WHITE SPRUCE GROWTH SENSITIVITY TO CLIMATE VARIABILITY
IN PURE AND MIXEDWOOD STANDS

by

Jéssica Chaves Cardoso
B.Sc., Federal Rural University of Rio de Janeiro, 2017

THESIS SUBMITTED IN PARTIAL FULFILLMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
IN
NATURAL RESOURCES AND ENVIRONMENTAL STUDIES

UNIVERSITY OF NORTHERN BRITISH COLUMBIA
April 2020
© Jéssica Chaves Cardoso, 2020

TABLE OF CONTENTS
Abstract ................................................................................................................................. IV
List of Figures: ........................................................................................................................ V
List of Tables:..................................................................................................................... VIII
1. Chapter 1: Introduction .................................................................................................. 1
1.1.
Tree and forest responses to climate .............................................................................. 1
1.2.
White spruce growth and climate sensitivity ................................................................. 6
1.3.
Thesis objectives ............................................................................................................ 9
2. Chapter 2: White spruce growth sensitivity to climate variability in pure and
mixedwood stands .................................................................................................................. 11
2.1.
Introduction.................................................................................................................. 12
2.2.
Materials and methods ................................................................................................. 16
2.2.1.
Study area description.............................................................................................. 16
2.2.2.
Dendrochronology ................................................................................................... 20
2.2.3.
Microclimate data collection ................................................................................... 21
2.2.4.
Statistical analysis .................................................................................................... 21
2.3.
Results ......................................................................................................................... 26
2.3.1.
Identification of important predictor variables for annual white spruce growth ..... 26
2.3.2.
Annual climate and tree growth relationships using full data set ............................ 28
2.3.3.
Climate and tree growth relationships across the treatments................................... 31
2.3.3.1. Annual climate and tree growth relationships across the treatments ....................... 31
2.3.3.2. Seasonal climate and tree growth relationships across the treatments .................... 35
2.3.3.3. Monthly climate and tree growth relationships across the treatments ..................... 38
2.4.
Discussion .................................................................................................................... 41
2.5.
Conclusion ................................................................................................................... 46
3. Chapter 3: White spruce sap flow sensitivity to climate variability in pure and
mixedwood stands .................................................................................................................. 47
3.1.
Introduction.................................................................................................................. 47
3.2.
Materials and methods ................................................................................................. 52
3.2.1.
Study area description.............................................................................................. 52
3.2.2.
Microclimate data collection ................................................................................... 52
3.2.3.
Sap flow measurements ........................................................................................... 53
3.2.4.
Statistical analysis .................................................................................................... 55
3.3.
Results ......................................................................................................................... 59
3.3.1.
Microclimate and sap flow comparison between the treatments ............................. 59
3.3.2.
Model selection of daily sap flow throughout the growing season ......................... 63
3.3.3.
Model selection of daily sap flow by season ........................................................... 65
3.3.4.
Mixed effect models of sap flow ............................................................................. 67
3.3.4.1. Climate and sap flow relationships throughout the growing season ....................... 67
3.3.4.2. Climate and sap flow relationships by season ......................................................... 69
3.4.
Discussion .................................................................................................................... 72
3.5.
Conclusion ................................................................................................................... 75
4.

Chapter 4: Conclusions ................................................................................................. 77
II

4.1.
4.2.
4.3.

Main findings and contributions .................................................................................. 78
Future research............................................................................................................. 84
Concluding remarks ..................................................................................................... 85

5.

Bibliography ................................................................................................................... 86

6.

Appendix 1: Study area ............................................................................................... 101

7. Appendix 2: Results for Chapter 2 ............................................................................ 107
7.1.
Random Forest (Liaw and Wiener 2002) results ....................................................... 107
7.2.
Mixed effect models: Full data – Annual .................................................................. 115
7.3.
Mixed effect models: Across the treatments – Annual .............................................. 120
7.4.
Mixed effect models: Across the treatments – Seasonal ........................................... 130
7.5.
Mixed effect models: Across the treatments – Monthly ............................................ 133
7.6.
Microclimate comparison between pure spruce and mixedwood treatments by month
using daily data from 1999 to 2017 ....................................................................................... 136
8. Appendix 3: Results for Chapter 3 ............................................................................ 138
8.1.
Microclimate comparison between pure spruce and mixedwood treatments by month
using daily data from 2007 to 2018 ....................................................................................... 138
8.2.
Mixed effect models: throughout the growing season ............................................... 141
8.3.
Mixed effect models: between seasons ..................................................................... 145
9.

Appendix 4: Metadata for Chapter 2 and Chapter 3 ............................................... 153

III

Abstract
It is prudent to understand how tree growth responds to climate variability to better
project their growth in the current and future changes in climate in boreal forests. In this
thesis, I studied how climate variables influence individual white spruce trees (Picea glauca
(Moench) Voss) over short and intermediate periods in pure and mixedwood stands in northeastern British Columbia. In Chapter 2, I studied the importance and the influence of annual,
seasonal, and monthly microclimate variables on the annual growth of white spruce trees in
pure and mixedwood stands. In Chapter 3, I studied the importance and the influence of
microclimate variables on sap flow of white spruce trees through different time scales in
these two stand types. My key finding in these two chapters is that stand composition and
structure are essential determinants of how spruce radial growth and sap flow respond to
fluctuations in climate variables, and how they will respond to projected future climate
scenarios. A combination of warmer temperatures and drought during summer will
negatively affect white spruce trees growth in pure and mixedwood stands in the studied
region. Spruce sap flow in both stand types is likely to increase as the climate warms,
increasing the demand for soil water. As this resource becomes less available, white spruce in
both stand types are likely to respond with processes that can compromise their physiological
integrity. White spruce growing in mixedwood stands might be more sensitive to drought
stress than in pure stands due to the higher competition for limiting resources (primarily
water). This thesis provides information of expected changes in tree growth to climate
variability and demonstrates the importance of appropriate site selection to plant spruce trees
and management of pure and mixedwood stands.

IV

List of Figures:
Figure 2.1 Study area location with Prince George and Fort St. John as references. Map
created using ESRI ArcGIS software. ............................................................................. 16
Figure 3.1 Sap flow velocity of tree T5 located in the pure spruce treatment, and tree T1
located in the mixedwood treatment over the 2007 growing season. Vertical lines
separate the seasons used in the study: Spring (April and May), early-summer (June and
July), and late-summer (August and September). ........................................................... 58
Figure 3.2 Sap flow velocity of tree T5 located in the pure spruce treatment (PS) with soil
water potential (SWP) in the PS, and tree T1 located in the mixedwood treatment (MW)
with SWP in the MW over the 2007 growing season. ..................................................... 61
Figure 3.3 Sap flow velocity of tree T5 located in the pure spruce treatment (PS) with air
temperature in the PS, and tree T1 located in the mixedwood treatment (MW) with air
temperature in the MW over the 2007 growing season. .................................................. 62
Figure 3.4 Comparison of daily mean of soil water potential (SWP) between pure spruce
treatment (in red) and mixedwood treatment (in green) from 2007 to 2018 using box
plots. Soil water potential ranges from 0 MPa to -1.5 MPa. Soil water potential of 0 MPa
indicates that the soil is in a state of saturation, increasingly negative values occur as the
soil becomes drier and water less available for the trees. The box plot visually shows the
distribution of the data and skewness through displaying the interquartile range (box),
median (horizontal line), whiskers (vertical lines) and outlines (circles). ....................... 63
Figure 6.1 Inga Overview Map. The treatments that I used in this study are the untreated
(mixedwood treatment, MW) with the plots A3, B7, C8, D1 and E7; and the herbicide
(pure white spruce PS treatment) with the plots A2, B1, C5, D4 and E1. Map source:
Powelson et al. (2016). .................................................................................................. 101
Figure 6.2 Inga Map. Mixedwood treatment with the plots A3, B7, C8, D1 and E7; and pure
white spruce treatment with the plots A2, B1, C5, D4 and E1. Climate stations are
located in the open area (west of plot A3), plot A3 and plot A2. Map image source:
Vivid - Canada, DigitalGlobe (2014). .......................................................................... 102
Figure 6.3 Illustration of pure spruce treatment- plot A2 (a) and mixedwood treatment- plot
A3 (b). Date of the pictures: September 23, 2018 (fall). ............................................... 103
Figure 6.4 Illustration of the climate station opening west of plot A3 (mixedwood treatment).
Area maintained in open condition by annual brushing; size is approximately 20 × 20 m.
Date of the picture: August 24, 2018 (summer). ........................................................... 104
Figure 6.5 Illustration of the view of plot A3 (mixedwood treatment) from the area
maintained in a open condition in the May (a) and September (b), 2019. .................... 105
Figure 6.6 Illustration of tree core extraction from a spruce tree in pure spruce treatmentplot A2 (a) and a spruce tree in mixedwood treatment- plot A3 (b). Date of the pictures:
August 24, 2018. ............................................................................................................ 106
Figure 6.7 Daily average photosynthetically active radiation (PAR) for the 2007 year at a
height of 3.0 m in the open area (west of plot A3) and at a height of 3.0 under the
broadleaf canopy in the mixedwood treatment plot A3................................................. 106
Figure 7.1 Relative importance of the 17 annual climatic predictor variables using the full
data set. Each variable’s mean minimum depth refers to the average minimal depth that
the variable within the Random Forest regression tree. Variable with lower mean
V

minimum depth indicates a split closer to the root of the tree and an increased
importance of the variable to spruce growth. The word “tree” refers to the regression
trees. The number of trees means the number of runs of regression (10000 runs). NA
values indicate the variable was not used in an individual regression tree. .................. 107
Figure 7.2 Relative importance of the 12 annual climatic predictor variables in the pure
spruce treatment. ............................................................................................................ 108
Figure 7.3 Relative importance of the 12 annual climatic predictor variables in the
mixedwood treatment. ................................................................................................... 109
Figure 7.4 Relative importance of the 20 out of 46 seasonal climatic predictor variables in
the pure spruce treatment. .............................................................................................. 110
Figure 7.5 Relative importance of the 20 out of 46 seasonal climatic predictor variables in
the mixedwood treatment. ............................................................................................. 111
Figure 7.6 Relative importance of the 20 out of 138 monthly climatic predictor variables in
the pure spruce treatment. .............................................................................................. 112
Figure 7.7 Relative importance of the 20 out of 138 monthly climatic predictor variables in
the mixedwood treatment. ............................................................................................. 113
Figure 7.8 Interaction between annual air temperature and annual sum of number of days
where air temperature is above 5 °C on spruce growth using the full data set. ............. 116
Figure 7.9 Interaction between annual air temperature and annual soil temperature on spruce
growth using the full data set. ........................................................................................ 116
Figure 7.10 Interaction between annual air temperature and annual solar radiation on spruce
growth using the full data set. ........................................................................................ 117
Figure 7.11 Interaction between annual sum of number of days where air temperature is
above 5 °C and solar radiation on spruce growth using the full data set. ...................... 117
Figure 7.12 Interaction between annual soil temperature and annual solar radiation on spruce
growth using the full data set. ........................................................................................ 118
Figure 7.13 Interaction between soil temperature and soil water potential on spruce growth
using the full data set. .................................................................................................... 118
Figure 7.14 Interaction between annual solar radiation and soil water potential on spruce
growth using the full data set. ........................................................................................ 119
Figure 7.15 Interaction between annual air temperature and annual rainfall on spruce growth
in the pure spruce treatment. .......................................................................................... 122
Figure 7.16 Interaction between annual air temperature and annual soil temperature on
spruce growth in the pure spruce treatment. .................................................................. 122
Figure 7.17 Interaction between annual air temperature and annual solar radiation on spruce
growth in the pure spruce treatment. ............................................................................. 123
Figure 7.18 Interaction between annual air temperature and annual soil water potential on
spruce growth in the pure spruce treatment. .................................................................. 123
Figure 7.19 Interaction between annual rainfall and annual soil temperature on spruce
growth in the pure spruce treatment. ............................................................................. 124
Figure 7.20 Interaction between annual rainfall and soil water potential on spruce growth in
the pure spruce treatment. .............................................................................................. 124
Figure 7.21 Interaction between annual soil temperature and annual soil water potential on
spruce in the pure spruce treatment. .............................................................................. 125
Figure 7.22 Interaction between annual solar radiation and annual soil water potential on
spruce growth in the pure spruce treatment. .................................................................. 125
VI

