the most extensive, spanning Central
and South America, sub-Saharan Africa, the Eastern Himalayas, Indo-Burma, and eastern
Australia.

90

Human footprint and conservation coverage in matrix areas with conservation potential
When we mapped the conservation potential of matrix areas against human footprint
categories (Gassert et al. 2023)—‘intact’, ‘low-disturbance’, and ‘highly modified’—we found
that these areas were predominantly highly modified, although notable subregional variations
existed (Fig. 4.4a–d). In parts of Brazil’s Amazon (Fig. 4.4a) and Atlantic Forest (Fig. 4.4b),
matrix areas with high, moderate, and low conservation potential were widespread and
predominantly highly modified. However, scattered remnants of intact and low-disturbance
matrix persisted, creating a mosaic of conditions. In the Western Congolian forest–savanna
region (Fig. 4.4c), matrix areas with low conservation potential remained largely intact, though
interspersed with more disturbed areas. In Sumatra’s forests (Fig. 4.4d), matrix areas with
moderate conservation potential dominated and were almost entirely highly modified, with only
small patches of intact and low-disturbance matrix remaining.

91

Figure 4.4: Human footprint within matrix areas with conservation potential. The map
shows the global distribution of matrix areas with high, moderate, and low conservation
potential, categorized by human pressure levels: intact, low-disturbance, and highly modified.
Insets show examples from Brazil’s Amazon (a) and Atlantic Forest (b), the Western Congolian
92

forest-savanna mosaic (c), and Sumatra’s forests (d). Human pressure levels were derived from
the 2015 global HFP map developed by Gassert et al. (2023) at a 100-m spatial resolution. For
details on thresholds used to define each human pressure level, see Methods.

Across all levels of conservation potential, matrix areas remained predominantly
unprotected and highly modified (Appendix C, Fig. C.4a). PAs covered only a fraction of matrix
areas with high conservation potential, with 49% of this land unprotected and highly modified,
and just 7% both protected and intact. Matrix areas with moderate and low conservation potential
showed similar patterns, with unprotected, highly modified conditions accounting for 58% and
54%, respectively. OECMs offered only limited additional coverage (Appendix C, Fig. C.4b),
and the matrix areas they covered were predominantly highly modified.
Combining PAs and OECMs—hereafter referred to as conserved areas—slightly
improved overall conservation coverage but did not substantially alter the broader pattern
(Appendix C, Fig. C.4c). Non-conserved and highly modified matrix areas remained the most
common condition, accounting for 48% of matrix areas with high conservation potential, 57%
with moderate conservation potential, and 52% with low conservation potential. Even where
matrix areas were conserved, they were predominantly within highly modified landscapes.
Among matrix areas with high conservation potential, 8% were both conserved and highly
modified, while only 7% were both conserved and intact. These trends were consistent across
other levels of conservation potential, with the majority of conserved matrix areas situated in
areas of high human pressure.

93

Discussion
Our study presents a global framework for identifying conservation opportunities within
the matrix, offering new insights into global patterns of the species richness of terrestrial
mammals within these areas. We find that species-rich matrix areas cover approximately 12% of
Earth’s land surface (excluding Antarctica), yet richness hotspots—largely concentrated in
tropical and subtropical regions—cover just around 1%. Importantly, more than half of all
terrestrial mammal species may occur within at least one of these hotspot types, with hotspots
identified solely by threatened species containing nearly as many species as all hotspot types
combined. Matrix areas with high conservation potential, defined by the overlap of all three
hotspot types, are limited in extent but potentially support disproportionately high species
richness, particularly in regions such as the Amazon Basin, the Colombian Tropical Andes, the
Brazilian Atlantic Forest, and the Albertine Rift. However, nearly half of these matrix areas
remain highly modified and poorly represented within existing protected or conserved area
networks, underscoring the urgent need to explicitly integrate matrix areas into global
biodiversity conservation strategies.
While our analysis partially covers regions traditionally recognized as biodiversity
hotspots (Myers et al. 2000, Ceballos and Ehrlich 2006, Jenkins et al. 2013), it offers a novel
perspective by focusing on species richness within the matrix. Rather than relying solely on
species range maps, we used AOH data (Lumbierres et al. 2022b) to differentiate between
suitable habitat and matrix areas within species’ ranges, an approach known to provide more
refined and ecologically relevant representations than estimates derived from geographic ranges
alone, particularly in extensively modified regions (Rondinini et al. 2011). This distinction
allowed us to identify conservation opportunities that might otherwise be overlooked if

94

biodiversity value is assumed to be confined solely to suitable habitats. By integrating HFP
(Gassert et al. 2023) and conservation coverage data (UNEP-WCMC and IUCN 2023), our
analysis offers a more nuanced picture of where biodiversity is not only concentrated but also
under pressure and poorly represented within existing formally recognized conservation
networks. Together, these elements position our framework as a complement to existing hotspot
analyses, extending their value by identifying overlooked components of the landscape that are
critical for biodiversity representation and long-term resilience (Brooks et al. 2006, Kremen et al.
2008, Mokany et al. 2020).
Our study advances global conservation assessments by applying a finer spatial grain
(100 meters) for AOH maps (Lumbierres et al. 2022b). This resolution represents a substantial
increase in spatial detail compared to influential global studies (Myers et al. 2000, Ceballos and
Ehrlich 2006, Jenkins et al. 2013), which typically relied on coarser spatial data, often ranging
from 10 to 100 km in grain size. We recognize, however, that finer spatial resolution alone does
not inherently guarantee greater ecological accuracy or realistic habitat suitability. Indeed, AOH
maps are simplified representations of habitat suitability, primarily suited for global biodiversity
assessments rather than detailed, species-specific management planning. Nevertheless, the highresolution analysis presented here provides enhanced localization of global species richness
patterns, revealing spatial details likely missed by coarser analyses. While this resolution is
particularly valuable for identifying general patterns in fragmented or rapidly changing
landscapes—where conservation strategies often depend on spatially explicit information
(Rouget 2003, Kremen et al. 2008, Tanhuanpää et al. 2023, Pla et al. 2024, Carruthers-Jones et
al. 2025)—we emphasize caution in interpreting these results as directly indicative of fine-scale
ecological processes or habitat suitability at local scales.

95

Empirical validations highlight that, although broadly accurate, AOH maps do not
capture all species occurrences. For terrestrial mammals, validation against known occurrence
points was possible for only a subset of species (8% of those mapped), owing to the limited
availability of high-accuracy point locality data (Lumbierres et al. 2022b). Within this subset,
~95% of maps performed better than random within species’ ranges—a minimum performance
benchmark rather than a high predictive accuracy metric. Consistent with these limitations, our
analyses reveal that matrix areas can contain high mapped species richness, reinforcing their
ecological relevance and underscoring the value of incorporating these areas into global
conservation planning beyond traditional habitat suitability boundaries.
Our findings challenge the expectation that species richness peaks exclusively in intact
habitats, suggesting instead that many highly modified matrix areas can also support relatively
high levels of terrestrial mammal richness. This pattern aligns with the well-established ‘winners
and losers’ framework, which highlights how some species benefit from human-driven landscape
changes due to increased resource availability, altered predation dynamics, and species filtering
(Fisher and Burton 2018, Kremen and Merenlender 2018, Tucker et al. 2021). Moreover, the
relationship between human footprint and ecological intactness is not always straightforward
(Belote et al. 2020). Many areas with a high human footprint still support high species richness
and retain relatively intact mammal assemblages, particularly in highly productive environments
or regions with a long history of human-wildlife coexistence (Belote et al. 2020, Torres-Romero
et al. 2020, Zhang et al. 2024). For example, in the Peace–Moberly region of northeast British
Columbia, Nitschke (2008) found that across an extensively developed landscape, 22% of
species experienced declines in modelled habitat while the remainder demonstrated increases,
indicating that while some species experienced habitat loss, others were able to persist or even

96

benefit under substantial human modification. Proximity to intact habitats may also play a role,
as species can spill over into adjacent matrix areas when habitat boundaries are permeable,
enabling temporary use or colonization of suboptimal but accessible habitats (Cook et al. 2002,
Brotons et al. 2003, Ewers and Didham 2006). These findings suggest that certain modified
habitats, particularly those embedded within heterogeneous or productive landscapes, can play a
significant role in sustaining terrestrial mammal biodiversity, challenging conventional
assumptions that equate human modification with low ecological value.
Our approach identified hotspots within the matrix based on species richness for all
terrestrial mammals in the database, mammals with declining population trends, and mammals
with threatened status. We acknowledged that these categories were not mutually exclusive, as
declining and threatened mammals constituted subsets of the broader ‘all terrestrial mammals’
group. In practice, a substantial proportion of threatened species are also experiencing population
declines, particularly those listed under IUCN Criterion A (population reduction) and, in many
cases, Criteria B or C, both of which require evidence of continuing decline. This close
alignment in listing criteria contributes to considerable overlap between these groups.
Nevertheless, this multi-faceted strategy allowed us to highlight matrix areas with the potential to
support overall biodiversity as well as species of conservation concern. While our analysis did
not aim to compare responses between these overlapping groups directly, we recognized that
future analyses intended to explicitly distinguish differences among these categories might
benefit from using mutually exclusive groups to enhance analytical clarity and interpretability.
Our results reveal limited spatial overlap among hotspot types—a trend consistent with
previous studies on biodiversity prioritization (Orme et al. 2005, Grenyer et al. 2006, Ceballos
and Ehrlich 2006, Jenkins et al. 2013). This suggests that different hotspot types reflect distinct

97

ecological patterns and conservation opportunities, reinforcing the value of multi-criteria
approaches in conservation planning (Marchese 2015). Building on this, previous studies have
shown that incorporating functional and phylogenetic diversity helps capture species’ ecological
roles, evolutionary history, and contributions to ecosystem resilience (Cadotte et al. 2011, Brum
et al. 2017, Stevens et al. 2019). Including these complementary dimensions could further refine
hotspot identification within matrix areas, supporting conservation strategies that prioritize not
only high species richness but also long-term ecological stability and evolutionary
distinctiveness.
Our results indicate that hotspots based on threatened species capture the largest
proportion of terrestrial mammal diversity within matrix areas compared to any other hotspot
type. Despite their smaller spatial extent, these hotspots encompass over half of all extant
terrestrial mammals, underscoring their disproportionate conservation value. This aligns with
previous studies highlighting the importance of threat-based prioritization to efficiently direct
conservation efforts toward species at the highest risk of extinction (Brooks et al. 2006, Allan et
al. 2019, Wolff et al. 2023, Huais et al. 2025), particularly in human-modified landscapes where
biodiversity loss is most acute (Newbold et al. 2015). While focusing on overall species richness
remains valuable, our findings suggest that threatened species hotspots within matrix areas
represent especially strategic targets for guiding restoration efforts, potentially yielding
substantial conservation returns for terrestrial mammal diversity.
Although our analysis focused on terrestrial mammals, the framework we developed is
readily transferable to other taxonomic groups. The growing availability of AOH maps (e.g.,
Nania et al. 2022, Lumbierres et al. 2022b, Neugarten et al. 2024) offers new opportunities to
assess species richness within matrix areas across broader components of biodiversity. Applying

98

this approach to additional taxa could reveal whether patterns of richness and conservation value
within matrix areas are consistent across groups, supporting more integrative and taxonomically
inclusive conservation strategies.
While our study underscores the importance of matrix areas for biodiversity conservation,
the extent to which these areas actively support species persistence remains a critical knowledge
gap. The ecological value of matrix areas depends fundamentally on their habitat structure,
resource availability, and species-specific ecological requirements (Prevedello and Vieira 2010,
Driscoll et al. 2013). Although species richness is frequently employed as a primary metric for
biodiversity assessments (Marchese 2015, Di Marco et al. 2018a), richness alone does not
necessarily reflect the functional capacity of matrix habitats to sustain viable populations or
distinguish whether these habitats primarily serve as corridors facilitating species dispersal.
Future research should thus complement our findings through detailed species-specific studies
examining population viability, habitat selection, and dispersal patterns within matrix areas.
Additionally, empirical validations leveraging independent observational datasets, such as
citizen-science platforms (e.g., iNaturalist), could confirm the actual usage of matrix habitats by
terrestrial mammals, providing critical evidence beyond predictions based solely on habitat
suitability models (e.g., AOH maps).
We acknowledge an inherent methodological scale mismatch between fine-scale
empirical studies, such as telemetry tracking, occupancy modelling, and camera trapping (e.g.,
Bassing et al. 2023), and broad-scale, global biodiversity analyses such as ours. Detailed studies
typically provide nuanced insights into habitat selection, species-specific ecological responses,
and fine-scale habitat dynamics, but are limited to smaller geographic extents or subsets of
species. Conversely, global-scale biodiversity assessments typically rely on broader proxies,

99

such as habitat suitability and species richness (e.g., Cohen and Jetz 2025), facilitating analyses
across extensive spatial scales and diverse taxonomic groups. Given this mismatch, future efforts
could strategically integrate insights from localized studies with global analyses to better
understand species persistence in matrix areas, ultimately refining global conservation strategies.
A further potential limitation of our global analysis is that geographic variation in species
diversity within matrix areas may partly reflect underlying natural variation in terrestrial
mammal richness across regions. In this study, we intentionally did not standardize or control
explicitly for these baseline regional differences, as our primary objective was to provide a
straightforward global assessment of species richness within matrix areas. Future research could
extend this framework by explicitly incorporating regional mammal richness patterns as
covariates or through other standardization methods. Such an approach would provide additional
insights into whether observed patterns in matrix areas differ meaningfully from general patterns
of regional biodiversity.
As primary habitats continue to be lost or modified, species may increasingly depend on
matrix areas for survival, making it necessary to reassess conservation priorities under dynamic
environmental conditions (Powers and Jetz 2019). However, the extent to which matrix areas
will serve as refugia or become further degraded remains uncertain. Predictive modeling
approaches that integrate climate-driven habitat shifts, landscape connectivity, and land-use
change could help identify matrix areas most likely to support species persistence under future
scenarios (Carvalho et al. 2011, Bateman et al. 2013, Guisan et al. 2013, Powers and Jetz 2019).
Such forward-looking strategies would enable more adaptive conservation efforts, allowing
decision-makers to anticipate species responses to environmental change rather than relying
solely on current biodiversity patterns.

100

Integrating matrix areas into conservation strategies aligns closely with global
biodiversity targets, notably the Kunming-Montreal Global Biodiversity Framework’s Target 2,
which calls for effective restoration of degraded ecosystems to enhance biodiversity, ecological
integrity, and connectivity (Maxwell et al. 2020, Allan et al. 2022, Obura 2023). Given that
expanding formal protected areas faces substantial socioeconomic and political constraints
(Watson et al. 2014, Schleicher et al. 2019, Sandbrook et al. 2023), matrix areas offer
complementary opportunities to achieve biodiversity restoration targets (Kremen and
Merenlender 2018, Galán-Acedo et al. 2019, Mokany et al. 2020). Conservation actions within
matrix areas inherently involve balancing biodiversity conservation with human land-use
demands, directly reflecting the ongoing debate between land-sharing approaches, which
integrate biodiversity conservation within human-dominated landscapes, and land-sparing
approaches, which intensify land use in some areas to allow strict biodiversity protection
elsewhere (Phalan et al. 2011, Fischer et al. 2014, Grass et al. 2019, Williams et al. 2021).
Critically, this perspective moves beyond the oversimplified dichotomy of habitat versus nonhabitat, recognizing that multi-use and managed landscapes frequently retain significant
ecological value and can actively sustain biodiversity (Tscharntke et al. 2005, Fischer et al. 2006,
Kremen and Merenlender 2018). Our findings thus support an adaptive combination of landsharing and land-sparing strategies, highlighting that even highly modified landscapes can retain
substantial biodiversity, particularly when strategically managed or restored (Daily et al. 2001,
2003, Benton et al. 2003, Hole et al. 2005, Gardner et al. 2009, Arroyo‐Rodríguez et al. 2020).
This integrated perspective emphasizes context-specific management interventions that leverage
the ecological realities and socioeconomic conditions of fragmented, multi-use systems (DeFries

101

et al. 2007, Gardner et al. 2009, Perfecto and Vandermeer 2010, Kremen and Merenlender 2018,
Mokany et al. 2020).
Our findings suggest that matrix areas—often overlooked in conservation planning—can
support substantial terrestrial mammal diversity. By identifying species-rich matrix areas and
mapping hotspots of species richness within them, our study expands the spatial scope of
conservation opportunities beyond traditional habitat boundaries. As habitat loss and
fragmentation continue to reshape ecosystems worldwide, prioritizing conservation action in
matrix areas offers a practical and urgently needed strategy to complement existing conservation
networks and help achieve global biodiversity targets. Integrating matrix areas into spatial
planning, habitat management, and restoration will be essential for building resilient, connected
landscapes that sustain biodiversity in a human-dominated world.

102

Chapter 5: Conclusions
In this dissertation, I investigated the role and conservation potential of the matrix for
supporting terrestrial mammal diversity. Specifically, I quantified how matrix condition mediates
the effects of habitat fragmentation on extinction risk, compared the predictive performance of
alternative habitat intactness models (patch-matrix, continuum, and hybrid models), and
identified hotspots of species richness within matrix areas. Despite long-standing conservation
paradigms prioritizing primary habitats (Margules and Pressey 2000, Brooks et al. 2006, Gibson
et al. 2011, Martin et al. 2012, Watson et al. 2018a, Di Marco et al. 2019), my findings clearly
demonstrate the profound and indispensable ecological value of the matrix.
Matrix condition emerged as a key driver of extinction risk for terrestrial mammals
globally, demonstrating greater predictive power than either habitat loss or habitat amount alone
(Chapter 2). Importantly, the negative effects of habitat fragmentation intensified under
conditions of high human pressure within the matrix, underscoring the critical need for
conservation strategies that enhance matrix quality alongside traditional conservation approaches
focused on primary habitats. Additionally, the hybrid habitat intactness model (Chapter 3)
consistently outperformed traditional patch-matrix and continuum models in predicting
extinction risk. By integrating continuous gradients of habitat quality with discrete habitat
patches, the hybrid model better captures ecological complexity, representing an important
methodological advancement for extinction risk assessments.
Moreover, my research provides a robust global assessment of species richness within the
matrix, identifying previously unrecognized hotspots that may disproportionately represent
global terrestrial mammal diversity (Chapter 4). Remarkably, these hotspots occupy only about
1% of Earth's terrestrial surface, yet potentially sustain over half of all terrestrial mammal

103

species. Matrix areas with the highest conservation potential—identified by the overlap among
hotspot types—are primarily concentrated in tropical strongholds such as parts of the Amazon
Basin, the Colombian Tropical Andes, the Brazilian Atlantic Forest, and the Albertine Rift. This
geographic concentration underscores the strategic value of targeted conservation in these areas,
where action could yield disproportionately high returns for global biodiversity conservation.
Furthermore, many of these areas with high conservation potential occur in landscapes facing
intense human pressure yet currently lack adequate protection through formal PAs or recognition
as OECMs, highlighting a critical conservation gap with substantial global implications.

Conservation implications
Given the current biodiversity crisis highlighted by recent landmark assessments (IPBES
2019, Obura 2023), effective conservation strategies must embrace multifunctional landscapes
(Mokany et al. 2020). My work demonstrates that matrix areas can buffer species from extinction
pressures and thus serve as valuable reservoirs of mammalian diversity (Chapters 2 and 3). These
findings directly align with Target 2 of the Kunming-Montreal Global Biodiversity Framework,
which aims to restore at least 30% of degraded terrestrial ecosystems by 2030 (Obura 2023).
Because improving matrix quality can yield higher conservation returns than solely manipulating
the size and configuration of remnant habitat patches (Donald and Evans 2006, Prugh et al. 2008,
Franklin and Lindenmayer 2009, Ruffell et al. 2017, Ramírez-Delgado et al. 2022), targeting
matrix areas identified as hotspots of species richness (Chapter 4) for restoration and improved
management can represent an effective and efficient strategy to partially achieve this ambitious
global restoration target, contributing to broader global biodiversity objectives. Moreover, the
spatially explicit nature of my approach provides opportunities for further refinement to

104

prioritize specific areas within these hotspots, enabling even more focused and effective
conservation action.

Limitations and future directions
While this dissertation advances global understanding of the ecological role and
conservation potential of the matrix for terrestrial mammals, certain methodological and
conceptual limitations offer clear opportunities for future research.
The global-scale analyses presented in Chapters 2 and 3 primarily relied on metrics
derived from human footprint data to predict extinction risk. While informative, these analyses
excluded other pressures influencing extinction risk, such as overhunting (Ripple et al. 2016,
Benítez-López et al. 2017), invasive species (Bellard et al. 2016, Doherty et al. 2016), disease
outbreaks (Daszak et al. 2000, Scheele et al. 2019), climate change (Urban 2015, Foden et al.
2019), and pollution (Hernández et al. 2016, McCune et al. 2019). The omission of these factors
likely increased uncertainty in extinction risk classifications, particularly for species at the
greatest risk of extinction. Future extinction risk analyses could benefit from integrating these
additional pressures, despite the methodological challenges associated with accurately mapping
and quantifying them at global scales (Joppa et al. 2016, Di Marco et al. 2018b).
While changes in habitat intactness and human pressure levels have been explicitly
considered (Chapters 2 and 3), extending these analyses to incorporate future dynamics, such as
climate-driven changes in habitat quality, shifting species distributions, and evolving land-use
pressures, also represents a compelling opportunity for future research. Incorporating dynamic
scenario modeling, including climate change projections and future land-use scenarios, would

105

allow researchers to better anticipate future shifts in extinction risk (Bateman et al. 2013, Powers
and Jetz 2019).
The global hotspots of species richness identified in Chapter 4 relied primarily on globalscale spatial analyses. Although robust and informative in identifying broad patterns of
mammalian richness within matrix areas, these global-scale assessments inherently cannot
capture detailed ecological processes such as local population viability, demographic dynamics
(e.g., survival and reproduction), species behavior, or fine-scale ecological interactions. As such,
the identified hotspots do not directly provide evidence of demographic persistence or ecological
stability of mammal populations within these areas. Empirical studies at finer ecological scales
typically involve more nuanced data and methods, evaluating habitat quality along continuous
gradients rather than simple patch-matrix distinctions, and focus on detailed habitat use and
interactions among a limited set of species (e.g., Bommel et al. 2022, Koetke et al. 2025). Future
research efforts should thus incorporate targeted empirical studies within selected hotspots, using
high-resolution data and detailed ecological assessments. Such focused research would
complement global findings, yielding deeper ecological insights that can refine conservation
strategies and improve their effectiveness at local and regional scales.
While this dissertation focused on terrestrial mammals for both ecological and practical
reasons, the scope could be expanded to other taxa in future work. At the time this research was
initiated, mammals were the only vertebrate group for which globally consistent, high-resolution
habitat data were available, making them the most suitable choice for large-scale analyses
(Rondinini et al. 2011). Comparable high-resolution datasets are now available for birds
(Lumbierres et al. 2022b), whose generally greater mobility and, in many cases, higher
sensitivity to habitat edges (Schmiegelow et al. 1997, Batáry and Báldi 2004, Laurance et al.

106

2011, Bregman et al. 2014) could produce different spatial patterns and conservation priorities.
Applying the analytical framework developed here to birds and other taxa would provide a
valuable test of the generality of these results and could yield complementary insights for global
conservation planning.
My global-scale analysis of species richness hotspots focused on taxonomic diversity
(species richness), without incorporating functional or phylogenetic dimensions of biodiversity.
Functional diversity, which captures ecological roles, resource-use strategies, and trait variation
among species, and phylogenetic diversity, which reflects evolutionary distinctiveness, are
increasingly recognized as critical for ecosystem resilience and adaptive capacity under
environmental change (Cadotte et al. 2011, Brum et al. 2017). Integrating these additional
biodiversity dimensions into hotspot analyses could provide more comprehensive ecological
criteria, helping identify matrix areas that can not only support high numbers of species but also
sustain critical ecosystem functions and evolutionary heritage.
Another limitation relates to practical considerations around restoration. Not all matrix
areas identified as hotspots (Chapter 4) may realistically become suitable for all mammal species
through targeted restoration efforts. Structural mismatches—such as attempting grassland
restoration for forest-dependent species—as well as economic and logistical constraints, could
limit the feasibility and effectiveness of restoration. Future analyses and management
interventions should thus carefully evaluate restoration potential by explicitly considering
species-specific habitat requirements alongside realistic socioeconomic and ecological
constraints.
Additionally, my global assessment of species richness within matrix areas (Chapter 4)
did not explicitly account for baseline regional variations in terrestrial mammal diversity. Thus,

107

the patterns identified may partially reflect inherent regional differences rather than solely
ecological responses to matrix conditions. Future global analyses would benefit from
standardizing or explicitly controlling for regional richness patterns, thereby enhancing the
interpretability and robustness of findings regarding the conservation potential of matrix areas at
a global scale.
Translating the global-scale biodiversity insights from this dissertation into practical
conservation outcomes demands bridging global analyses and local management actions. Globalscale conservation recommendations often face practical challenges in implementation at local
and regional scales, where ecological, socioeconomic, governance, and political contexts differ
widely (Knight et al. 2006, Watson et al. 2014, Di Minin and Toivonen 2015, Guerrero et al.
2018). Future research could help bridge this gap by establishing targeted, carefully designed
adaptive management experiments within selected hotspots of species richness in matrix areas
identified in this dissertation. Such adaptive experiments could aim to integrate ecological
monitoring, socioeconomic assessments, and governance analyses, potentially involving
approaches such as habitat restoration trials (Rey Benayas and Bullock 2012, Suding et al. 2015),
agroforestry practices (Bhagwat et al. 2008, Jose 2009), sustainable land-use initiatives
(Tscharntke et al. 2005, Kennedy et al. 2016), and community-based conservation interventions
(Berkes 2007, Koricha and Jemal Adem 2024). Although logistically challenging, rigorous
evaluation of these integrated management approaches could generate valuable, evidence-based
insights to support effective, contextually adapted, and scalable biodiversity conservation actions
in matrix areas globally.
The findings presented in this dissertation make a compelling case for revisiting and
revising conservation frameworks globally, embracing matrix areas as critical ecological assets

108

rather than marginal landscapes. Such integration would represent a transformative shift in
global biodiversity conservation strategies, better reflecting ecological realities in the
Anthropocene. As the global biodiversity crisis intensifies, recognizing and strategically
managing matrix areas presents one of the most promising pathways for maintaining
biodiversity, ecological connectivity, and ecosystem resilience in increasingly human-modified
landscapes.

109

References
Allan, J. R., Venter, O., Maxwell, S., Bertzky, B., Jones, K., Shi, Y. and Watson, J. E. M. 2017.
Recent increases in human pressure and forest loss threaten many Natural World Heritage
Sites. - Biological Conservation 206: 47–55.
Allan, J. R., Watson, J. E. M., Marco, M. D., O’Bryan, C. J., Possingham, H. P., Atkinson, S. C.
and Venter, O. 2019. Hotspots of human impact on threatened terrestrial vertebrates. PLOS Biology 17: e3000158.
Allan, J. R., Possingham, H. P., Atkinson, S. C., Waldron, A., Di Marco, M., Butchart, S. H. M.,
Adams, V. M., Kissling, W. D., Worsdell, T., Sandbrook, C., Gibbon, G., Kumar, K.,
Mehta, P., Maron, M., Williams, B. A., Jones, K. R., Wintle, B. A., Reside, A. E. and
Watson, J. E. M. 2022. The minimum land area requiring conservation attention to
safeguard biodiversity. - Science 376: 1094–1101.
Allouche, O., Tsoar, A. and Kadmon, R. 2006. Assessing the accuracy of species distribution
models: prevalence, kappa and the true skill statistic (TSS). - Journal of Applied Ecology
43: 1223–1232.
Arregoitia, L. D. V. 2016. Biases, gaps, and opportunities in mammalian extinction risk research.
- Mammal Review 46: 17–29.
Arroyo‐Rodríguez, V., Fahrig, L., Tabarelli, M., Watling, J. I., Tischendorf, L., Benchimol, M.,
Cazetta, E., Faria, D., Leal, I. R., Melo, F. P. L., Morante‐Filho, J. C., Santos, B. A.,
Arasa‐Gisbert, R., Arce‐Peña, N., Cervantes‐López, M. J., Cudney‐Valenzuela, S.,
Galán‐Acedo, C., San‐José, M., Vieira, I. C. G., Slik, J. W. F., Nowakowski, A. J. and
Tscharntke, T. 2020. Designing optimal human-modified landscapes for forest
biodiversity conservation. - Ecology Letters 23: 1404–1420.
Bahn, V. and McGill, B. J. 2013. Testing the predictive performance of distribution models. Oikos 122: 321–331.
Balmford, B., Green, R. E., Onial, M., Phalan, B. and Balmford, A. 2019. How imperfect can
land sparing be before land sharing is more favourable for wild species? - Journal of
Applied Ecology 56: 73–84.
Barlow, J., Lennox, G. D., Ferreira, J., Berenguer, E., Lees, A. C., Nally, R. M., Thomson, J. R.,
Ferraz, S. F. de B., Louzada, J., Oliveira, V. H. F., Parry, L., Ribeiro de Castro Solar, R.,
Vieira, I. C. G., Aragão, L. E. O. C., Begotti, R. A., Braga, R. F., Cardoso, T. M., de
Oliveira, R. C., Souza Jr, C. M., Moura, N. G., Nunes, S. S., Siqueira, J. V., Pardini, R.,
Silveira, J. M., Vaz-de-Mello, F. Z., Veiga, R. C. S., Venturieri, A. and Gardner, T. A.
2016. Anthropogenic disturbance in tropical forests can double biodiversity loss from
deforestation. - Nature 535: 144–147.
Barnosky, A. D., Matzke, N., Tomiya, S., Wogan, G. O. U., Swartz, B., Quental, T. B., Marshall,
C., McGuire, J. L., Lindsey, E. L., Maguire, K. C., Mersey, B. and Ferrer, E. A. 2011.
Has the Earth’s sixth mass extinction already arrived? - Nature 471: 51–57.
110