Figure 7.23 Interaction between annual air temperature and annual rainfall on spruce growth
in the mixedwood treatment. ......................................................................................... 126
Figure 7.24 Interaction between annual air temperature and annual soil temperature on
spruce growth in the mixedwood treatment................................................................... 126
Figure 7.25 Interaction between annual air temperature and annual soil water potential on
spruce growth in the mixedwood treatment................................................................... 127
Figure 7.26 Interaction between annual rainfall and annual soil temperature on spruce
growth in the mixedwood treatment. ............................................................................. 127
Figure 7.27 Interaction between annual rainfall and solar radiation on spruce growth in the
mixedwood treatment. ................................................................................................... 128
Figure 7.28 Interaction between annual soil temperature and annual solar radiation on spruce
growth in the mixedwood treatment. ............................................................................. 128
Figure 7.29 Interaction between annual soil temperature and annual soil water potential on
spruce growth in the mixedwood treatment................................................................... 129
Figure 7.30 Interaction between annual solar radiation and annual soil water potential on
spruce growth in the mixedwood treatment................................................................... 129
Figure 7.31 Interaction between air temperature and soil water potential during spring and
summer on spruce growth in the pure spruce treatment. ............................................... 131
Figure 7.32 Interaction between air temperature and soil water potential during spring and
summer on spruce growth in the mixedwood treatment. ............................................... 132
Figure 7.33 Interaction between air temperature and soil water potential in May and August
on spruce growth in the pure spruce treatment. ............................................................. 134
Figure 7.34 Interaction between air temperature and soil water potential in May and August
on spruce growth in the mixedwood treatment.............................................................. 135
Figure 7.35 Comparison of daily means of air temperature at height of 3.0 m between pure
spruce and mixedwood treatments by month from 1999 to 2017. The box plot visually
shows the distribution of the data and skewness through displaying the interquartile
range (box), median (horizontal line), whiskers (vertical lines) and outlines (circles). 136
Figure 7.36 Comparison of daily means of soil temperature at a depth of 15 cm between
pure spruce and mixedwood treatments by month from 1999 to 2017. ........................ 137
Figure 7.37 Comparison of daily mean of soil water potential (SWP) at a depth of 15 cm
between pure spruce and mixedwood treatments by month from 1999 to 2017. .......... 137
Figure 8.1 Comparison of daily mean of air temperature height of 3.0 m between pure spruce
and mixedwood treatments by month from 2007 to 2018. The box plot visually shows
the distribution of the data and skewness through displaying the interquartile range
(box), median (horizontal line), whiskers (vertical lines) and outlines (circles). .......... 138
Figure 8.2 Comparison of daily mean of soil temperature at a depth of 15 cm between pure
spruce and mixed by month from 2007 to 2018. ........................................................... 139
Figure 8.3 Daily sum of precipitation in the open area by month from 2007 to 2018. ........ 139
Figure 8.4 Daily mean of solar radiation in the open area by month from 2007 to 2018. The
box plot visually shows the distribution of the data and skewness through displaying the
interquartile range (box), median (horizontal line), whiskers (vertical lines) and outlines
(circles). ......................................................................................................................... 140
Figure 8.5 Interaction between solar radiation and rainfall in the pure spruce treatment (a)
and mixedwood treatment (b). ....................................................................................... 141
Figure 8.6 Interaction between air temperature and SWP in the pure spruce treatment (a) and
mixedwood treatment (b). .............................................................................................. 142
VII

Figure 8.7 Interaction between soil temperature and SWP in the pure spruce treatment (a)
and mixedwood treatment (b). ....................................................................................... 143
Figure 8.8 Interaction between solar radiation and SWP in the pure spruce treatment (a) and
mixedwood treatment (b). .............................................................................................. 144
Figure 8.9 Interaction between solar radiation and rainfall in spring (a), early-summer (b),
and late-summer (c) in the pure spruce treatment. ........................................................ 145
Figure 8.10 Interaction between air temperature and SWP in spring (a), early-summer (b),
and late-summer (c) in the pure spruce treatment. ........................................................ 146
Figure 8.11 Interaction between soil temperature and SWP in spring (a), early-summer (b),
and late-summer (c) in the pure spruce treatment. ........................................................ 147
Figure 8.12 Interaction between solar radiation and SWP in spring (a), early-summer (b),
and late-summer (c) in the pure spruce treatment. ........................................................ 148
Figure 8.13 Interaction between solar radiation and rainfall in spring (a), early-summer (b),
and late-summer (c) in the mixedwood treatment. ........................................................ 149
Figure 8.14 Interaction between air temperature and SWP in spring (a), early-summer (b),
and late-summer (c) in the mixedwood treatment. ........................................................ 150
Figure 8.15 Interaction between soil temperature and SWP in spring (a), early-summer (b),
and late-summer (c) in the mixedwood treatment. ........................................................ 151
Figure 8.16 Interaction between solar radiation and SWP in spring (a), early-summer (b),
and late-summer (c) in the mixedwood treatment. ........................................................ 152

List of Tables:
Table 2.1 Detailed information on white spruce trees in each plot of pure spruce and
mixedwood treatments from an inventory in 2018 of all spruce trees in each plot. As
with other information in this table, the basal area is for live white spruce trees only,
excluding the broadleaves................................................................................................ 19
Table 2.2 Mean and standard deviation (SD) for the total basal area of all species, spruces
and deciduous trees for plots A2 and A3 using the total basal area calculated in each
circular neighborhood plots. Basal area (m2/ha) calculated using the diameter at breast
height (DBH) of all the trees located in a circular neighborhood plot (N) of a radius of
5.98 m (0.011 ha) centered on selected eleven trees in the plot A2 and four trees in the
plot A3 in August of 2018. .............................................................................................. 20
Table 2.3. Description of equipment used to measure the microclimate variables. ............... 21
Table 2.4. Abbreviation and description of the microclimate predictor variables used for the
analyses. ........................................................................................................................... 22
Table 2.5. Mixed effect growth model structure and fitted coefficient values of the annual
climatic predictor variables for the full data set. Coefficient values were derived using
model averaging that included all candidate models that were within 4 AIC values of the
best model. The variables selected for inclusion in the best model are highlighted in
bold, with the gray shading for variables with a negative effect on tree growth. ............ 30
Table 2.6 Mixed effect growth model structure and fitted coefficient values of the annual
climatic predictor variables across the treatment types. Coefficient values were derived
using model averaging that included all candidate models that were within 4 AIC values

VIII

of the best model. The variables selected for inclusion in the best model are highlighted
in bold, with the gray shading for variables with a negative effect on tree growth......... 34
Table 2.7 Mixed effect growth model structure and fitted coefficient values of the seasonal
climatic predictor variables across the treatment types. Coefficient values were derived
using model averaging that included all candidate models that were within 4 AIC values
of the best model. The variables selected for inclusion in the best model are highlighted
in bold, with the gray shading for variables with a negative effect on tree growth......... 37
Table 2.8 Mixed effect growth model structure and fitted coefficient values of the monthly
climatic predictor variables across the treatment types. Coefficient values were derived
using model averaging that included all candidate models that were within 4 AIC values
of the best model. The variables selected for inclusion in the best model are highlighted
in bold, with the gray shading for variables with a negative effect on tree growth......... 40
Table 3.1 Description of equipment used to measure the microclimate variables. ................ 53
Table 3.2. Abbreviation and description of the microclimate predictor variables used for the
analyses. ........................................................................................................................... 55
Table 3.3 List of the models used for the model selection. Models 1 to 6 are the models
without interaction. Models 7 to 12 are the models with interaction terms. ................... 59
Table 3.4 Akaike information criterion (AIC) values of each of the 12 mixed effect models
fitted using the full data set and across the pure spruce and mixedwood treatments
throughout the growing season. The best models of SFV using the full data set and
across the treatments are highlighted in gray. The best models without interaction are
highlighted in bold. .......................................................................................................... 64
Table 3.5 Akaike information criterion (AIC) values of each of the 12 mixed effect models
fitted using the full data set and across the pure spruce and mixedwood treatments. The
best SFV models for each season across the treatments are highlighted in gray. Asterisk
(*) selected as best model but excludes SWP. ................................................................. 66
Table 3.6 Mixed effect growth model structure of selected model and fitted coefficient
values of climatic predictor variables using the full data set and across the treatment
types over the growing season. The variables (i.e. direct effect) that differ across the
treatment types regarding their negative or positive influence on SFV is highlighted in
gray. The interactions that differ across the treatment types regarding their directions
(i.e. positive or negative) is also highlighted in gray. ...................................................... 68
Table 3.7 Mixed effect growth model structure of selected model and fitted coefficient
values of climatic predictor variables across the treatment for each season. The variables
that differed regarding their negative or positive influence on SFV across the seasons in
each treatment are highlighted in gray, with dark gray for the season that differed from
the others. The interactions that differ between the seasons in each treatment regarding
their directions (i.e. positive or negative) are also highlighted in gray, with dark gray for
the season(s) that the interaction strongly differed from other(s). ................................... 71
Table 6.1 Total basal area (m2/ha) of all species, spruces and deciduous in circular
neighborhood plots (n= 11) centered on selected spruce trees (t= 11) in plot A2 (pure
spruce treatment) and in circular neighborhood plots (n= 4) centered on selected spruce
trees (t= 4) in plot A3 (mixedwood treatment). Basal area calculated using the diameter
at breast heigh (DBH) of all trees located in each circular neighborhood plot with radius
of 5.98 m (0.011 ha). Tree identification (ID) refers to the selected spruce trees (i.e.
focal trees) that circular neighborhood plots were centered. ......................................... 104
IX