Bassing, S. B., DeVivo, M., Ganz, T. R., Kertson, B. N., Prugh, L. R., Roussin, T., Satterfield,
L., Windell, R. M., Wirsing, A. J. and Gardner, B. 2023. Are we telling the same story?
Comparing inferences made from camera trap and telemetry data for wildlife monitoring.
- Ecological Applications 33: e2745.
Batáry, P. and Báldi, A. 2004. Evidence of an Edge Effect on Avian Nest Success. Conservation Biology 18: 389–400.
Bateman, B. L., Murphy, H. T., Reside, A. E., Mokany, K. and VanDerWal, J. 2013.
Appropriateness of full-, partial- and no-dispersal scenarios in climate change impact
modelling. - Diversity and Distributions 19: 1224–1234.
Bates, D., Mächler, M., Bolker, B. and Walker, S. 2015. Fitting Linear Mixed-Effects Models
Using lme4. - Journal of Statistical Software 67: 1–48.
Battin, J. 2004. When good animals love bad habitats: ecological traps and the conservation of
animal populations. - Conservation Biology 18: 1482–1491.
Bellard, C., Cassey, P. and Blackburn, T. M. 2016. Alien species as a driver of recent
extinctions. - Biology Letters 12: 20150623.
Belote, R. T., Faurby, S., Brennan, A., Carter, N. H., Dietz, M. S., Hahn, B., McShea, W. J. and
Gage, J. 2020. Mammal species composition reveals new insights into Earth’s remaining
wilderness. - Frontiers in Ecology and the Environment 18: 376–383.
Benítez-López, A., Alkemade, R., Schipper, A. M., Ingram, D. J., Verweij, P. A., Eikelboom, J.
A. J. and Huijbregts, M. A. J. 2017. The impact of hunting on tropical mammal and bird
populations. - Science 356: 180–183.
Benton, T. G., Vickery, J. A. and Wilson, J. D. 2003. Farmland biodiversity: is habitat
heterogeneity the key? - Trends in Ecology & Evolution 18: 182–188.
Berkes, F. 2007. Community-based conservation in a globalized world. - Proceedings of the
National Academy of Sciences 104: 15188–15193.
Betts, M. G., Wolf, C., Ripple, W. J., Phalan, B., Millers, K. A., Duarte, A., Butchart, S. H. M.
and Levi, T. 2017. Global forest loss disproportionately erodes biodiversity in intact
landscapes. - Nature 547: 441–444.
Betts, M. G., Wolf, C., Pfeifer, M., Banks-Leite, C., Arroyo-Rodríguez, V., Ribeiro, D. B.,
Barlow, J., Eigenbrod, F., Faria, D., Fletcher, R. J., Hadley, A. S., Hawes, J. E., Holt, R.
D., Klingbeil, B., Kormann, U., Lens, L., Levi, T., Medina-Rangel, G. F., Melles, S. L.,
Mezger, D., Morante-Filho, J. C., Orme, C. D. L., Peres, C. A., Phalan, B. T., Pidgeon,
A., Possingham, H., Ripple, W. J., Slade, E. M., Somarriba, E., Tobias, J. A., Tylianakis,
J. M., Urbina-Cardona, J. N., Valente, J. J., Watling, J. I., Wells, K., Wearn, O. R.,
Wood, E., Young, R. and Ewers, R. M. 2019. Extinction filters mediate the global effects
of habitat fragmentation on animals. - Science 366: 1236–1239.
111

Beyer, R. M. and Manica, A. 2020. Historical and projected future range sizes of the world’s
mammals, birds, and amphibians. - Nat Commun 11: 5633.
Beyer, H. L., Venter, O., Grantham, H. S. and Watson, J. E. M. 2020. Substantial losses in
ecoregion intactness highlight urgency of globally coordinated action. - Conservation
Letters 13: e12692.
Bhagwat, S. A., Willis, K. J., Birks, H. J. B. and Whittaker, R. J. 2008. Agroforestry: a refuge for
tropical biodiversity? - Trends in Ecology & Evolution 23: 261–267.
Bland, L. M., Orme, C. D. L., Bielby, J., Collen, B., Nicholson, E. and McCarthy, M. A. 2015a.
Cost-effective assessment of extinction risk with limited information. - Journal of
Applied Ecology 52: 861–870.
Bland, L. M., Collen, B., Orme, C. D. L. and Bielby, J. 2015b. Predicting the conservation status
of data-deficient species. - Conservation Biology 29: 250–259.
Boesing, A. L., Nichols, E. and Metzger, J. P. 2018. Biodiversity extinction thresholds are
modulated by matrix type. - Ecography 41: 1520–1533.
Bommel, J. K. van, Sun, C., Ford, A. T., Todd, M. and Burton, A. C. 2022. Coexistence or
conflict: Black bear habitat use along an urban-wildland gradient. - PLOS ONE 17:
e0276448.
Bregman, T. P., Sekercioglu, C. H. and Tobias, J. A. 2014. Global patterns and predictors of bird
species responses to forest fragmentation: Implications for ecosystem function and
conservation. - Biological Conservation 169: 372–383.
Breiman, L. 2001. Random Forests. - Machine Learning 45: 5–32.
Britnell, J. A., Zhu, Y., Kerley, G. I. H. and Shultz, S. 2023. Ecological marginalization is
widespread and increases extinction risk in mammals. - Proceedings of the National
Academy of Sciences 120: e2205315120.
Brooks, T. M., Mittermeier, R. A., da Fonseca, G. A. B., Gerlach, J., Hoffmann, M., Lamoreux,
J. F., Mittermeier, C. G., Pilgrim, J. D. and Rodrigues, A. S. L. 2006. Global Biodiversity
Conservation Priorities. - Science 313: 58–61.
Brooks, T. M., Pimm, S. L., Akçakaya, H. R., Buchanan, G. M., Butchart, S. H. M., Foden, W.,
Hilton-Taylor, C., Hoffmann, M., Jenkins, C. N., Joppa, L., Li, B. V., Menon, V.,
Ocampo-Peñuela, N. and Rondinini, C. 2019. Measuring Terrestrial Area of Habitat
(AOH) and Its Utility for the IUCN Red List. - Trends in Ecology & Evolution 34: 977–
986.
Brotons, L., Mönkkönen, M. and Martin, J. L. 2003. Are Fragments Islands? Landscape Context
and Density‐Area Relationships in Boreal Forest Birds. - The American Naturalist 162:
343–357.
112

Brudvig, L. A., Leroux, S. J., Albert, C. H., Bruna, E. M., Davies, K. F., Ewers, R. M., Levey, D.
J., Pardini, R. and Resasco, J. 2017. Evaluating conceptual models of landscape change. Ecography 40: 74–84.
Brum, F. T., Graham, C. H., Costa, G. C., Hedges, S. B., Penone, C., Radeloff, V. C., Rondinini,
C., Loyola, R. and Davidson, A. D. 2017. Global priorities for conservation across
multiple dimensions of mammalian diversity. - Proceedings of the National Academy of
Sciences 114: 7641–7646.
Bruner, A. G., Gullison, R. E., Rice, R. E. and da Fonseca, G. A. B. 2001. Effectiveness of Parks
in Protecting Tropical Biodiversity. - Science 291: 125–128.
Butchart, S. H. M., Clarke, M., Smith, R. J., Sykes, R. E., Scharlemann, J. P. W., Harfoot, M.,
Buchanan, G. M., Angulo, A., Balmford, A., Bertzky, B., Brooks, T. M., Carpenter, K.
E., Comeros-Raynal, M. T., Cornell, J., Ficetola, G. F., Fishpool, L. D. C., Fuller, R. A.,
Geldmann, J., Harwell, H., Hilton-Taylor, C., Hoffmann, M., Joolia, A., Joppa, L.,
Kingston, N., May, I., Milam, A., Polidoro, B., Ralph, G., Richman, N., Rondinini, C.,
Segan, D. B., Skolnik, B., Spalding, M. D., Stuart, S. N., Symes, A., Taylor, J., Visconti,
P., Watson, J. E. M., Wood, L. and Burgess, N. D. 2015. Shortfalls and Solutions for
Meeting National and Global Conservation Area Targets. - Conservation Letters 8: 329–
337.
Cadotte, M. W., Carscadden, K. and Mirotchnick, N. 2011. Beyond species: functional diversity
and the maintenance of ecological processes and services. - Journal of Applied Ecology
48: 1079–1087.
Cardillo, M., Mace, G. M., Jones, K. E., Bielby, J., Bininda-Emonds, O. R. P., Sechrest, W.,
Orme, C. D. L. and Purvis, A. 2005. Multiple causes of high extinction risk in large
mammal species. - Science 309: 1239–1241.
Cardillo, M., Mace, G. M., Gittleman, J. L., Jones, K. E., Bielby, J. and Purvis, A. 2008. The
predictability of extinction: biological and external correlates of decline in mammals. Proceedings of the Royal Society B: Biological Sciences 275: 1441–1448.
Cardillo, M., Skeels, A. and Dinnage, R. 2023. Priorities for conserving the world’s terrestrial
mammals based on over-the-horizon extinction risk. - Current Biology 33: 1381-1388.e6.
Cardinale, B. J., Duffy, J. E., Gonzalez, A., Hooper, D. U., Perrings, C., Venail, P., Narwani, A.,
Mace, G. M., Tilman, D., Wardle, D. A., Kinzig, A. P., Daily, G. C., Loreau, M., Grace,
J. B., Larigauderie, A., Srivastava, D. S. and Naeem, S. 2012. Biodiversity loss and its
impact on humanity. - Nature 486: 59–67.
Carruthers-Jones, J., Guetté, A., Carver, S., Lefebvre, T., Vallauri, D., Debeir, L., Aykroyd, T.,
Barthod, C., Cavallin, P., Vallée, S., Benest, F., Cherel, E., Kun, Z. and Debuf, O. 2025.
High-resolution naturalness mapping can support conservation policy objectives and
identify locations for strongly protected areas in France. - Commun Earth Environ 6: 1–
11.
113

Carvalho, S. B., Brito, J. C., Crespo, E. G., Watts, M. E. and Possingham, H. P. 2011.
Conservation planning under climate change: Toward accounting for uncertainty in
predicted species distributions to increase confidence in conservation investments in
space and time. - Biological Conservation 144: 2020–2030.
Cazalis, V., Di Marco, M., Butchart, S. H. M., Akçakaya, H. R., González-Suárez, M., Meyer,
C., Clausnitzer, V., Böhm, M., Zizka, A., Cardoso, P., Schipper, A. M., Bachman, S. P.,
Young, B. E., Hoffmann, M., Benítez-López, A., Lucas, P. M., Pettorelli, N., Patoine, G.,
Pacifici, M., Jörger-Hickfang, T., Brooks, T. M., Rondinini, C., Hill, S. L. L., Visconti, P.
and Santini, L. 2022. Bridging the research-implementation gap in IUCN Red List
assessments. - Trends in Ecology & Evolution 37: 359–370.
CBD 2010. COP 10 Decision X/2: Strategic Plan for Biodiversity 2011–2020 and the Aichi
Biodiversity Targets. Convention on Biological Diversity, Nagoya, Japan.
Ceballos, G. and Ehrlich, P. R. 2002. Mammal Population Losses and the Extinction Crisis. Science 296: 904–907.
Ceballos, G. and Ehrlich, P. R. 2006. Global mammal distributions, biodiversity hotspots, and
conservation. - Proceedings of the National Academy of Sciences 103: 19374–19379.
Ceballos, G., Ehrlich, P. R., Barnosky, A. D., García, A., Pringle, R. M. and Palmer, T. M. 2015.
Accelerated modern human–induced species losses: Entering the sixth mass extinction. Science Advances 1: e1400253.
Ceballos, G., Ehrlich, P. R. and Dirzo, R. 2017. Biological annihilation via the ongoing sixth
mass extinction signaled by vertebrate population losses and declines. - Proceedings of
the National Academy of Sciences 114: E6089–E6096.
Ceballos, G., Ehrlich, P. R. and Raven, P. H. 2020. Vertebrates on the brink as indicators of
biological annihilation and the sixth mass extinction. - Proceedings of the National
Academy of Sciences 117: 13596–13602.
Charrette, N. A., Cleary, D. F. R. and Mooers, A. Ø. 2006. Range-Restricted, Specialist Bornean
Butterflies Are Less Likely to Recover from Enso-Induced Disturbance. - Ecology 87:
2330–2337.
Cohen, J. 1988. Statistical Power Analysis for the Behavioral Sciences. - Academic Press, New
York.
Cohen, J. M. and Jetz, W. 2025. Fine-Grain Predictions Are Key to Accurately Represent
Continental-Scale Biodiversity Patterns. - Global Ecology and Biogeography 34: e13934.
Cook, W. M., Lane, K. T., Foster, B. L. and Holt, R. D. 2002. Island theory, matrix effects and
species richness patterns in habitat fragments. - Ecology Letters 5: 619–623.