Table 7.1 Random Forest (Liaw and Wiener 2002) selection and ranking of the 5 most
important annual predictor variables for tree growth using the full data set. ................ 113
Table 7.2 Random Forest selection and ranking of the 5 most important annual, seasonal and
monthly predictor variables for tree growth across the treatment types. ....................... 114
Table 7.3 Details of the best mixed effect growth models with and without two-way
interaction terms using annual climatic predictor variables for the full data set. .......... 115
Table 7.4 Details of the best mixed effect growth models with and without two-way
interaction terms using annual climatic predictor variables for the pure spruce treatment.
....................................................................................................................................... 120
Table 7.5 Details of the best mixed effect growth models with and without two-way
interaction terms using annual climatic predictor variables for the mixedwood treatment.
....................................................................................................................................... 121
Table 7.6 Details of the best mixed effect growth models with and without two-way
interaction terms using seasonal climatic predictor variables for the pure spruce and
mixedwood treatments. .................................................................................................. 130
Table 7.7 Details of the best mixed effect growth models with and without two-way
interaction terms using monthly climatic predictor variables for the pure spruce and
mixedwood treatments. .................................................................................................. 133
Table 9.1 Equipment descriptions for 1999-2018 climate stations. ...................................... 153
Table 9.2 Sap flow velocity data removed from the data set prior to analysis. .................... 154
Table 9.3 Years that measurements were included in the data set (x). Years that are not
checked with “x” were either not measured in the year or presented measurement error
over the entire year due to equipment problems and were not included in the data set.
Hourly measurement errors were removed from the data set. Colours correspond to
climate station locations. ............................................................................................... 155

X

1. Chapter 1: Introduction

1.1. Tree and forest responses to climate
Forest composition, growth and dynamics are influenced by weather and climate
through a variety of demographic, physiological, and ecosystem functions (Bonan 2002).
Climatic warming during the past century has led to a variety of responses by forest
ecosystems, such as changes in forest growth (Barber et al. 2000), carbon balances (Piao et
al. 2008, Arnone III et al. 2008), and phenology (Cleland et al. 2007). Northern sub-boreal,
boreal and subarctic regions (Lloyd et al. 2002, Lloyd and Fastie 2003) have experienced
some the largest warming impacts. In many areas, warming has been responsible for changes
in the hydrological balance leading to severe levels of summer drought (van Mantgem et al.
2009). Climate change is projected to continue to increase in the future (McCarthy et al.
2001, Cooper et al. 2002), with projected large impacts on forest ecosystems (Hamann and
Wang 2006, Cleland et al. 2007, Piao et al. 2008, Price et al. 2013, Luo and Chen 2015).
Gaining a better understanding of how trees and forest ecosystems respond to climate drivers
will provide important information to forest managers that will hopefully allow them to
develop improved adaptation or mitigation strategies.
Boreal forests cover approximately one-third of the global forest area and contain half
of the global forest carbon (Jiang et al. 2016). Climate change will play an important role in
the way forests will be managed in the future, even more for the boreal forests due to their
extent and role in global carbon dynamics. Canada has 34% of its territory covered with
forests, approximately 17% of global forest lands (FAO 2016). Forests are the dominant Cabsorbing component in the Northern Hemisphere (Pan et al. 2011). On the other hand,
forestry, a major component of Canada’s economy (FAO 2016), is one of the major
1

contributors to carbon emissions (Le Quéré et al. 2009), and global demands for forest
commodities are not likely to reduce. To avoid forest declines and an associated reductions in
forest ecosystem services (e.g. timber, carbon storage, and water regulation), forest
management practices need to adapt the forest to the ongoing environmental changes.
Projected changes in climate will likely pose challenges for forest management and
conservation (Elliott et al. 2015), and will influence tree growth (Goldblum and Rigg 2005,
Lloyd et al. 2013). Global warming from 2016-2035 is projected to result in temperature
increases between 0.3 to 0.7°C (IPCC 2013). Boreal forests are predicted to be especially
sensitive and vulnerable to changing climates. In Canada, climate has warmed and is
projected to warm further in the future, with increases in annual and seasonal temperature
over the country (Flato et al. 2019). Projections for 2055 in the Northeast Region of British
Columbia (Foord 2016), Canada, indicated an increase in mean annual temperature by 3.3°C,
with minimum temperatures increasing more than maximum temperatures. The number of
growing degree-days and the number of frost-free days are also projected to increase. Mean
annual precipitation is projected to increase by 10%, but precipitation as snow may decrease
by 10%. Despite the increase in summer precipitation, evaporation and climate moisture
deficit will increase as temperatures increase (Foord 2016).
Altered precipitation and warming air will possibly result in changes in tree growth
rates, mortality rates and species interactions (Konar et al. 2010, Zolkos et al. 2015). Changes
in climate will also have economic and ecological impacts resulting from events such as
drought-related mortality of forest trees, increased severity of fire, insect and disease
epidemics (Woods et al. 2005). Due to the ecological and economic importance of forests,
studies on forests response to climate change are very important. Even though many studies
have examined climate change effects on forests, there are still uncertainties regarding how
2

forests will respond to future climate. To better project how forests will develop, we need to
improve our understanding of how they grow, particularly under novel future climate
scenarios.
The study of tree growth has been a focus of scientific research for a long time, with
the first attempts in the form of yield tables in the early 19th century (Rohner et al. 2013).
Such studies have demonstrated the many ways that climate variables influence tree growth.
For example, Harley et al. (2011) studying slash pine growth in the United States found that
annual growth of slash pine is primarily influenced by water availability during the growing
season. Temperatures have a strong relationship with growth for white spruce, lodgepole
pine, and subalpine fir in British Columbia (B.C.) (Miyamoto et al. 2010). Researching
individual tree growth response to current climate will provide information about their
response to expected changes in climate and disturbances such as drought stress.
Drought-induced water stress has been identified as one of the main contributors of
widespread tree mortality, and growth decline in the western boreal forests of Canada (Hogg
et al. 2008, Ma et al. 2012). Drought conditions can lead to a reduction in the extent of
annual wood formation (Lautner 2013). Xylem cells may not expand fully because of the
lack of turgor pressure under water-deficit conditions (Steppe et al. 2015, Deslauriers et al.
2016). Balducci et al. (2015) revealed a lower wood density formed in black spruce seedlings
during droughts, reflecting a lower carbon allocation to cell wall formation, resulting in a
hydraulic system that is less able to cope with drought.
Studying which climatic variables are the most influential for tree growth will better
allow researchers and forest managers to project how trees will develop in the current and
future climate conditions. Studies have projected tree growth response to future climate based
in the past and current tree sensitivity to climate variables such as temperature and
3

precipitation. For example, a study on jack pine trees indicated that air temperature is a better
predictor of height growth than precipitation in eastern Canada and the United States
(Thomson and Parker 2008), whereas in B.C., Cortini et al. (2011) found that monthly
temperature and precipitation are both effective in predicting lodgepole pine (Pinus contorta
Dougl. ex Loud. var. latifolia) diameter growth. Elliott et al. (2015) found that hydroclimatic
variability influences the growth of eastern deciduous trees in the United States; suggesting
that if precipitation distributions change in the future, growth loss may be significant.
Tree growth responses to climate vary among species (e.g. Rehfeldt et al. 1999,
Miyamoto et al. 2010, Messaoud and Chen 2011, Legendre-Fixx et al. 2017). Some species
are more sensitive to climate or combinations of climatic variables than others (Clark et al.
2012). Furthermore, species under the same climatic conditions may differ in their growth
response according to site conditions (Elliott et al. 2015), such as soil type, slope position and
forest composition. A study conducted in southern interior British Columbia (B.C.), Canada,
showed that tree-ring width for all three species (Douglas-fir, lodgepole pine, and hybrid
white spruce) was primarily affected by climate variables from the year previous to the
growing season and only secondly by current year conditions; however, the critical months
varied between species and altitudes (Lo et al. 2010). Goldblum and Rigg (2005) found
differences in tree-ring sensitivity to monthly climate signals between the three species at the
deciduous-boreal forest ecotone in Canada. As the tree growth responses to climate differ
between species, the capacity to adapt to climate change is also expected to vary among
species. Besides climate, other site conditions such as stand dynamics, competition, and
silvicultural practices (e.g. brushing, spacing, and thinning), can also influence tree growth
(Lo et al. 2010, Linderholm and Linderholm 2014).

4

Competition is also a significant process driving forest dynamics (Coomes and Allen
2007, Kunstler et al. 2011, Sánchez-Salguero et al. 2015), and can influence how tree growth
responds to climate (Weber et al. 2008, Ruiz-Benito et al. 2013, Lebourgeois et al. 2014,
Madrigal-González and Zavala 2014, Fernández-de-Uña et al. 2015, Trouvé et al. 2015). For
example, Sánchez-Salguero et al. (2015), studying the effects of competition and climate on
three Scots pine stands in Spain, found that tree growth sensitivity to climate increased with
decreasing competition intensity. Climate change may intensify the effects of competition
(Metsaranta and Lieffers 2008, Luo and Chen 2015), which makes competition very
important in modeling how forests will respond to changes in climate. The increase in
average temperature, for example, may result in increased competition from other species
better suited to warmer climates (Spittlehouse 2008). Thus, the study of tree growth response
to climate should take into account species differences and their site conditions.
It is still unclear how mixewood stands influence tree growth under drought stress
compared with their performance in a pure stand (Pretzsch et al. 2010, Richards et al. 2010).
Studies show that mixedwood stands frequently improve resource supply, resource uptake,
and resource use efficiency, as a result, also tree and stand growth (Kelty 1992, Richards et
al. 2010). Niche complementarity can decrease competition for resources in mixedwood
versus pure stands (Morin et al. 2011). Another advantage is that trees species in mixedwood
stands can interact in such a way that one species exerts a positive effect and facilitates the
other species. An example of facilitation is the hydraulic lift by one species with the benefit
of water supply to the other (Brown et al. 2014). The stress-gradient theory hypothesizes that
facilitation prevails on poor sites, whereas on rich sites competition prevails (Callaway and
Walker 1997). Thus, the study of interactions among species in different site conditions are
essential for understanding, managing, and forecasting growth of forest stands.
5

1.2. White spruce growth and climate sensitivity
White spruce (Picea glauca (Moench) Voss) is found in all forested regions of
Canada except on the Pacific Coast. It is common in northern forests, occurring on a variety
of soils and under a wide range of climatic conditions (Nienstaedt and Zasada 1990, Farrar
1995), usually sharing the forest environment with trembling aspen, white birch, black
spruce, and balsam fir. White spruce trees are very important for the production of wood pulp
and lumber, and frequently planted for landscape and forestry purposes (Farrar 1995). White
spruce stands are a source of cover and food for many species such as moose, hares, red
squirrels, and spruce grouse, and have considerable value in maintaining soil stability and
watershed, and for recreation (Nienstaedt and Zasada 1990). White spruce is one of the most
productive and widespread forest types in Canada. Any climate-related change in white
spruce growth is likely to be an important factor in carbon sequestration in the boreal forest, a
region considered one of the planet’s major carbon sinks (Barber et al. 2000).
There are still uncertainties of how white spruce tree growth will respond to climate
change, but studies suggest that it will vary over a gradient of local climatic conditions
(Miyamoto et al. 2010). Goldblum and Rigg (2005) found that white spruce is likely to
benefit less in growth compared to sugar maple and balsam fir in a deciduous–boreal forest
ecotone in Ontario. They project that white spruce will experience a slight increase in growth
rate as a positive response to increasing winter temperatures and a decrease in growth rate as
negative response (January and February) to increasing winter precipitation. They identified
that only in September and April, both temperature and precipitation responses were positive
to white spruce growth, with the response during the remaining months being mixed (i.e.
negative or positive).

6

Declining growth and increased mortality of spruce in boreal forests have generally
been attributed to drought stress (Barber et al. 2000, Lloyd and Fastie 2003), and drought
stress is expected to increase, especially during summer, based in climate projections.
Productivity of pure spruce stands was found to increase from 40% to 60% in the absence of
drought, whereas in the presence of drought it declined by 20% or more (Johnston and
Williamson 2005). Spruce is identified as isohydric, which means that water consumption
and growth are reduced in the early phases of drought stress through stomata closure (Zang et
al. 2012, Pretzsch et al. 2013, Sullivan et al. 2017), increasing the probability of carbon
starvation (McDowell 2011, Kulaç et al. 2012).
Spruce growing in mixedwood stands might be more sensitive to drought stress
compared to pure stands due to higher competition for limiting resources (primarily water).
Unwanted vegetation could become a stronger competitor for water in white spruce stands
(Weber et al. 2008). In the boreal zone of B.C., the height growth of white spruce in for
future period 2005-2035 could potentially increase by around 3% on average where
vegetation control or mechanical site preparation is applied, whereas white spruce growth in
untreated stands (i.e. no site preparation or vegetation control) may suffer decreases in height
of up to 10% due to increased drought stress and shading (Cortini et al. 2011). Another study
in boreal mixtures of western Canada concluded, based on current global warming trends,
that an aspen canopy could limit the response of spruce to temperature increases due to light
limitations, and may be more competitive for water and light compared to past climates
(Cortini et al. 2012). However, tree growth under water stress in pure versus mixedwood
boreal stands is not well explored and understood.
Previous studies found that stand composition and structure influences the availability
of resources such as water and light throughout the year. The redistribution of soil water by
7

aspen (Populus tremuloides Michx.) root systems improves rooting-zone soil moisture
conditions (Brown et al. 2014), which can benefit spruce planted under established aspen
(Kabzems et al. 2016). Mixed species stands comprised of trees capable of hydraulical
redistribution have the potential to maintain high transpiration rates during periods of water
shortage (i.e. low rainfall periods) (Brown et al. 2014). Canopy openings resulting from the
senescence of deciduous trees in a mixed stand allow for the release of more shade-tolerant
slow growing conifers (Brassard and Chen 2006). Thus, studies of spruce growth sensitivity
to climate under different forest composition and structure are essential to better project
spruce trees growth under current and future climate conditions.
Past research has generally demonstrated that white spruce growth is climate sensitive
(Cortini et al. 2011, Lloyd et al. 2013). For example, Lloyd et al. (2013) found that total
precipitation during the previous summer and late winter seasons had a great influence on
white spruce growth in interior Alaska. Studies indicate that annual temperature and
precipitation influence white spruce growth and survival (Barber et al. 2000, Cortini et al.
2011, Lloyd et al. 2013, Lu et al. 2014). Temperature and precipitation are two factors that
can predict up to 45% of the variation in white spruce growth (Cortini et al. 2011). Better
growth rates were found with cooler, wetter years (Lloyd et al. 2013), and growth declines
were more common in warmer and drier parts of the boreal forest (Lloyd and Fastie 2003).
Few studies examined how climate influences tree growth over a very fine time scale.
Herrmann et al. (2016) revealed that changes in a tree’s stem circumference can be vary with
its water status, and that the amplitude of daytime shrinkage of the stem circumference was
significantly correlated with climate variables, sap flow, and evapotranspiration. The study of
the influence of climate variability on tree diameter on time scales of minutes to years can
provide important information for forest management. While many studies focus on finding
8

the climate variables that most influence annual tree growth, we know little about how spruce
growth responds to inter- and intra-annual variation in climate and if the response varies with
stand composition and structure. Furthermore, questions still remain about how inter and
intra-annual variation in climate variables influence white spruce growth in pure and
mixedwood stands in British Columbia and western Canada.

1.3. Thesis objectives
The purpose of this study was to examine how individual white spruce trees (Picea
glauca (Moench) Voss) are influenced by climate variables over short and intermediate time
periods in two different stand structures: pure white spruce stands and mixedwood stands
(white spruce, green alder, willow, and aspen). The study area is located north of Fort St.
John, B.C.. Specific objectives were as follows:
1. To analyze annual white spruce growth sensitivity to annual microclimate variables.
2. To analyze annual white spruce growth sensitivity to annual, seasonal and monthly
microclimate variables in pure white spruce stands versus mixedwood stands.
3. To analyze white spruce sap flow sensitivity to microclimate variables throughout the
growing season.
4. To analyze white spruce sap flow sensitivity to microclimate variables throughout the
growing season and between seasons in pure a white spruce stand versus a
mixedwood stand.
In addition to this introductory chapter, this thesis contains two data chapters written
as manuscripts that address the above questions, and a concluding chapter where data
chapters are synthesized. In the first data chapter (Chapter 2), I tackle the first and second
objective utilizing dendrochronology studies and microclimate data obtained from an on-site
9

station. In the second data chapter (Chapter 3), I tackle the third and fourth objectives,
utilizing sap flow measurements and microclimate data obtained from an on-site station.

10

2. Chapter 2: White spruce growth sensitivity to climate variability in pure and
mixedwood stands

Abstract: Projected changes in climate will likely influence tree growth and stand dynamics,
and pose challenges for forest management and conservation. To better project how forests
will develop, we need to improve our understanding of how trees grow, particularly under
novel future climate scenarios. Past research has demonstrated that white spruce (Picea
glauca (Moench) Voss) is climate sensitive. However, questions remain about how inter- and
intra-annual variations in climate variables influence annual spruce growth and whether stand
composition and structure can modify the realized microclimate conditions and the trees’
responses to weather stress. Using data from 1999 to 2017, I evaluated the importance and
the influence of annual, seasonal and monthly microclimate variables on the annual growth
of white spruce trees in pure and mixedwood stands located at the Inga Lake site, in northeastern British Columbia. First, I used Random Forest analyses to identify which climate
variables were most important to predict tree growth using the full data set, and each
treatment individually. Second, the best climate predictor variables were evaluated by fitting
tree growth to selected explanatory variables using a linear mixed effect model framework.
The order of importance of microclimate variables differed between pure and mixedwood
stands. Annual rainfall, soil water potential (SWP) in the spring and summer, air temperature
in May, and soil water potential in August, were found to impact tree growth, but the relative
importance and direct effects (positive or negative) of each variable differed between pure
and mixedwood stands. The influences of air temperature and SWP varied throughout the
year. Warm springs increased spruce growth, and warm summers decreased spruce growth in
both stand types. Spruce growth in pure stands had a positive relationship with soil water
11

potential during spring and summer, while spruce growth in the mixedwood stands had a
negative relationship. In both stand types, a combination of warmer temperatures and drought
during summer is likely to decrease wood production.
Keywords: Tree growth, climate variability, white spruce, drought.

2.1. Introduction
Projected changes in climate will likely pose challenges for forest management and
conservation (Elliott et al. 2015), and will influence tree growth (Goldblum and Rigg 2005,
Lloyd et al. 2013). Boreal forests are predicted to be especially sensitive and vulnerable to
changing climates. In Canada, the climate has warmed and is projected to warm further in the
future, with increases in annual and seasonal temperatures over the country (Flato et al.
2019). Projections for 2055 in the Northeast Region of British Columbia (Foord 2016),
Canada, indicated an increase in mean annual temperature by 3.3°C, with minimum
temperatures increasing more than maximum temperatures. The number of growing degreedays and the number of frost-free days are also projected to increase. Mean annual
precipitation is projected to increase by 10%, but precipitation as snow may decrease by
10%. Despite the increase in summer precipitation, evaporation and climate moisture deficit
will increase as air temperatures increase (Foord 2016).
Due to the ecological and economic importance of forests, studies of forest response
to climate change are very important. Even though many studies have examined climate
change effects on forests, there are still uncertainties regarding how forests will respond to
future climate. To better project how forests will develop, we need to improve our
understanding of how they grow, particularly under novel future climate scenarios.

12

The study of tree growth has been a focus of scientific research for a long time, with
the first attempts in the form of yield tables in the early 19th century (Rohner et al. 2013).
Such studies have demonstrated the many ways that climate variables influence tree growth.
For example, Harley et al. (2011) studying slash pine growth in the United States found that
annual growth of slash pine is primarily influenced by water availability during the growing
season. Temperatures have a strong relationship with growth for white spruce, lodgepole
pine, and subalpine fir in British Columbia (B.C.) (Miyamoto et al. 2010). Researching
individual tree growth response to current climate will provide information about their
response to expected changes in climate and disturbances such as drought stress.
Drought-induced water stress has been identified as one of the main contributors of
widespread tree mortality, and growth decline in the western boreal forests of Canada (Hogg
et al. 2008, Ma et al. 2012). Drought conditions can lead to a reduction in the extent of
annual wood formation (Lautner 2013). Xylem cells may not expand fully because of the
lack of turgor pressure under water-deficit conditions (Steppe et al. 2015, Deslauriers et al.
2016). Balducci et al. (2015) revealed a lower wood density formed in black spruce seedlings
during droughts, reflecting a lower carbon allocation to cell wall formation, resulting in a
hydraulic system that is less able to cope with drought.
Studying which climatic variables are the most influential on tree growth will better
allow researchers and forest managers to project how trees will develop in the current and
future climate conditions. Studies have projected tree growth response to future climate based
in the past and current tree sensitivity to climate variables such as temperature and
precipitation. For example, a study on jack pine trees indicated that air temperature is a better
predictor of height growth than precipitation in eastern Canada and the United States
(Thomson and Parker 2008), whereas in B.C., Cortini et al. (2011) found that monthly
13