114

Côté, I. M., Darling, E. S. and Brown, C. J. 2016. Interactions among ecosystem stressors and
their importance in conservation. - Proceedings of the Royal Society B: Biological
Sciences 283: 20152592.
Craigie, I. D., Baillie, J. E. M., Balmford, A., Carbone, C., Collen, B., Green, R. E. and Hutton,
J. M. 2010. Large mammal population declines in Africa’s protected areas. - Biological
Conservation 143: 2221–2228.
Crooks, K. R., Burdett, C. L., Theobald, D. M., Rondinini, C. and Boitani, L. 2011. Global
patterns of fragmentation and connectivity of mammalian carnivore habitat. Philosophical Transactions of the Royal Society B: Biological Sciences 366: 2642–2651.
Crooks, K. R., Burdett, C. L., Theobald, D. M., King, S. R. B., Marco, M. D., Rondinini, C. and
Boitani, L. 2017. Quantification of habitat fragmentation reveals extinction risk in
terrestrial mammals. - PNAS 114: 7635–7640.
Cushman, S. A., McKelvey, K. S., Hayden, J. and Schwartz, M. K. 2006. Gene Flow in Complex
Landscapes: Testing Multiple Hypotheses with Causal Modeling. - The American
Naturalist 168: 486–499.
Cutler, D. R., Edwards, T. C., Beard, K. H., Cutler, A., Hess, K. T., Gibson, J. and Lawler, J. J.
2007. Random forests for Classification in ecology. - Ecology 88: 2783–2792.
Daily, G. C., Ehrlich, P. R. and Sánchez-Azofeifa, G. A. 2001. Countryside Biogeography: Use
of Human-Dominated Habitats by the Avifauna of Southern Costa Rica. - Ecological
Applications 11: 1–13.
Daily, G. C., Ceballos, G., Pacheco, J., Suzán, G. and Sánchez‐Azofeifa, A. 2003. Countryside
biogeography of Neotropical mammals: conservation opportunities in agricultural
landscapes of Costa Rica. - Conservation Biology 17: 1814–1826.
Daszak, P., Cunningham, A. A. and Hyatt, A. D. 2000. Emerging Infectious Diseases of
Wildlife-- Threats to Biodiversity and Human Health. - Science 287: 443–449.
Davidson, A. D., Hamilton, M. J., Boyer, A. G., Brown, J. H. and Ceballos, G. 2009. Multiple
ecological pathways to extinction in mammals. - PNAS 106: 10702–10705.
De Vos, J. M., Joppa, L. N., Gittleman, J. L., Stephens, P. R. and Pimm, S. L. 2015. Estimating
the normal background rate of species extinction. - Conservation Biology 29: 452–462.
DeFries, R., Hansen, A., Turner, B. L., Reid, R. and Liu, J. 2007. Land Use Change Around
Protected Areas: Management to Balance Human Needs and Ecological Function. Ecological Applications 17: 1031–1038.
Devictor, V., Julliard, R. and Jiguet, F. 2008. Distribution of specialist and generalist species
along spatial gradients of habitat disturbance and fragmentation. - Oikos 117: 507–514.

115

Di Marco, M., Rondinini, C., Boitani, L. and Murray, K. A. 2013. Comparing multiple species
distribution proxies and different quantifications of the human footprint map,
implications for conservation. - Biological Conservation 165: 203–211.
Di Marco, M., Boitani, L., Mallon, D., Hoffmann, M., Iacucci, A., Meijaard, E., Visconti, P.,
Schipper, J. and Rondinini, C. 2014a. A Retrospective Evaluation of the Global Decline
of Carnivores and Ungulates. - Conservation Biology 28: 1109–1118.
Di Marco, M., Buchanan, G. M., Szantoi, Z., Holmgren, M., Grottolo Marasini, G., Gross, D.,
Tranquilli, S., Boitani, L. and Rondinini, C. 2014b. Drivers of extinction risk in African
mammals: the interplay of distribution state, human pressure, conservation response and
species biology. - Philosophical Transactions of the Royal Society B: Biological Sciences
369: 20130198.
Di Marco, M., Collen, B., Rondinini, C. and Mace, G. M. 2015. Historical drivers of extinction
risk: using past evidence to direct future monitoring. - Proceedings of the Royal Society
B: Biological Sciences 282: 20150928.
Di Marco, M., Watson, J. E. M., Currie, D. J., Possingham, H. P. and Venter, O. 2018a. The
extent and predictability of the biodiversity–carbon correlation. - Ecology Letters 21:
365–375.
Di Marco, M., Venter, O., Possingham, H. P. and Watson, J. E. M. 2018b. Changes in human
footprint drive changes in species extinction risk. - Nature Communications 9: 4621.
Di Marco, M., Ferrier, S., Harwood, T. D., Hoskins, A. J. and Watson, J. E. M. 2019. Wilderness
areas halve the extinction risk of terrestrial biodiversity. - Nature 573: 582–585.
Di Minin, E. and Toivonen, T. 2015. Global Protected Area Expansion: Creating More than
Paper Parks. - BioScience 65: 637–638.
Díaz, S., Settele, J., Brondízio, E. S., Ngo, H. T., Agard, J., Arneth, A., Balvanera, P., Brauman,
K. A., Butchart, S. H. M., Chan, K. M. A., Garibaldi, L. A., Ichii, K., Liu, J.,
Subramanian, S. M., Midgley, G. F., Miloslavich, P., Molnár, Z., Obura, D., Pfaff, A.,
Polasky, S., Purvis, A., Razzaque, J., Reyers, B., Chowdhury, R. R., Shin, Y.-J.,
Visseren-Hamakers, I., Willis, K. J. and Zayas, C. N. 2019. Pervasive human-driven
decline of life on Earth points to the need for transformative change. - Science 366:
eaax3100.
Didham, R. K., Kapos, V. and Ewers, R. M. 2012. Rethinking the conceptual foundations of
habitat fragmentation research. - Oikos 121: 161–170.
Dinerstein, E., Olson, D., Joshi, A., Vynne, C., Burgess, N. D., Wikramanayake, E., Hahn, N.,
Palminteri, S., Hedao, P., Noss, R., Hansen, M., Locke, H., Ellis, E. C., Jones, B., Barber,
C. V., Hayes, R., Kormos, C., Martin, V., Crist, E., Sechrest, W., Price, L., Baillie, J. E.
M., Weeden, D., Suckling, K., Davis, C., Sizer, N., Moore, R., Thau, D., Birch, T.,
Potapov, P., Turubanova, S., Tyukavina, A., de Souza, N., Pintea, L., Brito, J. C.,
Llewellyn, O. A., Miller, A. G., Patzelt, A., Ghazanfar, S. A., Timberlake, J., Klöser, H.,
116

Shennan-Farpón, Y., Kindt, R., Lillesø, J.-P. B., van Breugel, P., Graudal, L., Voge, M.,
Al-Shammari, K. F. and Saleem, M. 2017. An Ecoregion-Based Approach to Protecting
Half the Terrestrial Realm. - BioScience 67: 534–545.
Dirzo, R., Young, H. S., Galetti, M., Ceballos, G., Isaac, N. J. B. and Collen, B. 2014.
Defaunation in the Anthropocene. - Science 345: 401–406.
Doherty, T. S., Glen, A. S., Nimmo, D. G., Ritchie, E. G. and Dickman, C. R. 2016. Invasive
predators and global biodiversity loss. - Proceedings of the National Academy of
Sciences 113: 11261–11265.
Donald, P. F. and Evans, A. D. 2006. Habitat connectivity and matrix restoration: the wider
implications of agri-environment schemes. - Journal of Applied Ecology 43: 209–218.
Driscoll, D. A., Banks, S. C., Barton, P. S., Lindenmayer, D. B. and Smith, A. L. 2013.
Conceptual domain of the matrix in fragmented landscapes. - Trends in Ecology &
Evolution 28: 605–613.
Dudley, N. 2008. Guidelines for applying protected area management categories. - IUCN.
Ellis, E. C. 2019. Sharing the land between nature and people. - Science 364: 1226–1228.
Ellis, E. C. 2021. Land Use and Ecological Change: A 12,000-Year History. - Annual Review of
Environment and Resources 46: 1–33.
Ellis, E. C., Gauthier, N., Klein Goldewijk, K., Bliege Bird, R., Boivin, N., Díaz, S., Fuller, D.
Q., Gill, J. L., Kaplan, J. O., Kingston, N., Locke, H., McMichael, C. N. H., Ranco, D.,
Rick, T. C., Shaw, M. R., Stephens, L., Svenning, J.-C. and Watson, J. E. M. 2021.
People have shaped most of terrestrial nature for at least 12,000 years. - Proceedings of
the National Academy of Sciences 118: e2023483118.
ESRI 2021. ArcGIS Pro, Version 2.8.2.
ESRI 2022. ArcGIS Pro, Version 2.5.0.
ESRI 2024. ArcGIS Pro, Version 3.4.0.
Estes, J. A., Terborgh, J., Brashares, J. S., Power, M. E., Berger, J., Bond, W. J., Carpenter, S.
R., Essington, T. E., Holt, R. D., Jackson, J. B. C., Marquis, R. J., Oksanen, L., Oksanen,
T., Paine, R. T., Pikitch, E. K., Ripple, W. J., Sandin, S. A., Scheffer, M., Schoener, T.
W., Shurin, J. B., Sinclair, A. R. E., Soulé, M. E., Virtanen, R. and Wardle, D. A. 2011.
Trophic Downgrading of Planet Earth. - Science 333: 301–306.
Ewers, R. M. and Didham, R. K. 2006. Confounding factors in the detection of species responses
to habitat fragmentation. - Biological Reviews 81: 117–142.
Fahrig, L. 2001. How much habitat is enough? - Biological Conservation 100: 65–74.

117

Fahrig, L. 2003. Effects of Habitat Fragmentation on Biodiversity. - Annu. Rev. Ecol. Evol. Syst.
34: 487–515.
Fahrig, L. 2013. Rethinking patch size and isolation effects: the habitat amount hypothesis. Journal of Biogeography 40: 1649–1663.
Fahrig, L. 2017. Ecological responses to habitat fragmentation per se. - Annu. Rev. Ecol. Evol.
Syst. 48: 1–23.
Fahrig, L. 2019. Habitat fragmentation: A long and tangled tale. - Global Ecology and
Biogeography 28: 33–41.
Fahrig, L. and Rytwinski, T. 2009. Effects of roads on animal abundance: an empirical review
and synthesis. - Ecology and Society in press.
Fahrig, L., Baudry, J., Brotons, L., Burel, F. G., Crist, T. O., Fuller, R. J., Sirami, C.,
Siriwardena, G. M. and Martin, J.-L. 2011. Functional landscape heterogeneity and
animal biodiversity in agricultural landscapes. - Ecology Letters 14: 101–112.
Fahrig, L., Arroyo-Rodríguez, V., Bennett, J. R., Boucher-Lalonde, V., Cazetta, E., Currie, D. J.,
Eigenbrod, F., Ford, A. T., Harrison, S. P., Jaeger, J. A. G., Koper, N., Martin, A. E.,
Martin, J.-L., Metzger, J. P., Morrison, P., Rhodes, J. R., Saunders, D. A., Simberloff, D.,
Smith, A. C., Tischendorf, L., Vellend, M. and Watling, J. I. 2019. Is habitat
fragmentation bad for biodiversity? - Biological Conservation 230: 179–186.
Faurby, S., Davis, M., Pedersen, R. Ø., Schowanek, S. D., Antonelli1, A. and Svenning, J.-C.
2018. PHYLACINE 1.2: the phylogenetic atlas of mammal macroecology. - Ecology 99:
2626–2626.
Ferreira, A. S., Peres, C. A., Bogoni, J. A. and Cassano, C. R. 2018. Use of agroecosystem
matrix habitats by mammalian carnivores (Carnivora): a global-scale analysis. - Mammal
Review 48: 312–327.
Fischer, J. and Lindenmayer, D. B. 2006. Beyond fragmentation: the continuum model for fauna
research and conservation in human-modified landscapes. - Oikos 112: 473–480.
Fischer, J. and Lindenmayer, D. B. 2007. Landscape modification and habitat fragmentation: a
synthesis. - Global Ecology and Biogeography 16: 265–280.
Fischer, J., Lindenmayer, D. B. and Manning, A. D. 2006. Biodiversity, ecosystem function, and
resilience: ten guiding principles for commodity production landscapes. - Frontiers in
Ecology and the Environment 4: 80–86.
Fischer, J., Abson, D. J., Butsic, V., Chappell, M. J., Ekroos, J., Hanspach, J., Kuemmerle, T.,
Smith, H. G. and von Wehrden, H. 2014. Land Sparing Versus Land Sharing: Moving
Forward. - Conservation Letters 7: 149–157.

118

Fisher, J. T. and Burton, A. C. 2018. Wildlife winners and losers in an oil sands landscape. Frontiers in Ecology and the Environment 16: 323–328.
Fisher, D. O., Blomberg, S. P. and Owens, I. P. F. 2003. Extrinsic versus intrinsic factors in the
decline and extinction of Australian marsupials. - Proceedings of the Royal Society of
London. Series B: Biological Sciences 270: 1801–1808.
Fletcher, R. J., Didham, R. K., Banks-Leite, C., Barlow, J., Ewers, R. M., Rosindell, J., Holt, R.
D., Gonzalez, A., Pardini, R., Damschen, E. I., Melo, F. P. L., Ries, L., Prevedello, J. A.,
Tscharntke, T., Laurance, W. F., Lovejoy, T. and Haddad, N. M. 2018. Is habitat
fragmentation good for biodiversity? - Biological Conservation 226: 9–15.
Fletcher, R. J., Smith, T. A. H., Troy, S., Kortessis, N., Turner, E. C., Bruna, E. M. and Holt, R.
D. 2024. The Prominent Role of the Matrix in Ecology, Evolution, and Conservation. Annual Review of Ecology, Evolution, and Systematics 55: 423–447.
Foden, W. B., Young, B. E., Akçakaya, H. R., Garcia, R. A., Hoffmann, A. A., Stein, B. A.,
Thomas, C. D., Wheatley, C. J., Bickford, D., Carr, J. A., Hole, D. G., Martin, T. G.,
Pacifici, M., Pearce-Higgins, J. W., Platts, P. J., Visconti, P., Watson, J. E. M. and
Huntley, B. 2019. Climate change vulnerability assessment of species. - WIREs Climate
Change 10: e551.
Franklin, J. F. and Lindenmayer, D. B. 2009. Importance of matrix habitats in maintaining
biological diversity. - PNAS 106: 349–350.
Galán-Acedo, C., Arroyo-Rodríguez, V., Andresen, E., Verde Arregoitia, L., Vega, E., Peres, C.
A. and Ewers, R. M. 2019. The conservation value of human-modified landscapes for the
world’s primates. - Nature Communications 10: 152.
Galetti, M., Moleón, M., Jordano, P., Pires, M. M., Guimarães Jr., P. R., Pape, T., Nichols, E.,
Hansen, D., Olesen, J. M., Munk, M., de Mattos, J. S., Schweiger, A. H., Owen-Smith,
N., Johnson, C. N., Marquis, R. J. and Svenning, J.-C. 2018. Ecological and evolutionary
legacy of megafauna extinctions. - Biological Reviews 93: 845–862.
Gardner, T. A., Barlow, J., Chazdon, R., Ewers, R. M., Harvey, C. A., Peres, C. A. and Sodhi, N.
S. 2009. Prospects for tropical forest biodiversity in a human-modified world. - Ecology
Letters 12: 561–582.
Gassert, F., Venter, O., Watson, J. E. M., Brumby, S. P., Mazzariello, J. C., Atkinson, S. C. and
Hyde, S. 2023. An operational approach to near real time global high resolution mapping
of the terrestrial Human Footprint. - Front. Remote Sens. in press.
Geldmann, J., Manica, A., Burgess, N. D., Coad, L. and Balmford, A. 2019. A global-level
assessment of the effectiveness of protected areas at resisting anthropogenic pressures. Proceedings of the National Academy of Sciences 116: 23209–23215.