temperature and precipitation are both effective in predicting lodgepole pine diameter
growth. Elliott et al. (2015) found that hydroclimatic variability influences the growth of
eastern deciduous trees in the United States; suggesting that if precipitation distributions
change in the future, growth loss may be significant.
Tree growth responses to climate vary among species (e.g. Rehfeldt et al. 1999;
Miyamoto et al. 2010, Messaoud and Chen 2011, Legendre-Fixx et al. 2017). Some species
are more sensitive to climate or combinations of climatic variables than others (Clark et al.
2012). Furthermore, species under the same climatic conditions may differ in their growth
response according to site conditions (Elliott et al. 2015), such as soil type, slope position and
forest composition. For example, a study of the effects of competition and climate on three
scots pine stands in Spain, indicated that tree growth sensitivity to climate increased with
decreasing competition intensity (Sánchez-Salguero et al. 2015). Climate change may
intensify the effects of competition (Metsaranta and Lieffers 2008, Luo and Chen 2015). The
increase in average temperature, for example, may result in increased competition from other
species better suited to warmer climates (Spittlehouse 2008). Thus, the study of tree growth
response to climate should take into account individual species and their site conditions.
Studies indicate that annual temperature and precipitation influence white spruce
growth and survival (Barber et al. 2000, Cortini et al. 2011, Lloyd et al. 2013, Lu et al. 2014).
Temperature and precipitation are two factors that can predict up to 45% of the variation in
white spruce growth (Cortini et al. 2011). Better growth rates were found with cooler, wetter
years (Lloyd et al. 2013), and growth declines were more common in warmer and drier parts
of the boreal forest (Lloyd and Fastie 2003). There are still uncertainties of how white spruce
tree growth will respond to climate change, but studies suggest that it will vary over a
gradient of local climatic conditions (Miyamoto et al. 2010). In the boreal zone of B.C., the
14

height growth of white spruce in for future period 2005-2035 could potentially increase by
around 3% on average where vegetation control or mechanical site preparation is applied,
white spruce growth in untreated stands (i.e. no site preparation or vegetation control) may
suffer decreases in height of up to 10% due to increased drought stress (Cortini et al. 2011).
While many studies focus on the relationship between annual climate variables and
annual spruce growth, we know little about how annual spruce growth responds to within
year climate variability and if the response varies with stand composition and structure.
Furthermore, questions remain about how inter and intra-annual variations in climate
variables influence annual white spruce growth in British Columbia and western Canada.
Therefore, the objectives of this study were to:
1. Identify the important annual microclimate variables influencing annual white spruce
growth.
2. Determine the relationships between annual microclimate variables and annual white
spruce growth.
3. Identify the important annual, seasonal and monthly climate variables influencing
annual white spruce growth in pure white spruce stands versus mixedwood stands.
4. Determine and compare the relationships between annual, seasonal and monthly
climate variables and annual white spruce growth in pure white spruce stands versus
mixedwood stands.

By assessing individual white spruce growth responses to climate in the pure and
mixedwood stands, I provide information that will be useful in modeling and managing these
stands across western Canada under current and future climate conditions.

15

2.2. Materials and methods
2.2.1. Study area description
The Inga Lake research site (56°37’N, 121°38’W) is located 60 km northwest of Fort
St. John, British Columbia, in the Peace variant of the moist warm subzone of the Boreal
White and Black Spruce zone (BWBSmw, Delong et al. 2011) (Figure 2.1). Slope ranges
from 0 to 15% and aspect is variable on gently rolling terrain. Mean elevation at the site is
890 m above sea level. The soils are Orthic Gray or Gleyed Solonetzic Gray Luvisols with
approximately 15% coarse fragments and silt loam to clay loam soil textures (Lord and
Green 1973). The soil moisture regime (SMR) grades from mesic (occasionally submesic) in
upper and mid-slope positions to subhygric in lower slope positions, and soil nutrient regime
(SNR) is medium to rich. Forest floor depth estimated prior to treatment installation was 5
cm (Powelson et al. 2016).

Figure 2.1 Study area location with Prince George and Fort St. John as references. Map created using
ESRI ArcGIS software.
16

The cold, continental climate has mean annual temperatures of 1.5°C, mean annual
precipitation of 483 mm, mean summer precipitation of 309 mm, precipitation as snowfall
143 mm, and average frost-free period of 99 days (Delong et al. 2011, Wang et al. 2011,
Powelson et al. 2016). There is no harvest history for this site, but it was periodically burned
by ranchers until the 1950s to improve grazing, a practice that contributed to the development
of abundant willow (Salix spp.) and green alder (Alnus crispa [Ait] Pursh) (Powelson et al.
2016).
The Inga Lake study site was installed in 1987 to study the effects of silvicultural site
preparation on white spruce establishment and growth. The Inga Lake site supported wellestablished shrub-dominated plant communities prior to installation of the experimental
treatments. In the winter of 1986-1987, an area of approximately 20 ha was mechanically
sheared with a brush blade mounted on a crawler tractor to remove above-ground vegetation.
Vegetation was piled into windrows, leaving the site free of woody vegetation when the
experimental treatments were installed.
In the spring of 1987 the experiment was laid out as a randomized block design with
5 blocks (A, B, C, D, and E) following a topographic sequence from well-drained hilltop
(Block A, submesic to mesic) to a moist toe-slope position (Block E, subhygric). Then, eight
site preparation treatments were randomly applied to treatment plots (Appendix 6, Figure
6.1). I investigated two of these experimental treatments: a mixedwood treatment (MW) and
pure white spruce treatment (PS), each with five replications (plots). The mixedwood
treatment includes the plots A3, B7, C8, D1 and E7; and the pure white spruce treatment
includes the plots A2, B1, C5, D4 and E1 (Appendix 6, Figure 6.2 and 6.3).
In the first week of June 1988, each replication was planted with 48 two-year-old
white spruce seedlings (PSB 313 2+0 stock), totaling 240 white spruce seedlings in each
17

treatment. In each replication, the 48 trees were tagged using a unique identification, and
were on average, 23 cm tall and 0.4 cm in diameter at the time of planting. Seedlings were
planted to the root collar without screefing. Alder (Alnus crispa [Ait] Pursh), willow (Salix
spp.) and aspen (Populus tremuloides Michx.) trees resprouted vigorously following winter
shearing on all treatment plots. Shapes and sizes of the replicate plots were somewhat
variable; but all were approximately 0.052 ha.
In the PS treatment, glyphosate (Vision R) was broadcast applied at a rate of 2.14 kg
acid equivalent per hectare (ae/ha) in August 20, 1990. However, the treatment failed to
control willow, which had been defoliated by insects in the year of herbicide application.
Manual cutting treatments were subsequently applied 4, 6, 8, 9, 11, and 14 years after
planting to remove competing vegetation. More details of the Inga Lake study area,
experimental design and establishment can be found in Powelson et al. (2016). Inventory of
all spruce trees planted in each plot in 2018 showed significant differences between the two
treatments, for example, the mean spruce basal area of all plots in the PS was 26.6 m2/ha,
whereas in the MW it was 3.8 m2/ha (Table 2.1).

18

Table 2.1 Detailed information on white spruce trees in each plot of pure spruce and mixedwood
treatments from an inventory in 2018 of all spruce trees in each plot. As with other information in this
table, the basal area is for live white spruce trees only, excluding the broadleaves.

Measurements of White Spruce Trees

Treatment
Pure Spruce

Mixedwood

Plot
A2
B1
C5
D4
E1
A3
B7
C8
D1
E7

Density
(stems/ha)
Alive Dead
1152 50
1274 85
1110 159
1178 201
1207 140
691
817
1216 501
1004 171
841
505
1070 357

Survival
(% alive)

Mean
height (cm)

Mean
DBH (cm)

Basal area
(m2/ha)

96%
94%
88%
85%
90%
46%
71%
85%
63%
75%

1107
1058
971
1153
1146
351
601
749
349
554

17.1
16.0
15.6
17.5
16.8
4.8
6.9
9.2
4.1
6.2

27.3
27.0
21.7
29.2
27.7
1.3
5.5
7.5
1.1
3.8

In 2018, the MW treatment was dominated by 8 to 12 m tall willow and aspen canopy
overtopping 4 to 6 m tall planted spruce. The pure spruce treatment had a closed canopy of
white spruce over 10 m tall, with a very minor presence of tall shrubs or other tree species
(Haeussler et al. submitted). In August of 2018, I measured the diameter at breast height
(DBH) of all the trees located in circular neighborhood plots with radii of 5.98 m (0.011 ha)
centered on eleven selected trees in plot A2 and four trees in the plot A3. I calculated basal
area of each species and the total basal area of all species, spruces, and deciduous for each
circular neighborhood plot using these measurements (Appendix 6, Table 6.1). I calculated
the mean and standard deviation (SD) of the total basal area of the circular neighborhood
plots for the basal area of each plot (Table 2.2). Total basal area of all species in the two plots
were similar, but with different species composition.

19

Table 2.2 Mean and standard deviation (SD) for the total basal area of all species, spruces and
deciduous trees for plots A2 and A3 using the total basal area calculated in each circular
neighborhood plots. Basal area (m 2/ha) calculated using the diameter at breast height (DBH) of all the
trees located in a circular neighborhood plot (N) of a radius of 5.98 m (0.011 ha) centered on selected
eleven trees in the plot A2 and four trees in the plot A3 in August of 2018.

Basal (DBH) area (m2/ha)
Treatment

Total (all species)

Spruces

Deciduous trees

Plot

N

Mean

SD

Mean

SD

Mean

SD

Pure Spruce A2

11

35.870

4.884

34.736

5.547

1.134

2.142

A3

4

34.846

14.653

0.896

0.634

33.950

14.267

Mixedwood

2.2.2. Dendrochronology
I measured annual white spruce growth by using dendrochronology methods. In
August of 2017, I selected six random white spruce trees in each replicate plot for a total of
30 trees in each of the two treatments. Using a 5 mm increment borer (tree-coring tool), I
extracted two increment cores from each selected white spruce tree in a 90˚ angle, on the
north and west directions, to capture the variation in growth in each individual tree and
minimize the impact on the sampled trees (Appendix 6, Figure 6.6). The core from the north
and west directions were collected at a tree height of 1.20 m and 1.25 m, respectively. Cores
were air-dried and secured to wood core mounts, then sanded and polished using increasingly
fine sandpaper.
I dotted cores following the methods of Stokes and Smiley (1968), and visually crossdated using the list method (Yamaguchi 1991). I measured ring width on each core to the
nearest 0.001 cm using Windendro® 2012, and verified the visual cross-dating with the
dendrochronology program COFECHA (Holmes 1983, Grissino-Mayer 2001). Tree-ring
analyses were done in the Statistical Software R (R Development Core Team 2011) and
explained in detail in the statistical analysis section of this Chapter.
20

2.2.3. Microclimate data collection
Microclimate data were obtained from an on-site climate station installed at Inga
Lake. A variety of microclimate variables (Table 2.3; Appendix 9, Tables 9.1 and 9.3) were
measured from 1999 to 2017 every hour (standard time) at plots A2 and A3 and at the
climate station opening, and recorded on a data logger (models CR10X and CR10, Campbell
Scientific). At the climate station opening (Appendix 6, Figure 6.4), microclimate variables
included air temperature, solar radiation and rainfall. At PS (plot A2) and MW (plot A3),
microclimate variables included air temperature, soil temperature, and soil water potential.

Table 2.3. Description of equipment used to measure the microclimate variables.
Variable

Position (cm) d

Sensor make/model

Sensor type

Solar radiation a

+300

Li-Cor/LI200S

Silicone pyranometer

Rainfall a

+60 or +80

Sierra Misco/2501 or TE525m

Tipping bucket

-2.5, -15 and -50

Home built/twisted soldered wire

Cu-Co thermocouple

Soil temperatureb
Soil water potential
Air temperaturec
a

b

-2.5, -15, and -50 Campbell Sci/model 223

Gypsum Block

+130 and +300

Cu-Co thermocouple

Home/fine wire 36AWG

Variables measured in climate station opening west of plot A3 (mixedwood treatment). Area

maintained in open condition by annual brushing; size is approximately 20 × 20 m. b Variables
measured in climate station located within plots A2 (pure spruce treatment) and A3 (mixedwood
treatment). c Variable measured in climate station opening, plots A2 and A3. d Values for height (+)
indicate height (cm) above the ground surface (regardless of whether mineral soil or organic
material). Values for soil depth (-) indicate depth from the mineral soil forest floor interface.

2.2.4. Statistical analysis
I summarized the hourly microclimate measurements from 1999 to 2017 into annual,
seasonal and monthly means or sums (Table 2.4).

21

Table 2.4. Abbreviation and description of the microclimate predictor variables used for the analyses.
Variable (abbreviation)

Descriptionb

Rain_opena

Sum of precipitation at a height of 0.6 m or 0.8 m (mm)

SolRad_3.0m_opena

Sum of solar radiation at a height of 3.0 m (KW/m2)

ndays_AirTmp0_1.3m

Sum of number of days with air temperature below 0 °C

ndays_AirTmp0_1.3m_opena

Sum of number of days with air temperature below 0 °C

ndays_AirTmp5_1.3m

Sum of number of days with air temperature above 5 °C

ndays_AirTmp5_1.3m_opena

Sum of number of days with air temperature above 5 °C

SoilTmp_2.5cm

Mean of soil temperature at a depth of 2.5 cm (°C)

SoilTmp_15cm

Mean of soil temperature at a depth of 15 cm (°C)

SoilTmp_50cm

Mean of soil temperature at a depth of 50 cm (°C)

AirTmp_1.3m

Mean of air temperature at a height of 1.3 m (°C)

AirTmp_3.0m

Mean of air temperature at a height of 3.0 m (°C)
a

Mean of air temperature at a height of 1.3 m (°C)

AirTmp_3.0m_opena

Mean of air temperature at a height of 3.0 m (°C)

SWP_2.5cm

Mean of soil water potential at a depth of 2.5 cm (MPa)

SWP_15cm

Mean of soil water potential at a depth of 15 cm (MPa)

SWP_50cm

Mean of soil water potential at a depth of 50 cm (MPa)

Time scale (abbreviation)

Description

Variable name_Annual

Annually

Variable name_Spring

Spring season (March, April, and May)

Variable name_Summer

Sumer season (June, July, and August)

Variable name_Fall

Fall season (September, October, and November)

Variable name_Winter

Winter season (December, January, and February)

Variable name_Jan, Feb, Mar, Apr,

Monthly: January to December

AirTmp_1.3m_open

May, Jun, Jul, Aug, Sep, Oct, Nov, Dec
a

Variables measured in climate station opening west of plot A3 (mixedwood treatment). Area

maintained in open condition by annual brushing; size is approximately 20 × 20 m. The other
variables were measured within plots A2 (pure spruce treatment) and A3 (mixedwood treatment).
b

Values for height indicate height above the ground surface (regardless of whether mineral soil or

organic material). Values for soil depth indicate depth from the mineral soil forest floor interface.

22

The response variable was the standardized ring widths. I determined the mean of
ring-width series of the north and west tree cores to obtain a single measurement per tree
(package dpIR, Bunn 2008). Then, I standardized (detrended) ring-widths of each tree using
the dendrochronology program library (package dpIR, Bunn 2008) in the Statistical Software
R (R Development Core Team 2011). I applied a smoothing spline curve to accentuate the
climate-related signal by reducing the effects of stand dynamics (e.g. competition and
disturbance). I applied the smoothing spline curve to the response variable ring-widths with
rigidity determined by one parameter: frequency response f at a wavelength of an average of
20 years. I chose to use the smoothing spline since it showed a better fit to my data compared
to other detrending methods. The smoothing spline is a function defined by piecewise
polynomials, in other words, the curve that best fits the data. As the spline is so flexible and
fits the data so well, there is the risk of inadvertently removing climate effects on growth
(Sullivan et al. 2016). Standardization transforms the raw ring width data into ring width
index (rwi) values. Ring width index values do not refer to absolute measurements of growth.
Instead, rwi represents relative radial growth rates fluctuating around a mean of 1.0.
To fit white spruce growth to climatic predictor variables, I used a two-step
procedure. First, I used Random Forest (Liaw and Wiener 2002) analyses to identify which
climate variables were most important to predict tree growth using the full data set, and each
treatment individually. Second, the best climate predictor variables were evaluated by fitting
tree growth to selected explanatory climate variables using a linear mixed effect model
framework.
To minimize concerns of using climate data from Block A and applying it to data
from Blocks A to E, before the model selection, I analyzed the amount of variance that was
captured by the random effect versus the fixed effects. I found that across the board the plot
23

random effects accounted for less than 2% of the variation of the model residuals. I also
compared the Akaike information criterion (AIC) between a model without blocks as a
random factor and a similar model with blocks (block A to E) using all data set, and for each
treatment. I found that for all the cases there was no significant difference in AIC values
between the model with and without block as random factor.
I performed the Random Forest analysis on three sets of predictor variables (annual,
seasonal, and monthly), and whether or not it used the full data set (i.e. all data) or across the
treatment types. Due to the low accuracy of rainfall data from October through April, I
calculated precipitation as the sum of rainfall in the growing season months (May to
September) only. I evaluated Random Forest regression trees using the Cran R package
“randomForest” (Liaw and Wiener 2002), with 10,000 runs each. I used the variable’s mean
depth from the 10,000 runs to rank the most important variables.
To identify the important annual climate variables for annual spruce growth, I applied
Random Forest using the full data set that included tree growth of PS and MW (total of 10
plots) as response variables. For the climate predictor variables, I used 17 annual climate
variables, including those measured in the climate station opening, plots A2, and A3. Then, I
chose the five most important distinct climate variables based on their mean depth ranking.
I fitted tree growth to the five climate variables using a linear mixed models
framework (R package nlme; Pinheiro et al. 2019) with and without two-way interactions to
determine the relationship between annual climate and annual spruce growth. I included
individual trees (n = 6) nested within plots (r = 5) nested within treatments (t = 2) in all
models as random factors. I performed a Multimodel selection using the MuMIn R package
(Bartoń 2019). Thus, all potential models that could be generated using the five selected
explanatory climatic predictor variables were evaluated and ranked according to their AIC. I
24

applied model averaging (Burnham and Anderson 2002) to estimate coefficients for the
explanatory predictor variables. Model averaging includes just those candidate models whose
AIC delta value lies within four of the best model.
To identify the important annual, seasonal and monthly climate variables for annual
spruce growth in pure and mixed stands, I performed the Random Forest analysis on three
sets (annual, seasonal, monthly) of predictor variables for tree growth on each treatment type.
I used the climate variables measured on plot A2 for tree growth in the PS treatment (total of
5 plots), and the climate variables measured on plot A3 for tree growth in the MW treatment
(total of 5 plots). Due to the large number of climate variables, I selected the five most
important variables in each set by scanning the 20 variables’ mean depth ranking (Appendix
7, Figures 7.1-7.7).
To compare spruce growth response to the predictor variables in PS versus MW, I
selected explanatory climate variables to include in linear mixed effects tree growth models
for each treatment type based on the Random Forest results and my evaluation of the
ecological relevance of the potential climatic predictor variables. Restricting the model to
only a few explanatory variables would reduce the “noise” related to the extensive data set of
climate variables analyzed and related issues of collinearity between similar independent
variables. This avoids overparameterizing the model giving the limited number of
observations of tree growth index available. When the same variable but different
measurement positions (e.g. SWP_15 and SWP_50) presented a similar rank in the minimal
depth, I chose the variable that contains the most complete data set for the analyses.
I fitted a mixed effect model with and without two-way interaction terms using the
five annual, four seasonal, and four monthly predictor variables for MW and PS separately. I
included individual trees (n = 6) nested within plots (r = 5) in all models as the random
25

factor. I also applied the previous multimodel inference and model averaging procedure to
analyze the data. However, all potential models that could be generated using the selected set
of explanatory climate variables were evaluated and ranked according to their AIC.
I analyzed individual (i.e. direct) climate effects on spruce growth by examining
whether coefficients generated by the mixed effect models were positive or negative. To
examine interactions among climate variables on spruce growth, I plotted the marginal means
interaction from the models (R package emmeans: Lenth et al. 2020). My interpretation was
based on the direction and strength (slope) of the interaction.

2.3. Results
2.3.1. Identification of important predictor variables for annual white spruce
growth
For the full data set, the Random Forest analysis found that annual soil temperature,
the number of days with air temperature above 5°C, air temperature, solar radiation and soil
water potential (SWP) were the five most important annual predictor variables for spruce
growth (Appendix 7.1, Table 7.1).
Random Forest results showed that PS and MW had four of the same top five annual
predictor variables (Appendix 7.1, Table 7.2). However, the order of importance of these
variables differed between them. In PS, Random Forest analysis selected annual rainfall as
the most important predictor variable, followed by the number of days with air temperature is
above 5°C, soil temperature, solar radiation, and air temperature. In MW, Random Forest
analysis selected annual solar radiation as the most important variable, followed by soil water
potential, air temperature, soil temperature, and rainfall. The two variables that differed

26

between these two treatments among the five top predictor variables were the number of days
with air temperature above 5°C for PS, and SWP for MW.
Random Forest analysis showed that the top five seasonal climatic predictor variables
differed across the treatment types (Appendix 7.1, Table 7.2). In PS, the five most important
variables were the number of days with air temperature above 5°C in the summer, air
temperature in the summer and fall, and soil temperature in the fall and winter. In contrast, in
MW, the variables include air temperature in the winter, solar radiation in the winter, the
number of fall days with air temperature below 0°C and the number of fall days with air
temperature above 5°C, and soil water potential in the spring.
Random Forest analyses also indicated different top five monthly climatic predictor
variables for the two treatments (Appendix 7.1, Table 7.2). In PS, the five most important
predictor variables were soil temperature in January, solar radiation in February, May and
August, and soil water potential in October. In MW, the variables were the number of days
when the air temperature was above 5°C in November and September, rainfall in September,
solar radiation in December and August.
In summary, Random Forests analysis indicated that four out of five most important
annual variables for predicting white spruce growth were the same across the treatment types
but that the five most important variables differed when using seasonal and monthly
variables. Random Forests results indicated that solar radiation, soil temperature, air
temperature, and soil water potential were important not only between years but also within
the year, especially in spring, summer and fall.