119

Gibson, L., Lee, T. M., Koh, L. P., Brook, B. W., Gardner, T. A., Barlow, J., Peres, C. A.,
Bradshaw, C. J. A., Laurance, W. F., Lovejoy, T. E. and Sodhi, N. S. 2011. Primary
forests are irreplaceable for sustaining tropical biodiversity. - Nature 478: 378–381.
Grass, I., Loos, J., Baensch, S., Batáry, P., Librán-Embid, F., Ficiciyan, A., Klaus, F., Riechers,
M., Rosa, J., Tiede, J., Udy, K., Westphal, C., Wurz, A. and Tscharntke, T. 2019. Landsharing/-sparing connectivity landscapes for ecosystem services and biodiversity
conservation. - People and Nature 1: 262–272.
Green, R. E., Cornell, S. J., Scharlemann, J. P. W. and Balmford, A. 2005. Farming and the fate
of wild nature. - Science 307: 550–555.
Grenyer, R., Orme, C. D. L., Jackson, S. F., Thomas, G. H., Davies, R. G., Davies, T. J., Jones,
K. E., Olson, V. A., Ridgely, R. S., Rasmussen, P. C., Ding, T.-S., Bennett, P. M.,
Blackburn, T. M., Gaston, K. J., Gittleman, J. L. and Owens, I. P. F. 2006. Global
distribution and conservation of rare and threatened vertebrates. - Nature 444: 93–96.
Guerrero, A., Bennett, N., Wilson, K., Carter, N., Gill, D., Mills, M., Ives, C., Selinske, M.,
Larrosa, C., Bekessy, S., Januchowski-Hartley, F., Travers, H., Wyborn, C. and Nuno, A.
2018. Achieving the promise of integration in social-ecological research: a review and
prospectus. - Ecology and Society in press.
Guisan, A., Tingley, R., Baumgartner, J. B., Naujokaitis-Lewis, I., Sutcliffe, P. R., Tulloch, A. I.
T., Regan, T. J., Brotons, L., McDonald-Madden, E., Mantyka-Pringle, C., Martin, T. G.,
Rhodes, J. R., Maggini, R., Setterfield, S. A., Elith, J., Schwartz, M. W., Wintle, B. A.,
Broennimann, O., Austin, M., Ferrier, S., Kearney, M. R., Possingham, H. P. and
Buckley, Y. M. 2013. Predicting species distributions for conservation decisions. Ecology Letters 16: 1424–1435.
Haddad, N. M., Brudvig, L. A., Clobert, J., Davies, K. F., Gonzalez, A., Holt, R. D., Lovejoy, T.
E., Sexton, J. O., Austin, M. P., Collins, C. D., Cook, W. M., Damschen, E. I., Ewers, R.
M., Foster, B. L., Jenkins, C. N., King, A. J., Laurance, W. F., Levey, D. J., Margules, C.
R., Melbourne, B. A., Nicholls, A. O., Orrock, J. L., Song, D.-X. and Townshend, J. R.
2015. Habitat fragmentation and its lasting impact on Earth’s ecosystems. - Science
Advances 1: e1500052.
Haila, Y. 2002. A conceptual genealogy of fragmentation research: from island biogeography to
landscape ecology. - Ecological Applications 12: 321–334.
Hall, L. S., Krausman, P. R. and Morrison, M. L. 1997. The Habitat Concept and a Plea for
Standard Terminology. - Wildlife Society Bulletin (1973-2006) 25: 173–182.
Hanski, I. 2015. Habitat fragmentation and species richness. - Journal of Biogeography 42: 989–
993.
Hanski, I., Zurita, G. A., Bellocq, M. I. and Rybicki, J. 2013. Species–fragmented area
relationship. - Proceedings of the National Academy of Sciences 110: 12715–12720.
120

Hanson, J. O. 2022. wdpar: Interface to the World Database on ProtectedAreas. - JOSS 7: 4594.
Henle, K., Davies, K. F., Kleyer, M., Margules, C. and Settele, J. 2004. Predictors of Species
Sensitivity to Fragmentation. - Biodiversity and Conservation 13: 207–251.
Hernández, D. L., Vallano, D. M., Zavaleta, E. S., Tzankova, Z., Pasari, J. R., Weiss, S.,
Selmants, P. C. and Morozumi, C. 2016. Nitrogen Pollution Is Linked to US Listed
Species Declines. - BioScience 66: 213–222.
Hoffmann, M., Hilton-Taylor, C., Angulo, A., Böhm, M., Stuart, S. N., et al. 2010. The impact of
conservation on the status of the world’s vertebrates. - Science 330: 1503–1509.
Hoffmann, M., Belant, J. L., Chanson, J. S., Cox, N. A., Lamoreux, J., Rodrigues, A. S. L.,
Schipper, J. and Stuart, S. N. 2011. The changing fates of the world’s mammals. Philosophical Transactions of the Royal Society B: Biological Sciences in press.
Hole, D. G., Perkins, A. J., Wilson, J. D., Alexander, I. H., Grice, P. V. and Evans, A. D. 2005.
Does organic farming benefit biodiversity? - Biological Conservation 122: 113–130.
Huais, P. Y., Osorio-Olvera, L., Cordier, J. M., Tomba, A. N., Soberón, J., Loyola, R. and Nori,
J. 2025. Rethinking Global Hotspots for Threatened Terrestrial Vertebrates. - Global
Ecology and Biogeography 34: e13942.
IPBES 2019. Global assessment report on biodiversity and ecosystem services of the
Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services.
IUCN 2012. IUCN Red List of Threatened Species. Version 2012-1.
IUCN 2016. A global standard for the identification of Key Biodiversity Areas: version 1.0. IUCN.
IUCN 2020. Guidelines for using A global standard for the identification of Key Biodiversity
Areas: version 1.1. - IUCN.
IUCN 2021. IUCN Red List of Threatened Species. Version 2021-3.
IUCN 2024. IUCN Red List of Threatened Species. Version 2024-2.
IUCN 2025. IUCN Red List of Threatened Species. Version 2025-1.
Jaureguiberry, P., Titeux, N., Wiemers, M., Bowler, D. E., Coscieme, L., Golden, A. S., Guerra,
C. A., Jacob, U., Takahashi, Y., Settele, J., Díaz, S., Molnár, Z. and Purvis, A. 2022. The
direct drivers of recent global anthropogenic biodiversity loss. - Science Advances 8:
eabm9982.
Jenkins, C. N., Pimm, S. L. and Joppa, L. N. 2013. Global patterns of terrestrial vertebrate
diversity and conservation. - Proceedings of the National Academy of Sciences 110:
E2602–E2610.
121

Jetz, W., Carbone, C., Fulford, J. and Brown, J. H. 2004. The scaling of animal space use. Science 306: 266–268.
Jones, K. E., Bielby, J., Cardillo, M., Fritz, S. A., O’Dell, J., Orme, C. D. L., Safi, K., Sechrest,
W., Boakes, E. H., Carbone, C., Connolly, C., Cutts, M. J., Foster, J. K., Grenyer, R.,
Habib, M., Plaster, C. A., Price, S. A., Rigby, E. A., Rist, J., Teacher, A., BinindaEmonds, O. R. P., Gittleman, J. L., Mace, G. M. and Purvis, A. 2009. PanTHERIA: a
species-level database of life history, ecology, and geography of extant and recently
extinct mammals. - Ecology 90: 2648–2648.
Jones, K. R., Venter, O., Fuller, R. A., Allan, J. R., Maxwell, S. L., Negret, P. J. and Watson, J.
E. M. 2018. One-third of global protected land is under intense human pressure. - Science
360: 788–791.
Joppa, L. N., O’Connor, B., Visconti, P., Smith, C., Geldmann, J., Hoffmann, M., Watson, J. E.
M., Butchart, S. H. M., Virah-Sawmy, M., Halpern, B. S., Ahmed, S. E., Balmford, A.,
Sutherland, W. J., Harfoot, M., Hilton-Taylor, C., Foden, W., Minin, E. D., Pagad, S.,
Genovesi, P., Hutton, J. and Burgess, N. D. 2016. Filling in biodiversity threat gaps. Science 352: 416–418.
Jose, S. 2009. Agroforestry for ecosystem services and environmental benefits: an overview. Agroforest Syst 76: 1–10.
Kennedy, C. M., Hawthorne, P. L., Miteva, D. A., Baumgarten, L., Sochi, K., Matsumoto, M.,
Evans, J. S., Polasky, S., Hamel, P., Vieira, E. M., Develey, P. F., Sekercioglu, C. H.,
Davidson, A. D., Uhlhorn, E. M. and Kiesecker, J. 2016. Optimizing land use decisionmaking to sustain Brazilian agricultural profits, biodiversity and ecosystem services. Biological Conservation 204: 221–230.
Kennedy, C. M., Oakleaf, J. R., Theobald, D. M., Baruch-Mordo, S. and Kiesecker, J. 2019.
Managing the middle: A shift in conservation priorities based on the global human
modification gradient. - Global Change Biology 25: 811–826.
Kissling, W. D., Dalby, L., Fløjgaard, C., Lenoir, J., Sandel, B., Sandom, C., Trøjelsgaard, K.
and Svenning, J.-C. 2014. Establishing macroecological trait datasets: digitalization,
extrapolation, and validation of diet preferences in terrestrial mammals worldwide. Ecology and Evolution 4: 2913–2930.
Knight, A. T., Driver, A., Cowling, R. M., Maze, K., Desmet, P. G., Lombard, A. T., Rouget, M.,
Botha, M. A., Boshoff, A. F., Castley, J. G., Goodman, P. S., Mackinnon, K., Pierce, S.
M., Sims-Castley, R., Stewart, W. I. and Von Hase, A. 2006. Designing Systematic
Conservation Assessments that Promote Effective Implementation: Best Practice from
South Africa. - Conservation Biology 20: 739–750.
Koetke, L. J., Hodder, D. P., Johnson, C. J. and Anderson, M. 2025. Risk of mortality and
maternal status are key determinants of the dynamic use of cover by moose in a
landscape characterized by forest harvesting. - Oecologia 207: 64.
122

Koricha, H. G. and Jemal Adem, M. 2024. Investigated the role of community based approaches
for biodiversity conservation and socio-economic development in Bale Mountains
National Park, Southeast Ethiopia. - Sci Rep 14: 12241.
Krausman, P. R. and Morrison, M. L. 2016. Another plea for standard terminology. - The Journal
of Wildlife Management 80: 1143–1144.
Kremen, C. and Merenlender, A. M. 2018. Landscapes that work for biodiversity and people. Science 362: eaau6020.
Kremen, C., Cameron, A., Moilanen, A., Phillips, S. J., Thomas, C. D., Beentje, H., Dransfield,
J., Fisher, B. L., Glaw, F., Good, T. C., Harper, G. J., Hijmans, R. J., Lees, D. C., Louis,
E., Nussbaum, R. A., Raxworthy, C. J., Razafimpahanana, A., Schatz, G. E., Vences, M.,
Vieites, D. R., Wright, P. C. and Zjhra, M. L. 2008. Aligning Conservation Priorities
Across Taxa in Madagascar with High-Resolution Planning Tools. - Science 320: 222–
226.
Kuhn, M. 2008. Building predictive models in R using the caret package. - J Stat Softw 28: 1–26.
Laurance, W. F. 1991. Ecological correlates of extinction proneness in Australian tropical rain
forest mammals. - Conservation Biology 5: 79–89.
Laurance, W. F. 2008. Theory meets reality: How habitat fragmentation research has transcended
island biogeographic theory. - Biological Conservation 141: 1731–1744.
Laurance, W. F. and Useche, D. C. 2009. Environmental synergisms and extinctions of tropical
species. - Conservation Biology 23: 1427–1437.
Laurance, W. F., Camargo, J. L. C., Luizão, R. C. C., Laurance, S. G., Pimm, S. L., Bruna, E.
M., Stouffer, P. C., Bruce Williamson, G., Benítez-Malvido, J., Vasconcelos, H. L., Van
Houtan, K. S., Zartman, C. E., Boyle, S. A., Didham, R. K., Andrade, A. and Lovejoy, T.
E. 2011. The fate of Amazonian forest fragments: A 32-year investigation. - Biological
Conservation 144: 56–67.
Law, E. A. and Wilson, K. A. 2015. Providing context for the land-sharing and land-sparing
debate. - Conservation Letters 8: 404–413.
Li, B.-L. and Archer, S. 1997. Weighted mean patch size: a robust index for quantifying
landscape structure. - Ecological Modelling 102: 353–361.
Liaw, A. and Wiener, M. 2002. Classification and regression by randomForest. - R news 2: 18–
22.
Lucas, P. M., González‐Suárez, M. and Revilla, E. 2019. Range area matters, and so does spatial
configuration: predicting conservation status in vertebrates. - Ecography 42: 1103–1114.

123

Lumbierres, M., Dahal, P. R., Di Marco, M., Butchart, S. H. M., Donald, P. F. and Rondinini, C.
2022a. Translating habitat class to land cover to map area of habitat of terrestrial
vertebrates. - Conservation Biology 36: e13851.
Lumbierres, M., Dahal, P. R., Soria, C. D., Di Marco, M., Butchart, S. H. M., Donald, P. F. and
Rondinini, C. 2022b. Area of Habitat maps for the world’s terrestrial birds and mammals.
- Sci Data 9: 749.
MacArthur, R. H. and Wilson, E. O. 1967. The Theory of Island Biogeography. - Princeton
University Press.
Marchese, C. 2015. Biodiversity hotspots: A shortcut for a more complicated concept. - Global
Ecology and Conservation 3: 297–309.
Margules, C. R. and Pressey, R. L. 2000. Systematic conservation planning. - Nature 405: 243–
253.
Maron, M., Simmonds, J. S. and Watson, J. E. M. 2018. Bold nature retention targets are
essential for the global environment agenda. - Nature Ecology & Evolution 2: 1194–
1195.
Martin, L. J., Blossey, B. and Ellis, E. 2012. Mapping where ecologists work: biases in the
global distribution of terrestrial ecological observations. - Frontiers in Ecology and the
Environment 10: 195–201.
Maxwell, S. L., Fuller, R. A., Brooks, T. M. and Watson, J. E. M. 2016. Biodiversity: the
ravages of guns, nets and bulldozers. - Nature News 536: 143.
Maxwell, S. L., Cazalis, V., Dudley, N., Hoffmann, M., Rodrigues, A. S. L., Stolton, S.,
Visconti, P., Woodley, S., Kingston, N., Lewis, E., Maron, M., Strassburg, B. B. N.,
Wenger, A., Jonas, H. D., Venter, O. and Watson, J. E. M. 2020. Area-based
conservation in the twenty-first century. - Nature 586: 217–227.
May, S. A. and Norton, T. W. 1996. Influence of fragmentation and disturbance on the potential
impact of feral predators on native fauna in Australian forest ecosystems. - Wildl. Res.
23: 387–400.
McCune, J. L., Colla, S. R., Coristine, L. E., Davy, C. M., Flockhart, D. T. T., Schuster, R. and
Orihel, D. M. 2019. Are we accurately estimating the potential role of pollution in the
decline of species at risk in Canada? - FACETS 4: 598–614.
McKinney, M. L. and Lockwood, J. L. 1999. Biotic homogenization: a few winners replacing
many losers in the next mass extinction. - Trends in Ecology & Evolution 14: 450–453.
McNab, B. K. 1986. The influence of food habits on the energetics of eutherian mammals. Ecological Monographs 56: 1–19.

124

Mendenhall, C. D., Karp, D. S., Meyer, C. F. J., Hadly, E. A. and Daily, G. C. 2014. Predicting
biodiversity change and averting collapse in agricultural landscapes. - Nature 509: 213–
217.
Miller-Rushing, A. J., Primack, R. B., Devictor, V., Corlett, R. T., Cumming, G. S., Loyola, R.,
Maas, B. and Pejchar, L. 2019. How does habitat fragmentation affect biodiversity? A
controversial question at the core of conservation biology. - Biological Conservation 232:
271–273.
Mokany, K., Ferrier, S., Harwood, T. D., Ware, C., Di Marco, M., Grantham, H. S., Venter, O.,
Hoskins, A. J. and Watson, J. E. M. 2020. Reconciling global priorities for conserving
biodiversity habitat. - Proceedings of the National Academy of Sciences 117: 9906–9911.
Molnar, C. and Schratz, P. 2020. iml: Interpretable Machine Learning.
Morante‐Filho, J. C., Arroyo‐Rodríguez, V., Pessoa, M. de S., Cazetta, E. and Faria, D. 2018.
Direct and cascading effects of landscape structure on tropical forest and non-forest
frugivorous birds. - Ecological Applications 28: 2024–2032.
Morris, D. W. 2003. Toward an ecological synthesis: a case for habitat selection. - Oecologia
136: 1–13.
Mu, H., Li, X., Wen, Y., Huang, J., Du, P., Su, W., Miao, S. and Geng, M. 2022. A global record
of annual terrestrial Human Footprint dataset from 2000 to 2018. - Sci Data 9: 176.
Murphy, M. A., Evans, J. S. and Storfer, A. 2010. Quantifying Bufo boreas connectivity in
Yellowstone National Park with landscape genetics. - Ecology 91: 252–261.
Murray, K. A., Arregoitia, L. D. V., Davidson, A., Marco, M. D. and Fonzo, M. M. I. D. 2014.
Threat to the point: improving the value of comparative extinction risk analysis for
conservation action. - Global Change Biology 20: 483–494.
Myers, N., Mittermeier, R. A., Mittermeier, C. G., da Fonseca, G. A. B. and Kent, J. 2000.
Biodiversity hotspots for conservation priorities. - Nature 403: 853–858.
Nania, D., Lumbierres, M., Ficetola, G. F., Falaschi, M., Pacifici, M. and Rondinini, C. 2022.
Maps of area of habitat for Italian amphibians and reptiles. - Nature Conservation 49:
117–129.
Neugarten, R. A., Chaplin-Kramer, R., Sharp, R. P., Schuster, R., Strimas-Mackey, M.,
Roehrdanz, P. R., Mulligan, M., van Soesbergen, A., Hole, D., Kennedy, C. M., Oakleaf,
J. R., Johnson, J. A., Kiesecker, J., Polasky, S., Hanson, J. O. and Rodewald, A. D. 2024.
Mapping the planet’s critical areas for biodiversity and nature’s contributions to people. Nat Commun 15: 261.
Newbold, T., Hudson, L. N., Hill, S. L. L., Contu, S., Lysenko, I., Senior, R. A., Börger, L.,
Bennett, D. J., Choimes, A., Collen, B., Day, J., De Palma, A., Díaz, S., EcheverriaLondoño, S., Edgar, M. J., Feldman, A., Garon, M., Harrison, M. L. K., Alhusseini, T.,
125

Ingram, D. J., Itescu, Y., Kattge, J., Kemp, V., Kirkpatrick, L., Kleyer, M., Correia, D. L.
P., Martin, C. D., Meiri, S., Novosolov, M., Pan, Y., Phillips, H. R. P., Purves, D. W.,
Robinson, A., Simpson, J., Tuck, S. L., Weiher, E., White, H. J., Ewers, R. M., Mace, G.
M., Scharlemann, J. P. W. and Purvis, A. 2015. Global effects of land use on local
terrestrial biodiversity. - Nature 520: 45–50.
Newbold, T., Hudson, L. N., Arnell, A. P., Contu, S., Palma, A. D., Ferrier, S., Hill, S. L. L.,
Hoskins, A. J., Lysenko, I., Phillips, H. R. P., Burton, V. J., Chng, C. W. T., Emerson, S.,
Gao, D., Pask-Hale, G., Hutton, J., Jung, M., Sanchez-Ortiz, K., Simmons, B. I.,
Whitmee, S., Zhang, H., Scharlemann, J. P. W. and Purvis, A. 2016. Has land use pushed
terrestrial biodiversity beyond the planetary boundary? A global assessment. - Science
353: 288–291.
Nitschke, C. R. 2008. The cumulative effects of resource development on biodiversity and
ecological integrity in the Peace-Moberly region of Northeast British Columbia, Canada.
- Biodivers Conserv 17: 1715–1740.
O’Bryan, C. J., Garnett, S. T., Fa, J. E., Leiper, I., Rehbein, J. A., Fernández-Llamazares, Á.,
Jackson, M. V., Jonas, H. D., Brondizio, E. S., Burgess, N. D., Robinson, C. J., Zander,
K. K., Molnár, Z., Venter, O. and Watson, J. E. M. 2021. The importance of Indigenous
Peoples’ lands for the conservation of terrestrial mammals. - Conservation Biology 35:
1002–1008.
Obura, D. 2023. The Kunming-Montreal Global Biodiversity Framework: Business as usual or a
turning point? - One Earth 6: 77–80.
Oliveira-Filho, A. T. and Ratter, J. A. 2002. 6. Vegetation Physiognomies and Woody Flora of
the Cerrado Biome. - In: Oliveira, P. S. and Marquis, R. J. (eds), The Cerrados of Brazil:
Ecology and Natural History of a Neotropical Savanna. Columbia University Press, pp.
91–120.
Orme, C. D. L., Davies, R. G., Burgess, M., Eigenbrod, F., Pickup, N., Olson, V. A., Webster, A.
J., Ding, T.-S., Rasmussen, P. C., Ridgely, R. S., Stattersfield, A. J., Bennett, P. M.,
Blackburn, T. M., Gaston, K. J. and Owens, I. P. F. 2005. Global hotspots of species
richness are not congruent with endemism or threat. - Nature 436: 1016–1019.
Ovaskainen, O. and Hanski, I. 2003. How much does an individual habitat fragment contribute to
metapopulation dynamics and persistence? - Theoretical Population Biology 64: 481–
495.
Pacifici, M., Di Marco, M. and Watson, J. E. M. 2020. Protected areas are now the last
strongholds for many imperiled mammal species. - Conservation Letters 13: e12748.
Paine, R. T. 1969. A Note on Trophic Complexity and Community Stability. - The American
Naturalist 103: 91–93.

126

Pardini, R., Bueno, A. de A., Gardner, T. A., Prado, P. I. and Metzger, J. P. 2010. Beyond the
fragmentation threshold hypothesis: regime shifts in biodiversity across fragmented
landscapes. - PLOS ONE 5: e13666.
Peres, C. A. 2001. Synergistic Effects of Subsistence Hunting and Habitat Fragmentation on
Amazonian Forest Vertebrates. - Conservation Biology 15: 1490–1505.
Perfecto, I. and Vandermeer, J. 2008. Biodiversity conservation in tropical agroecosystems. Annals of the New York Academy of Sciences 1134: 173–200.
Perfecto, I. and Vandermeer, J. 2010. The agroecological matrix as alternative to the landsparing/agriculture intensification model. - Proceedings of the National Academy of
Sciences 107: 5786–5791.
Perrin, A., Rein, F., Christe, P. and Pellet, J. 2025. Habitat fragmentation impact on insect
diversity: opposing forces at patch and landscape levels. - Landsc Ecol 40: 113.
Pfeifer, M., Lefebvre, V., Peres, C. A., Banks-Leite, C., Wearn, O. R., Marsh, C. J., Butchart, S.
H. M., Arroyo-Rodríguez, V., Barlow, J., Cerezo, A., Cisneros, L., D’Cruze, N., Faria,
D., Hadley, A., Harris, S. M., Klingbeil, B. T., Kormann, U., Lens, L., Medina-Rangel,
G. F., Morante-Filho, J. C., Olivier, P., Peters, S. L., Pidgeon, A., Ribeiro, D. B.,
Scherber, C., Schneider-Maunoury, L., Struebig, M., Urbina-Cardona, N., Watling, J. I.,
Willig, M. R., Wood, E. M. and Ewers, R. M. 2017. Creation of forest edges has a global
impact on forest vertebrates. - Nature 551: 187–191.
Phalan, B. T. 2018. What have we learned from the land sparing-sharing model? - Sustainability
10: 1760.
Phalan, B., Onial, M., Balmford, A. and Green, R. E. 2011. Reconciling Food Production and
Biodiversity Conservation: Land Sharing and Land Sparing Compared. - Science 333:
1289–1291.
Pillay, R., Watson, J. E. M., Hansen, A. J., Jantz, P. A., Aragon-Osejo, J., Armenteras, D.,
Atkinson, S. C., Burns, P., Ervin, J., Goetz, S. J., González-del-Pliego, P., Robinson, N.
P., Supples, C., Virnig, A. L. S., Williams, B. A. and Venter, O. 2022. Humid tropical
vertebrates are at lower risk of extinction and population decline in forests with higher
structural integrity. - Nat Ecol Evol 6: 1840–1849.
Pimm, S. L., Jenkins, C. N., Abell, R., Brooks, T. M., Gittleman, J. L., Joppa, L. N., Raven, P.
H., Roberts, C. M. and Sexton, J. O. 2014. The biodiversity of species and their rates of
extinction, distribution, and protection. - Science 344: 1246752.
Pla, M., Hermoso, V., Pino, J. and Brotons, L. 2024. Mapping pressures on biodiversity: the
contribution of thematic detail to decision making. - Biodivers Conserv 33: 1079–1098.
Powers, R. P. and Jetz, W. 2019. Global habitat loss and extinction risk of terrestrial vertebrates
under future land-use-change scenarios. - Nat. Clim. Chang. 9: 323–329.
127

Prevedello, J. A. and Vieira, M. V. 2010. Does the type of matrix matter? A quantitative review
of the evidence. - Biodivers Conserv 19: 1205–1223.
Price, B., McAlpine, C. A., Kutt, A. S., Phinn, S. R., Pullar, D. V. and Ludwig, J. A. 2009.
Continuum or discrete patch landscape models for savanna birds? Towards a pluralistic
approach. - Ecography 32: 745–756.
Protected Planet 2025. Calculating protected area and OECM coverage. - Protected Planet
Prugh, L. R., Hodges, K. E., Sinclair, A. R. E. and Brashares, J. S. 2008. Effect of habitat area
and isolation on fragmented animal populations. - PNAS 105: 20770–20775.
Purvis, A., Gittleman, J. L., Cowlishaw, G. and Mace, G. M. 2000. Predicting extinction risk in
declining species. - Proceedings of the Royal Society of London. Series B: Biological
Sciences 267: 1947–1952.
R Core Team 2021. R: A Language and Environment for Statistical Computing, Version 4.1.0.
R Core Team 2023. R: A Language and Environment for Statistical Computing, Version 4.3.2.
Ramírez-Delgado, J. P., Di Marco, M., Watson, J. E. M., Johnson, C. J., Rondinini, C., Corredor
Llano, X., Arias, M. and Venter, O. 2022. Matrix condition mediates the effects of habitat
fragmentation on species extinction risk. - Nat Commun 13: 595.
Ramírez-Delgado, J. P., Di Marco, M., Johnson, C. J., Watson, J. E. M., Beyer, H., de Assis
Barros, L., Pillay, R. and Venter, O. In press. Global comparison of habitat intactness
models for predicting extinction risk in terrestrial mammals. - Ecography.
Ramírez-Delgado, J. P., Di Marco, M., Johnson, C. J., Watson, J. E. M. and Venter, O. In
preparation. Mapping the conservation potential of the matrix for terrestrial mammals.
Reider, I. J., Donnelly, M. A. and Watling, J. I. 2018. The influence of matrix quality on species
richness in remnant forest. - Landscape Ecol 33: 1147–1157.
Rey Benayas, J. M. and Bullock, J. M. 2012. Restoration of Biodiversity and Ecosystem Services
on Agricultural Land. - Ecosystems 15: 883–899.
Ricketts, T. H. 2001. The Matrix Matters: Effective Isolation in Fragmented Landscapes. - The
American Naturalist 158: 87–99.
Ries, L., Jr, R. J. F., Battin, J. and Sisk, T. D. 2004. Ecological Responses to Habitat Edges:
Mechanisms, Models, and Variability Explained. - Annual Review of Ecology,
Evolution, and Systematics 35: 491–522.
Ripple, W. J., Bradshaw, G. A. and Spies, T. A. 1991. Measuring forest landscape patterns in the
cascade range of Oregon, USA. - Biological Conservation 57: 73–88.