27

2.3.2. Annual climate and tree growth relationships using full data set
The best full data set model without interaction (i.e. individual effects) included all
five top annual variables (Table 2.5). Higher air temperature and solar radiation were related
to increased spruce growth, while higher number of days with air temperature above 5°C,
SWP and higher soil temperature reduced spruce growth. The best annual interaction model
included all five individual variables and seven two-way interactions (Table 2.5; Appendix
7.2, Table 7.3, Figures 7.8-7.14).
Increases in annual air temperature negatively influenced spruce growth when annual
number of days with air temperature above 5°C was high, and positively influenced spruce
growth when annual number of days with air temperature above 5°C was low (Appendix 7.2,
Figure 7.8). The increase in air temperature positively influenced spruce growth when the
annual soil temperature was high, and negatively influenced spruce growth when soil
temperature low (Appendix 7.2, Figure 7.9), and the same was true for high and low solar
radiation (Appendix 7.2, Figure 7.10).
The increase in the annual number of days with air temperature above 5°C negatively
influenced spruce growth when the annual solar radiation was low, and positively influenced
spruce growth when solar radiation was high (Appendix 7.2, Figure 7.11). When annual solar
radiation was low, increases in annual soil temperature had a positive influence on spruce
growth. However, when annual solar radiation was high, increases in annual soil temperature
had a negative influence on spruce growth (Appendix 7.2, Figure 7.12).
Increases in annual soil temperature had a positive influence on spruce growth when
annual SWP was high and a strong negative influence on spruce growth when annual SWP
was low (Appendix 7.2, Figure 7.13). Increases in annual solar radiation decreased spruce

28

growth when annual SWP was low, and increased spruce growth when annual SWP was high
(Appendix 7.2, Figure 7.14).

29

Table 2.5. Mixed effect growth model structure and fitted coefficient values of the annual climatic
predictor variables for the full data set. Coefficient values were derived using model averaging that
included all candidate models that were within 4 AIC values of the best model. The variables selected
for inclusion in the best model are highlighted in bold, with the gray shading for variables with a
negative effect on tree growth.
Full data set
Annual variable (without interaction)

Coefficient

Intercept

1.3968

AirTmp_3.0m_Annual

0.0257

ndays_AirTmp5_1.3m_Annual

-0.0026

SoilTmp_50cm_Annual

-0.0764

SolRad_3.0m_open_Annual

0.0039

SWP_15cm_Annual

-0.3308

Annual variable (with interaction)

Coefficient

(Intercept)

2.0984

AirTmp_3.0m_Annual

0.1838

ndays_AirTmp5_1.3m_Annual

-0.0130

SoilTmp_50cm_Annual

0.2585

SolRad_3.0m_open_Annual

-0.0358

SWP_15cm_Annual

-6.5810

AirTmp_3.0m_Annual:ndays_AirTmp5_1.3m_Annual

-0.0046

AirTmp_3.0m_Annual:SoilTmp_50cm_Annual

0.0319

AirTmp_3.0m_Annual:SolRad_3.0m_open_Annual

0.0082

ndays_AirTmp5_1.3m_Annual:SolRad_3.0m_open_Annual

0.0004

SoilTmp_50cm_Annual:SolRad_3.0m_open_Annual

-0.0069

SoilTmp_50cm_Annual:SWP_15cm_Annual

0.9475

SolRad_3.0m_open_Annual:SWP_15cm_Annual

0.0673

AirTmp_3.0m_Annual:SWP_15cm_Annual

-0.1058

ndays_AirTmp5_1.3m_Annual:SWP_15cm_Annual

-0.0072

ndays_AirTmp5_1.3m_Annual:SoilTmp_50cm_Annual

0.0008

30

2.3.3. Climate and tree growth relationships across the treatments
Akaike information criterion values indicated that the best annual and inter-annual
climate predictors differed between PS and MW treatment. The estimated coefficients for the
explanatory climate variables (i.e. full model) using model averaging, indicated that annual
rainfall, SWP in the spring and summer, air temperature in May, and SWP in August differed
in respect to their individual direct effects (positive or negative) on tree growth between the
two treatments (Tables 2.6-2.7).

2.3.3.1.

Annual climate and tree growth relationships across the

treatments
Annual air temperature, rainfall, soil temperature and solar radiation were important
predictor variables included in both the PS and MW models, whereas annual SWP was only
included in the best PS model. Rainfall was the only climate variable that differed regarding
its influence on tree growth between the two treatments. Higher annual rainfall positively
influenced tree growth in PS and negatively influenced tree growth in MW (Table 2.6;
Appendix 7.3, Table 7.4-7.5).
The best interaction model for PS and the best model for MW included all the
individual variables and eight two-way interactions, but differed regarding two of their
interactions (Table 2.6). The plotted interactions between the annual climate variables on
spruce growth for PS and MW is in Appendix 7.3 (Figures 7.15-7.22) and Appendix 7.3
(Figures 7.23-7.29), respectively.
The increase in annual air temperature positively influenced spruce growth when
annual rain was low or high in both treatments. When the annual soil temperature was low or
high in PS, the increase in annual air temperature had a positive effect on spruce growth, with
31

a stronger positive influence on tree growth when annual soil temperature was high. On the
other hand, in MW, the increase in annual air temperature negatively influenced spruce
growth when the annual soil temperature was low, and positively influenced spruce growth
when soil temperature was high.
In PS, the increase in annual air temperature positively influenced spruce growth
when annual solar radiation was low, and negatively influenced spruce growth when annual
solar radiation was high. The interaction between annual air temperature and solar radiation
was not included in the best model for MW.
The increase in annual air temperature positively influenced spruce growth in PS
when annual SWP was high or low, with a stronger positive influence when annual SWP was
low. Conversely, in MW, the increase in annual air temperature positively influenced spruce
growth when annual SWP was low, and negatively influenced spruce growth when SWP was
high.
The increase in annual rainfall had a positive influence on spruce growth when annual
soil temperature was low and a strong negative influence on spruce growth when soil
temperature was high in PS. In MW, the increase in annual rainfall had a positive influence
on spruce growth when annual soil temperature was low or high, with a stronger positive
influence when annual soil temperature was high.
In MW, the increase in annual rainfall positively influenced spruce growth when solar
radiation was low, and negatively influenced spruce growth when solar radiation was high.
The interaction between annual rainfall and solar radiation was not included in the best
model for PS.
The increase in annual rainfall had a negative relationship with spruce growth in PS
when annual SWP was low or high, with a higher negative relationship when annual SWP
32

was high. The interaction between annual rainfall and SWP was not included in the best
model for MW. In MW, the increased in annual soil temperature positively influenced spruce
growth when annual solar radiation was low or high, with a stronger positive influence when
solar radiation was low. The interaction between annual soil temperature and solar radiation
was not included in the best model for PS.
In PS, the increase in annual soil temperature had a negative influence on spruce
growth when annual SWP was low, and had a positive influence on spruce growth when
annual SWP was high. In contrast, in MW, when annual SWP was either low or high, annual
soil temperature positively influenced spruce growth, with a greater positive influence on
spruce growth when annual SWP was low. For both treatments, the increase in annual solar
radiation positively influenced spruce growth when annual SWP was low, and negatively
influenced spruce growth when annual SWP was high.

33

Coefficient
0.8158
0.0251
0.0006
-0.0844
0.0028
-0.9732
Coefficient
-7.3413
0.8625
0.0089
1.8948
0.0598
-21.3316
-0.0006
0.0621
-0.0180
-0.7196
-0.0028
0.00001
-0.0113
-0.0066
7.9998
-0.0492

Pure Spruce
Annual variable (without interaction)
(Intercept)
AirTmp_3.0m_Annual
Rain_open_Annual
SoilTmp_50cm_Annual
SolRad_3.0m_open_Annual
SWP_15cm_Annual

Annual variable (with interaction)
(Intercept)
AirTmp_3.0m_Annual
Rain_open_Annual
SoilTmp_50cm_Annual
SolRad_3.0m_open_Annual
SWP_15cm_Annual
AirTmp_3.0m_Annual:Rain_open_Annual
AirTmp_3.0m_Annual:SoilTmp_50cm_Annual
AirTmp_3.0m_Annual:SolRad_3.0m_open_Annual
AirTmp_3.0m_Annual:SWP_15cm_Annual
Rain_open_Annual:SoilTmp_50cm_Annual
Rain_open_Annual:SolRad_3.0m_open_Annual
Rain_open_Annual:SWP_15cm_Annual
SoilTmp_50cm_Annual:SolRad_3.0m_open_Annual
SoilTmp_50cm_Annual:SWP_15cm_Annual
SolRad_3.0m_open_Annual:SWP_15cm_Annual

Annual variable (with interaction)
(Intercept)
AirTmp_3.0m_Annual
Rain_open_Annual
SoilTmp_50cm_Annual
SolRad_3.0m_open_Annual
SWP_15cm_Annual
AirTmp_3.0m_Annual:Rain_open_Annual
AirTmp_3.0m_Annual:SoilTmp_50cm_Annual
AirTmp_3.0m_Annual:SolRad_3.0m_open_Annual
AirTmp_3.0m_Annual:SWP_15cm_Annual
Rain_open_Annual:SoilTmp_50cm_Annual
Rain_open_Annual:SolRad_3.0m_open_Annual
Rain_open_Annual:SWP_15cm_Annual
SoilTmp_50cm_Annual:SolRad_3.0m_open_Annual
SoilTmp_50cm_Annual:SWP_15cm_Annual
SolRad_3.0m_open_Annual:SWP_15cm_Annual

Mixedwood
Annual variable (without interaction)
(Intercept)
AirTmp_3.0m_Annual
Rain_open_Annual
SoilTmp_50cm_Annual
SolRad_3.0m_open_Annual
SWP_15cm_Annual

34

Coefficient
1.5043
-2.6123
0.0043
-0.8173
0.1085
-0.9248
0.0014
0.3508
0.0018
-6.1089
0.0004
-0.0002
-0.0013
-0.0060
-1.8932
0.3503

Coefficient
1.2172
0.0301
-0.0002
-0.0879
0.0020
-0.1356

variables selected for inclusion in the best model are highlighted in bold, with the gray shading for variables with a negative effect on tree growth.

Coefficient values were derived using model averaging that included all candidate models that were within 4 AIC values of the best model. The

Table 2.6 Mixed effect growth model structure and fitted coefficient values of the annual climatic predictor variables across the treatment types.

2.3.3.2.

Seasonal climate and tree growth relationships across the

treatments
When using the seasonal climatic predictor variables in the mixed effect models, I
found that PS and MW differed regarding the variables included in the best model (Table 2.7;
Appendix 7.4, Table 7.6). The best model of PS included air temperature during the spring
and SWP during the summer. In contrast, the best model of MW included air temperature
during the summer and SWP during the spring.
The only variable that differed between the two treatments regarding its influence on
tree growth is the spring and summer SWP, which positively influenced tree growth in PS
and negatively influenced tree growth in MW. Air temperature in the spring showed a
positive relationship with tree growth, where air temperature in the summer showed a
negative relationship with tree growth in both treatments.
When adding an interaction term in the mixed effect model using the seasonal
climatic predictor variables, the best model for PS included the air temperature during the
spring, SWP during the spring and summer and the interaction between air temperature and
SWP during the spring. The best model for MW included all the variables and interactions of
the initial full model structure (Table 2.7; Appendix 7.2.3, Table 7.6). The plotted
interactions between the air temperature and SWP variables during spring and summer on
spruce growth for PS and MW are in Appendix 7.4, Figures 7.31 and Figure 7.32,
respectively.
In PS, the increase in air temperature during the spring had a slightly negative
influence on spruce growth when SWP in the spring was high and a strong positive influence
on spruce growth when SWP in the spring was low. I found a similar interaction between air
35

temperature and SWP during the summer in PS, but this interaction was not included in the
best model.
In MW, the increase in air temperature in the spring positively influenced spruce
growth when SWP in the spring was high, and negatively influenced spruce growth when
SWP in the spring was low. In contrast, the increase in air temperature during the summer in
MW had a negative influence on spruce growth when SWP was either high or low, with a
stronger negative influence when SWP was low.