128

Ripple, W. J., Estes, J. A., Beschta, R. L., Wilmers, C. C., Ritchie, E. G., Hebblewhite, M.,
Berger, J., Elmhagen, B., Letnic, M., Nelson, M. P., Schmitz, O. J., Smith, D. W.,
Wallach, A. D. and Wirsing, A. J. 2014. Status and Ecological Effects of the World’s
Largest Carnivores. - Science 343: 1241484.
Ripple, W. J., Abernethy, K., Betts, M. G., Chapron, G., Dirzo, R., Galetti, M., Levi, T., Lindsey,
P. A., Macdonald, D. W., Machovina, B., Newsome, T. M., Peres, C. A., Wallach, A. D.,
Wolf, C. and Young, H. 2016. Bushmeat hunting and extinction risk to the world’s
mammals. - Royal Society Open Science 3: 160498.
Ripple, W. J., Wolf, C., Newsome, T. M., Hoffmann, M., Wirsing, A. J. and McCauley, D. J.
2017. Extinction risk is most acute for the world’s largest and smallest vertebrates. Proceedings of the National Academy of Sciences 114: 10678–10683.
Riva, F. and Fahrig, L. 2023. Landscape-scale habitat fragmentation is positively related to
biodiversity, despite patch-scale ecosystem decay. - Ecology Letters 26: 268–277.
Rodrigues, A. S. L., Andelman, S. J., Bakarr, M. I., Boitani, L., Brooks, T. M., Cowling, R. M.,
Fishpool, L. D. C., da Fonseca, G. A. B., Gaston, K. J., Hoffmann, M., Long, J. S.,
Marquet, P. A., Pilgrim, J. D., Pressey, R. L., Schipper, J., Sechrest, W., Stuart, S. N.,
Underhill, L. G., Waller, R. W., Watts, M. E. J. and Yan, X. 2004. Effectiveness of the
global protected area network in representing species diversity. - Nature 428: 640–643.
Rondinini, C., Wilson, K. A., Boitani, L., Grantham, H. and Possingham, H. P. 2006. Tradeoffs
of different types of species occurrence data for use in systematic conservation planning.
- Ecology Letters 9: 1136–1145.
Rondinini, C., Di Marco, M., Chiozza, F., Santulli, G., Baisero, D., Visconti, P., Hoffmann, M.,
Schipper, J., Stuart, S. N., Tognelli, M. F., Amori, G., Falcucci, A., Maiorano, L. and
Boitani, L. 2011. Global habitat suitability models of terrestrial mammals. - Philosophical
Transactions of the Royal Society B: Biological Sciences 366: 2633–2641.
Rondinini, C., Marco, M. D., Visconti, P., Butchart, S. H. M. and Boitani, L. 2014. Update or
outdate: long-term viability of the IUCN Red List. - Conservation Letters 7: 126–130.
Rouget, M. 2003. Measuring conservation value at fine and broad scales: implications for a
diverse and fragmented region, the Agulhas Plain. - Biological Conservation 112: 217–
232.
Ruffell, J., Banks‐Leite, C. and Didham, R. K. 2016. Accounting for the causal basis of
collinearity when measuring the effects of habitat loss versus habitat fragmentation. Oikos 125: 117–125.
Ruffell, J., Clout, M. N. and Didham, R. K. 2017. The matrix matters, but how should we
manage it? Estimating the amount of high-quality matrix required to maintain
biodiversity in fragmented landscapes. - Ecography 40: 171–178.

129

Runge, C. A., Watson, J. E. M., Butchart, S. H. M., Hanson, J. O., Possingham, H. P. and Fuller,
R. A. 2015. Protected areas and global conservation of migratory birds. - Science 350:
1255–1258.
Sandbrook, C., Albury-Smith, S., Allan, J. R., Bhola, N., Bingham, H. C., Brockington, D.,
Byaruhanga, A. B., Fajardo, J., Fitzsimons, J., Franks, P., Fleischman, F., Frechette, A.,
Kakuyo, K., Kaptoyo, E., Kuemmerle, T., Kalunda, P. N., Nuvunga, M., O’Donnell, B.,
Onyai, F., Pfeifer, M., Pritchard, R., Ramos, A., Rao, M., Ryan, C. M., Shyamsundar, P.,
Tauli, J., Tumusiime, D. M., Vilaça, M., Watmough, G. R., Worsdell, T. and Zaehringer,
J. G. 2023. Social considerations are crucial to success in implementing the 30×30 global
conservation target. - Nat Ecol Evol 7: 784–785.
Saura, S. and Rubio, L. 2010. A common currency for the different ways in which patches and
links can contribute to habitat availability and connectivity in the landscape. - Ecography
33: 523–537.
Scheele, B. C., Pasmans, F., Skerratt, L. F., Berger, L., Martel, A., Beukema, W., Acevedo, A.
A., Burrowes, P. A., Carvalho, T., Catenazzi, A., De la Riva, I., Fisher, M. C., Flechas, S.
V., Foster, C. N., Frías-Álvarez, P., Garner, T. W. J., Gratwicke, B., Guayasamin, J. M.,
Hirschfeld, M., Kolby, J. E., Kosch, T. A., La Marca, E., Lindenmayer, D. B., Lips, K.
R., Longo, A. V., Maneyro, R., McDonald, C. A., Mendelson, J., Palacios-Rodriguez, P.,
Parra-Olea, G., Richards-Zawacki, C. L., Rödel, M.-O., Rovito, S. M., Soto-Azat, C.,
Toledo, L. F., Voyles, J., Weldon, C., Whitfield, S. M., Wilkinson, M., Zamudio, K. R.
and Canessa, S. 2019. Amphibian fungal panzootic causes catastrophic and ongoing loss
of biodiversity. - Science 363: 1459–1463.
Schipper, J., Chanson, J. S., Chiozza, F., Cox, N. A., Young, B. E., et al. 2008. The status of the
world’s land and marine mammals: diversity, threat, and knowledge. - Science 322: 225–
230.
Schleicher, J., Zaehringer, J. G., Fastré, C., Vira, B., Visconti, P. and Sandbrook, C. 2019.
Protecting half of the planet could directly affect over one billion people. - Nat Sustain 2:
1094–1096.
Schmiegelow, F. K. A., Machtans, C. S. and Hannon, S. J. 1997. Are Boreal Birds Resilient to
Forest Fragmentation? An Experimental Study of Short-Term Community Responses. Ecology 78: 1914–1932.
Schmitz, O. J., Hambäck, P. A. and Beckerman, A. P. 2000. Trophic Cascades in Terrestrial
Systems: A Review of the Effects of Carnivore Removals on Plants. - The American
Naturalist 155: 141–153.
Segan, D. B., Murray, K. A. and Watson, J. E. M. 2016. A global assessment of current and
future biodiversity vulnerability to habitat loss–climate change interactions. - Global
Ecology and Conservation 5: 12–21.

130

Smith, F. A., Lyons, S. K., Ernest, S. K. M., Jones, K. E., Kaufman, D. M., Dayan, T., Marquet,
P. A., Brown, J. H. and Haskell, J. P. 2003. Body mass of late Quaternary mammals. Ecology 84: 3403–3403.
Soberón, J. 2007. Grinnellian and Eltonian niches and geographic distributions of species. Ecology Letters 10: 1115–1123.
Sodhi, N. S., Koh, L. P., Brook, B. W. and Ng, P. K. L. 2004. Southeast Asian biodiversity: an
impending disaster. - Trends in Ecology & Evolution 19: 654–660.
Soria, C. D., Pacifici, M., Di Marco, M., Stephen, S. M. and Rondinini, C. 2021. COMBINE: a
coalesced mammal database of intrinsic and extrinsic traits. - Ecology 102: e03344.
Stevens, R. D., Rowe, R. J. and Badgley, C. 2019. Gradients of mammalian biodiversity through
space and time. - Journal of Mammalogy 100: 1069–1086.
Suding, K., Higgs, E., Palmer, M., Callicott, J. B., Anderson, C. B., Baker, M., Gutrich, J. J.,
Hondula, K. L., LaFevor, M. C., Larson, B. M. H., Randall, A., Ruhl, J. B. and Schwartz,
K. Z. S. 2015. Committing to ecological restoration. - Science 348: 638–640.
Swets, J. A. 1988. Measuring the Accuracy of Diagnostic Systems. - Science 240: 1285–1293.
Swets, J. A. 2014. Signal Detection Theory and ROC Analysis in Psychology and Diagnostics:
Collected Papers. - Psychology Press.
Tacutu, R., Craig, T., Budovsky, A., Wuttke, D., Lehmann, G., Taranukha, D., Costa, J.,
Fraifeld, V. E. and de Magalhães, J. P. 2013. Human ageing genomic resources:
integrated databases and tools for the biology and genetics of ageing. - Nucleic Acids
Research 41: D1027–D1033.
Tanhuanpää, T., Mikkonen, N., Kujala, H., Heinaro, E., Mäyrä, J. and Kumpula, T. 2023. Input
data resolution affects the conservation prioritization outcome of spatially sparse
biodiversity features. - Ambio 52: 1793–1803.
Terborgh, J., Lopez, L., Nuñez, P., Rao, M., Shahabuddin, G., Orihuela, G., Riveros, M.,
Ascanio, R., Adler, G. H., Lambert, T. D. and Balbas, L. 2001. Ecological Meltdown in
Predator-Free Forest Fragments. - Science 294: 1923–1926.
Thatte, P., Chandramouli, A., Tyagi, A., Patel, K., Baro, P., Chhattani, H. and Ramakrishnan, U.
2020. Human footprint differentially impacts genetic connectivity of four wide-ranging
mammals in a fragmented landscape. - Diversity and Distributions 26: 299–314.
Torchiano, M. 2017. effsize: Efficient Effect Size Computation. R package version 0.7.1. in
press.
Torres-Romero, E. J., Giordano, A. J., Ceballos, G. and López-Bao, J. V. 2020. Reducing the
sixth mass extinction: Understanding the value of human-altered landscapes to the

131

conservation of the world’s largest terrestrial mammals. - Biological Conservation 249:
108706.
Tracewski, Ł., Butchart, S. H. M., Marco, M. D., Ficetola, G. F., Rondinini, C., Symes, A.,
Wheatley, H., Beresford, A. E. and Buchanan, G. M. 2016. Toward quantification of the
impact of 21st-century deforestation on the extinction risk of terrestrial vertebrates. Conservation Biology 30: 1070–1079.
Tscharntke, T., Klein, A. M., Kruess, A., Steffan-Dewenter, I. and Thies, C. 2005. Landscape
perspectives on agricultural intensification and biodiversity – ecosystem service
management. - Ecology Letters 8: 857–874.
Tucker, M. A., Ord, T. J. and Rogers, T. L. 2014. Evolutionary predictors of mammalian home
range size: body mass, diet and the environment. - Global Ecology and Biogeography 23:
1105–1114.
Tucker, M. A., Böhning-Gaese, K., Fagan, W. F., Fryxell, J. M., Mueller, T., et al. 2018. Moving
in the Anthropocene: global reductions in terrestrial mammalian movements. - Science
359: 466–469.
Tucker, M. A., Santini, L., Carbone, C. and Mueller, T. 2021. Mammal population densities at a
global scale are higher in human-modified areas. - Ecography 44: 1–13.
UNEP-WCMC and IUCN 2023. Protected Planet: The World Database on Protected Areas
(WDPA) and World Database on Other Effective Area-based Conservation Measures
(WD-OECM).
UNEP-WCMC and IUCN 2025. Protected Planet: The World Database on Protected Areas
(WDPA) and World Database on Other Effective Area-based Conservation Measures
(WD-OECM).
Urban, M. C. 2015. Accelerating extinction risk from climate change. - Science 348: 571–573.
Venter, O., Sanderson, E. W., Magrach, A., Allan, J. R., Beher, J., Jones, K. R., Possingham, H.
P., Laurance, W. F., Wood, P., Fekete, B. M., Levy, M. A. and Watson, J. E. M. 2016.
Sixteen years of change in the global terrestrial human footprint and implications for
biodiversity conservation. - Nature Communications 7: 12558.
Verboom, J., Snep, R. P. H., Stouten, J., Pouwels, R., Pe’er, G., Goedhart, P. W., Adrichem, M.
H. C. van, Alkemade, J. R. M. and Jones-Walters, L. M. 2014. Using Minimum Area
Requirements (MAR) for assemblages of mammal and bird species in global biodiversity
assessments. in press.
Verde Arregoitia, L. D., Blomberg, S. P. and Fisher, D. O. 2013. Phylogenetic correlates of
extinction risk in mammals: species in older lineages are not at greater risk. - Proceedings
of the Royal Society B: Biological Sciences 280: 20131092.

132

Villard, M.-A. and Metzger, J. P. 2014. Beyond the fragmentation debate: a conceptual model to
predict when habitat configuration really matters. - Journal of Applied Ecology 51: 309–
318.
Visconti, P., Bakkenes, M., Baisero, D., Brooks, T., Butchart, S. H. M., Joppa, L., Alkemade, R.,
Di Marco, M., Santini, L., Hoffmann, M., Maiorano, L., Pressey, R. L., Arponen, A.,
Boitani, L., Reside, A. E., van Vuuren, D. P. and Rondinini, C. 2016. Projecting Global
Biodiversity Indicators under Future Development Scenarios. - Conservation Letters 9:
5–13.
Viveiros de Castro, E. B. and Fernandez, F. A. S. 2004. Determinants of differential extinction
vulnerabilities of small mammals in Atlantic forest fragments in Brazil. - Biological
Conservation 119: 73–80.
Ward, M., Saura, S., Williams, B., Ramírez-Delgado, J. P., Arafeh-Dalmau, N., Allan, J. R.,
Venter, O., Dubois, G. and Watson, J. E. M. 2020. Just ten percent of the global
terrestrial protected area network is structurally connected via intact land. - Nat Commun
11: 4563.
Watling, J. I., Nowakowski, A. J., Donnelly, M. A. and Orrock, J. L. 2011. Meta-analysis reveals
the importance of matrix composition for animals in fragmented habitat. - Global
Ecology and Biogeography 20: 209–217.
Watson, D. M. 2002. A conceptual framework for studying species composition in fragments,
islands and other patchy ecosystems. - Journal of Biogeography 29: 823–834.
Watson, J. E. M. and Venter, O. 2019. Mapping the Continuum of Humanity’s Footprint on
Land. - One Earth 1: 175–180.
Watson, J. E. M., Whittaker, R. J. and Freudenberger, D. 2005. Bird community responses to
habitat fragmentation: how consistent are they across landscapes? - Journal of
Biogeography 32: 1353–1370.
Watson, J. E. M., Dudley, N., Segan, D. B. and Hockings, M. 2014. The performance and
potential of protected areas. - Nature 515: 67–73.
Watson, J. E. M., Shanahan, D. F., Di Marco, M., Allan, J., Laurance, W. F., Sanderson, E. W.,
Mackey, B. and Venter, O. 2016. Catastrophic Declines in Wilderness Areas Undermine
Global Environment Targets. - Current Biology 26: 2929–2934.
Watson, J. E. M., Evans, T., Venter, O., Williams, B., Tulloch, A., Stewart, C., Thompson, I.,
Ray, J. C., Murray, K., Salazar, A., McAlpine, C., Potapov, P., Walston, J., Robinson, J.
G., Painter, M., Wilkie, D., Filardi, C., Laurance, W. F., Houghton, R. A., Maxwell, S.,
Grantham, H., Samper, C., Wang, S., Laestadius, L., Runting, R. K., Silva-Chávez, G. A.,
Ervin, J. and Lindenmayer, D. 2018a. The exceptional value of intact forest ecosystems. Nat Ecol Evol 2: 599–610.

133

Watson, J. E. M., Venter, O., Lee, J., Jones, K. R., Robinson, J. G., Possingham, H. P. and Allan,
J. R. 2018b. Protect the last of the wild. - Nature 563: 27–30.
Williams, B. A., Venter, O., Allan, J. R., Atkinson, S. C., Rehbein, J. A., Ward, M., Di Marco,
M., Grantham, H. S., Ervin, J., Goetz, S. J., Hansen, A. J., Jantz, P., Pillay, R.,
Rodríguez-Buriticá, S., Supples, C., Virnig, A. L. S. and Watson, J. E. M. 2020. Change
in terrestrial human footprint drives continued loss of intact ecosystems. - One Earth 3:
371–382.
Williams, D. R., Clark, M., Buchanan, G. M., Ficetola, G. F., Rondinini, C. and Tilman, D. 2021.
Proactive conservation to prevent habitat losses to agricultural expansion. - Nat Sustain 4:
314–322.
Wilman, H., Belmaker, J., Simpson, J., Rosa, C. de la, Rivadeneira, M. M. and Jetz, W. 2014.
EltonTraits 1.0: species-level foraging attributes of the world’s birds and mammals. Ecology 95: 2027–2027.
Wilson, M. C., Chen, X.-Y., Corlett, R. T., Didham, R. K., Ding, P., Holt, R. D., Holyoak, M.,
Hu, G., Hughes, A. C., Jiang, L., Laurance, W. F., Liu, J., Pimm, S. L., Robinson, S. K.,
Russo, S. E., Si, X., Wilcove, D. S., Wu, J. and Yu, M. 2016. Habitat fragmentation and
biodiversity conservation: key findings and future challenges. - Landscape Ecol 31: 219–
227.
Woinarski, J. C. Z., Burbidge, A. A. and Harrison, P. L. 2015. Ongoing unraveling of a
continental fauna: Decline and extinction of Australian mammals since European
settlement. - PNAS 112: 4531–4540.
Wolff, N. H., Visconti, P., Kujala, H., Santini, L., Hilbers, J. P., Possingham, H. P., Oakleaf, J.
R., Kennedy, C. M., Kiesecker, J., Fargione, J. and Game, E. T. 2023. Prioritizing global
land protection for population persistence can double the efficiency of habitat protection
for reducing mammal extinction risk. - One Earth 6: 1564–1575.
WWF 2024. Living Planet Report 2024 -A System in Peril.
Zhang, C., Wang, Y., Chang, J., Li, J., Pan, S., Yang, B., Zhan, X. and Dai, Q. 2024. Global
patterns of human–wildlife spatial associations and implications for differentiating
conservation strategies. - Conservation Biology 38: e14279.

134

A. Appendix A: Supplementary material for “Chapter 2: Matrix condition mediates
the effects of habitat fragmentation on species extinction risk”

Figure A.1: Transition matrix of extinction risk categories for terrestrial mammals between
1996 and 2020. The plot shows the transition matrix of the first and last Red List category
reported between 1996 and 2020. The colour scheme represents the relative frequency of
individual species in each transition. Acronyms refer to the IUCN Red List categories, including
Least Concern (LC), Near Threatened (NT), Vulnerable (VU), Endangered (EN), and Critically
Endangered (CR). The relative frequency was arcsine square-root-transformed for visual
purposes and to avoid variance instability when handling proportions close to zero.

135

Figure A.2: Sensitivity analysis of selected variables for the prediction of extinction risk
transitions in terrestrial mammals. Relative importance of each predictor when the extent of
suitable habitat is represented by high and medium habitat suitability combined, and the extent of
the matrix by unsuitable habitat alone (a). Relative importance of each predictor when the extent
of suitable habitat is represented by high habitat suitability, and the extent of the matrix by
medium habitat suitability and unsuitable habitat combined (b). Variables are colour-coded
according to their broad class (human pressure, environment, and life-history). The description of

136

each variable is given in Table 2.1 of the main manuscript. High levels of the human footprint
(HFP) included values of 3 or above.

Table A.1: Cross-validation results of the Random Forest models for the prediction of
extinction risk transitions in terrestrial mammals. *Cross-validation results when the extent
of suitable habitat is represented by high and medium habitat suitability combined, and the extent
of the matrix by unsuitable habitat alone. †Cross-validation results when the extent of suitable
habitat is represented by high habitat suitability, and the extent of the matrix by medium habitat
suitability and unsuitable habitat combined. The cross-validation was performed in terms of
proportion of correctly classified species (accuracy), proportion of correctly classified high-risk
species (sensitivity), proportion of correctly classified low-risk species (specificity), and the true
skill statistic (TSS = sensitivity + specificity – 1).
Model

Accuracy (%)

Sensitivity (%)

Specificity (%)

TSS

81.2

60.5

90.9

0.51

82.0

61.1

91.7

0.53

Extinction risk transitions ~ High and
medium suitability combined*
Extinction risk transitions ~ Medium
suitability and unsuitable combined†

137

Figure A.3: Distribution of the matrix condition in low-risk and high-risk species at a
global scale. The plot represents the frequency distribution of the extent of high human footprint
values within the matrix of 4,329 species (68% low-risk species and 32% high-risk species).
Species with ranges that did not overlap with the human footprint (2% of 4,426 terrestrial
mammals in our sample) were excluded from the calculations. Blue and red lines refer to lowrisk and high-risk species, as reported in legend. Dashed lines denote mean of distributions. High
levels of the human footprint (HFP) included values of 3 or above.

138

Figure A.4: Distribution of the matrix condition in low-risk and high-risk species at the
scale of individual biogeographic realms. The plots represent the frequency distribution of the
extent of high human footprint values within the matrix of 1.109 species in the Neotropical realm
(68% low-risk species and 32% high-risk species), 990 species in the Afrotropical realm (65%
low-risk species and 35% high-risk species), 596 species in the Indo-Malay realm (52% low-risk
species and 48% high-risk species), 552 species in the Australasia realm (58% low-risk species
and 42% high-risk species), 498 species in the Palearctic realm (79% low-risk species and 21%
high-risk species), and 286 species in the Nearctic realm (83% low-risk species and 17% highrisk species). The Oceania realm included a total of 4 species only (17% low-risk species and
83% high-risk species), which are not depicted on the map. Species with ranges that overlapped
with two or more realms (9% of 4,426 terrestrial mammals in our sample) were excluded from
the calculations. Blue and red lines refer to low-risk and high-risk species, as reported in legend.

139

Dashed lines denote mean of distributions. High levels of the human footprint (HFP) included
values of 3 or above.

Figure A.5: Effect size (Cohen’s d statistic) of the degree of habitat fragmentation and the
degree of patch isolation between low-risk and high-risk species with a low-quality matrix
and a high-quality matrix. Effect size of the degree of fragmentation and the degree of patch
isolation between low-risk and high-risk species with a matrix of low-quality habitat (a). Effect
size of the degree of fragmentation and the degree of patch isolation between low-risk and highrisk species with a matrix of high-quality habitat (b). Green and yellow points represent the
effect sizes or the standardized difference of mean values (Cohen’s d statistic) of the degree of
140

fragmentation and the degree of patch isolation between low-risk and high-risk species,
respectively. Error bars represent 95% confidence intervals for each effect size calculated. Lowquality matrices included species with proportions > 84.2% of the extent of their matrix
overlapping with high human footprint values (n = 1,815 low-risk species and 1,027 high-risk
species). High-quality matrices included species with proportions < 15.8% of the extent of their
matrix overlapping with high human footprint values (n = 60 low-risk species and 29 high-risk
species). High levels of the human footprint included values of 3 or above.

141

B. Appendix B: Supplementary material for “Chapter 3: Global comparison of habitat
intactness models for predicting extinction risk for terrestrial mammals”

Supplementary methods B.1: Transformation of the human footprint (HFP) to habitat
quality. Following previous work (Beyer et al. 2020), we transformed the HFP maps to quality
using an exponential function (𝑤𝑖 = 𝑒𝑥𝑝(−𝛾𝐻𝐹𝑃𝑖 )), such that habitat quality is 1 when the HFP
is 0, and habitat quality is 20% when the HFP is 4. We applied this transformation based on the
assumptions listed below: (i) A HFP value of 0 represents areas free of any human influence (i.e.
the remaining wilderness areas on land) and the highest habitat quality; (ii) A HFP value of 4
represents areas of low human pressure, such as pasture lands with other human disturbances
nearby (e.g., roads), and habitat quality should have declined to lower levels at this point; and
(iii) HFP values above 8 represent areas of high human pressure levels, such as intensive
agriculture and urbanized areas, and habitat quality should be approaching values of 0 at this
point. We achieved these criteria using a value of 𝛾 = 0.4023595.

142

Figure B.1: Effects of the 5th, 10th, 50th, 90th, and 95th percentiles of the continuum and
hybrid habitat intactness models on extinction risk for terrestrial mammals classified as
habitat generalists and specialists. Each variable is colour-coded according to the habitat
model (continuum or hybrid model) it represents. Effect sizes were estimated while holding other
variables known to influence extinction risk (see Table 3.1) at their mean. All variables were
standardized (z-transformed) before analysis to ensure comparable results. Error bars represent
95% confidence intervals. Confidence intervals not containing zero (dashed green line)
indicate statistical significance.

143

Figure B.2: Correlation between variables used to predict extinction risk for terrestrial
mammals classified as habitat generalists and habitat specialists. Pearson’s correlation
coefficient (r) was used to quantify relationships between variables. Acronyms: patch-matrix
model (PMM), reduction in the patch-matrix model (RPMM), continuum model (CM), reduction
in the continuum model (RCM), hybrid model (HM), reduction in the hybrid model (RHM),
range size (RS), gestation length (GL), and weaning age (WA). Detailed descriptions of these
variables are provided in Table 3.1.

144

Figure B.3: Predictive performance of extinction risk models for both habitat generalists
and habitat specialists as a function of three habitat intactness models and how they
decrease over time. AUC values were estimated from mixed-effects logistic regression models
using a spatially blocked cross-validation and a ten-fold cross-validation scheme (see Methods).
AUCs above 0.5 (dashed green line) indicate that model performance is better than random
chance. All extinction risk models accounted for the effects of other variables known to
influence extinction risk (see Table 1). Error bars represent 95% confidence intervals. Acronyms:
extinction risk model (ER), patch-matrix model (PMM), continuum model (CM), hybrid model
(HM), reduction in the patch-matrix model (RPMM), reduction in the continuum model (RCM),
and reduction in the hybrid model (RHM).

145

Figure B.4: Predictive performance of extinction risk models for both habitat generalists
and habitat specialists as a function of three habitat intactness models and their temporal
decline, before and after excluding threatened species assessed solely under IUCN Red List
Criterion B. AUC values were estimated from mixed-effects logistic regression models using a
spatially blocked cross-validation scheme (see Methods). AUCs above 0.5 (dashed green line)
indicate that model performance is better than random chance. All extinction risk models
accounted for the effects of other variables known to influence extinction risk (see Table 1).
Error bars represent 95% confidence intervals. Acronyms: extinction risk model (ER), patchmatrix model (PMM), continuum model (CM), hybrid model (HM), reduction in the patchmatrix model (RPMM), reduction in the continuum model (RCM), and reduction in the hybrid
model (RHM).

146

Figure B.5: Predictive performance of extinction risk models for both habitat generalists
and habitat specialists as a function of three habitat intactness models and their decline
over time, before and after excluding threatened species assessed solely under IUCN Red
List Criterion B. AUC values were estimated from mixed-effects logistic regression models
using a ten-fold cross-validation scheme (see Methods). AUCs above 0.5 (dashed green line)
indicate that model performance is better than random chance. All extinction risk models
accounted for the effects of other variables known to influence extinction risk (see Table 1).
Error bars represent 95% confidence intervals. Acronyms: extinction risk model (ER), patchmatrix model (PMM), continuum model (CM), hybrid model (HM), reduction in the patchmatrix model (RPMM), reduction in the continuum model (RCM), and reduction in the hybrid
model (RHM).

147

Figure B.6: Effects of three habitat intactness models and their decline over time on
extinction risk for terrestrial mammals classified as habitat generalists and habitat
specialists, before and after excluding threatened species assessed solely under IUCN Red
List Criterion B. Effect sizes were estimated while holding other variables known to influence
extinction risk (see Table 3.1) at their mean. All variables were standardized (z-transformed)
before analysis to ensure comparable results. Error bars represent 95% confidence intervals. The
statistical significance of the coefficients is indicated by the absence of overlap between their
148

confidence intervals and zero (dashed green line). Acronyms: patch-matrix model (PMM),
continuum model (CM), hybrid model (HM), reduction in the patch-matrix model (RPMM),
reduction in the continuum model (RCM), and reduction in the hybrid model (RHM).

149

C. Appendix C: Supplementary material for “Chapter 4: Mapping the conservation
potential of matrix habitats for terrestrial mammals”

Figure C.1: Species richness in matrix areas for terrestrial mammals. Maps show species
richness within matrix areas for all species (a), species with declining population trends (b), and
150

species with threatened status (c). The dataset includes 5,439 species in the all-mammal group,
1,765 in the declining population trend group, and 1,212 in the threatened status group.

Figure C.2: Species richness in suitable habitats for terrestrial mammals. The map shows
species richness within suitable habitats for the all-mammal group.

Table C.1: Geographic coverage and species representation of hotspot types under
different hotspot definitions. The table includes the total area and number of species whose
matrix area overlap with each hotspot type, based on two different thresholds: the top 2.5% and
the top 10% of pixels with the highest richness values. Numbers in parentheses indicate the
percentage of species-rich matrix areas (out of 15,922,965 km² globally) or the percentage of
extant terrestrial mammal species (out of 5,808 species worldwide).
2.5% criterion
Hotspot type

Area (km²)

10% criterion

Number of

Area (km²)

species
All species

393,255 (2.5)

1,431 (24.6) 1,613,280 (10.1)

Number of
species
2,410 (41.5)

151

Declining trend

358,140 (2.2)

2,529 (43.5) 1,728,984 (10.9)

3,175 (54.7)

Threatened status

396,109 (2.5)

2,906 (50.0)

1,381,567 (8.7)

3,631 (62.5)

864,005 (5.4)

3,078 (53.0) 2,852,778 (17.9)

3,792 (65.3)

Total in all
hotspots

Figure C.3: Overlap of hotspot types and conservation potential in matrix areas across
different hotspot definitions. Venn diagrams illustrate the spatial overlap of hotspot types (a, c)

152

and species richness overlap across these hotspot types (b, d) under two different thresholds: the
top 2.5% and the top 10% of pixels with the highest richness values for each species group (all
mammals, mammals with declining population trends, and mammals with threatened status).

Figure C.4: Conservation status and human footprint within matrix areas with
conservation potential. The figure shows the proportion of matrix hotspots classified as high-,
moderate-, and low-priority that fall within protected areas (PAs) (a), other effective area-based
153

conservation measures (OECMs) (b), and the combined coverage of PAs and OECMs (c). Bars
are categorized by human footprint level: intact, low-disturbance, and highly modified.

154