36

Coefficient
1.0207
0.0151
0.0494
0.1030
-0.0019
Coefficient
1.1300
-0.0035
0.3946
-0.0084
0.5594
-0.0967
-0.1166

Pure Spruce
Seasonal variable (without interaction)
(Intercept)
AirTmp_3.0m_Spring
SWP_15cm_Spring
SWP_15cm_Summer
AirTmp_3.0m_Summer

Seasonal variable (with interaction)
(Intercept)
AirTmp_3.0m_Spring
SWP_15cm_Spring
AirTmp_3.0m_Summer
SWP_15cm_Summer
AirTmp_3.0m_Spring:SWP_15cm_Spring
AirTmp_3.0m_Summer:SWP_15cm_Summer

Seasonal variable (with interaction)
(Intercept)
AirTmp_3.0m_Spring
SWP_15cm_Spring
AirTmp_3.0m_Summer
SWP_15cm_Summer
AirTmp_3.0m_Spring:SWP_15cm_Spring
AirTmp_3.0m_Summer:SWP_15cm_Summer

Mixedwood
Seasonal variable (without interaction)
(Intercept)
AirTmp_3.0m_Spring
SWP_15cm_Spring
SWP_15cm_Summer
AirTmp_3.0m_Summer

Coefficient
1.2374
0.2180
-5.6627
-0.0660
-3.5513
1.8833
0.2510

Coefficient
1.3947
0.0019
-0.6575
-0.0309
-0.0341

37

variables selected for inclusion in the best model are highlighted in bold, with the gray shading for variables with a negative effect on tree growth.

Coefficient values were derived using model averaging that included all candidate models that were within 4 AIC values of the best model. The

Table 2.7 Mixed effect growth model structure and fitted coefficient values of the seasonal climatic predictor variables across the treatment types.

2.3.3.3.

Monthly climate and tree growth relationships across the

treatments
Across the two treatments, SWP during May and air temperature during August were
both variables included in the best model for PS and the best model for MW (Table 2.8;
Appendix 7.5, Table 7.7). Both SWP during May and air temperature in August showed a
negative relationship with spruce growth in both treatment types. Air temperature during
May and SWP in August variables differ between the two treatments regarding their
inclusion in the best model and their relationship with spruce growth.
Air temperature during May was only included in the best model of PS, with a
positive relationship with spruce growth. In contrast, air temperature in May in MW
presented a negative relationship with tree growth. Soil water potential in August was only
included in the best model of MW and presented a positive relationship with spruce growth.
In contrast, SWP during August in PS had a negative relationship with spruce growth.
When adding an interaction term in the mixed effect model, air temperature in May is
the only variable that differed between the best models of each treatment since it was just
included in the best model of PS. The interaction between air temperature and SWP in
August was included in the best model of PS and the best model for MW and showed a
similar relationship with spruce growth. The plotted interactions between air temperature and
SWP variables in May and August for PS and MW on spruce growth are in Appendix 7.5,
Figure 7.33 and Figure 7.34, respectively.
In PS, the increase in air temperature in August negatively influenced spruce growth
when SWP was high, and positively influenced spruce growth when SWP was low.

38

Similarly, in MW, the increase in air temperature in August had a negative influence on
spruce growth when SWP was high, and had a positive influence when SWP was low.
In PS, the increase in air temperature in May in PS positively influenced spruce
growth when SWP was low or high, with a stronger increase when SWP was high. In
contrast, the increase in air temperature in May in MW had a positive influence on spruce
growth when SWP in May was high, and had a negative influence on spruce growth when
SWP was low. However, none of the best models included the interaction between air
temperature and SWP in May.

39

-0.1252
-0.0110
-0.0075
Coefficient
1.4659
0.0268
-0.2557
-0.0535
1.2265
0.0355
-0.0875

Monthly variable (with interaction)
(Intercept)
AirTmp_3.0m_May
SWP_15cm_May
AirTmp_3.0m_Aug
SWP_15cm_Aug
AirTmp_3.0m_May:SWP_15cm_May
AirTmp_3.0m_Aug:SWP_15cm_Aug

0.9675
0.0157

Coefficient

Pure Spruce
Monthly variable (without interaction)
(Intercept)
AirTmp_3.0m_May
SWP_15cm_May
AirTmp_3.0m_Aug
SWP_15cm_Aug
Monthly variable (with interaction)
(Intercept)
AirTmp_3.0m_May
SWP_15cm_May
AirTmp_3.0m_Aug
SWP_15cm_Aug
AirTmp_3.0m_May:SWP_15cm_May
AirTmp_3.0m_Aug:SWP_15cm_Aug

Mixedwood
Monthly variable (without interaction)
(Intercept)
AirTmp_3.0m_May
SWP_15cm_May
AirTmp_3.0m_Aug
SWP_15cm_Aug

Coefficient
1.9660
0.0045
-1.3066
-0.0767
1.4374
0.0776
-0.1028

Coefficient
1.0358
-0.0029
-1.3338
-0.0125
0.0372

40

variables selected for inclusion in the best model are highlighted in bold, with the gray shading for variables with a negative effect on tree growth.

Coefficient values were derived using model averaging that included all candidate models that were within 4 AIC values of the best model. The

Table 2.8 Mixed effect growth model structure and fitted coefficient values of the monthly climatic predictor variables across the treatment types.

2.4. Discussion
The positive relationship between annual air temperature and annual solar radiation
on white spruce growth found in the model using the full data set and across the treatments is
not consistent with much published literature in white spruce growth in the boreal zone of
B.C. (Cortini et al. 2011). Another study in the Alaskan boreal forest using mature and old
stands found that radial growth of white spruce had decreased with increasing temperature,
concluding that temperature-induced drought stress is reducing white spruce productivity at
northern latitudes (Barber et al. 2000). The negative relationship between annual SWP and
spruce growth was also surprising since I expected that more soil water available in a year
would positively influence spruce growth. At treeline sites in interior Alaska, better growth
rates were found with cooler, wetter years (Lloyd et al. 2013), and growth declines in
response to warming temperatures were more common in warmer and drier parts of the
boreal forest (Lloyd and Fastie 2003). Another unexpected result found in the model using
the full data set is the negative relationship between the annual number of days when the air
temperature was above 5°C and spruce growth. I expected tree growth to occur when the air
temperature is above 5°C, and this is a well-known threshold used among researchers to
calculate growing degree days (Cortini et al. 2011).
Rainfall was the only annual climate variable for which the relationship with spruce
growth differed by treatment, with a positive relationship in pure spruce stands and a
negative relationship in the mixedwood stands. Based on past white spruce studies, I
expected that the increase in annual rainfall would positively influence spruce growth in both
treatment types. It is unclear why spruce growth is responding negatively to annual rainfall in
the mixedwood stand. A possible explanation is that high annual rainfall also means cloudy
41

days, accentuating the shading of spruce growing under a deciduous canopy (willow-aspen
canopy), further reducing access to light. However, since the rainfall measurement was taken
in the open area and not within each treatment type, the SWP variable may better inform
water availability to trees within each treatment.
When using seasonal and monthly climate variables, my study indicated that the
influence of SWP and air temperature on spruce growth varies throughout the year. Together
with the interpretation of the relationship between annual climate variables and spruce
growth, my results suggest that seasonal and monthly climate predictor variables are more
suitable to understand annual tree growth responses to climate than annual climate variables.
Other studies also concluded that finer temporal scale better explain tree growth than more
coarse climatic variables. For instance, Thomson and Parker (2008) found that climate
variables in a monthly level (i.e. August minimum temperature and January maximum
temperature) were better correlated with jack pine (Pinus banksiana Lamb.) than seasonal
and annual averages by scanning a total of 65 climate variables at annual, seasonal and
monthly levels. Cortini et al. (2011) showed that monthly climate variables had a stronger
relationship to the growth of lodgepole pine and white spruce than seasonal and annual
variables.
The relationships found between air temperature and spruce growth during spring and
summer are consistent with previous studies. In both stand types, warm springs increase
spruce growth, and warm summers decrease spruce growth. A study of white spruce growth
at treeline areas in Alaska indicated that high mean temperatures in July decreased the
growth of white spruce, whereas warm springs increased tree growth (Wilmking et al. 2004).
Barber et al. (2000) showed that ring-width chronologies of white spruce were strongly
negatively correlated with summer monthly mean temperatures in the interior of Alaska, and
42

the negative relationship was related to reduced CO2 uptake and higher water loss during
photosynthesis. The increase in air temperature tends to increase evapotranspiration from soil
and plant tissues, inducing stomatal closure and reducing net photosynthesis to minimize
water losses in response to moisture stress (Kozlowski and Pallardy 2002).
An interesting finding of my study is the different influences of SWP during spring
and during summer on spruce growth in the two stand types. Spruce growth in pure stands
has a positive relationship with SWP during spring and summer, while spruce growth in
mixedwoods has a negative relationship. When comparing microsites, mixedwood stands
have the lowest SWP during the summer. I infer that the negative relationship between
spruce growth and SWP during spring and summer in the mixedwoods stands might be
consistent with a drought stress mechanism. Spruce trees might not be benefiting from the
water available in the soil, and competing trees might be the main factor influencing the
availability of this resource. A recent study in Alaska indicated that, in early spring,
deciduous trees are capable of taking up 21–25% of snowmelt water while conifers take up
less than 1% during that period (Young-Robertson et al. 2016). Moreover, a previous study at
Inga Lake and nearby sites (Cortini et al. 2011) suggested that controlling unwanted
vegetation in white spruce stands enhanced white spruce growth by increasing resource
availability, and that drought stress related to warmer and drier summers could make
vegetation stronger competitors for water in white spruce stands.
The interaction between air temperature and SWP during spring on spruce growth in
the mixedwoods indicates that spruce growth positively responds to an increase in spring air
temperature as long there is sufficient water available in the soil (i.e. high SWP). Otherwise,
spruce growth is negatively influenced by the increase in air temperature. Early in the
growing season, soil moisture can be sufficient for growth even at higher temperatures. Soil
43

moisture usually dwindles as summer progresses, and the lack of moisture becomes more
stressful for tree growth at high temperatures (D’Arrigo et al. 2004).
The interaction between air temperature and SWP during summer on spruce growth
in the mixedwood stand might indicate drought stress. White spruce growing in the
mixedwood stand might be suffering from drought stress due to the increase in air
temperature even when SWP is high, with a stronger negative effect of the increase in air
temperature on spruce growth when SWP is low. This would indicate that warm temperatures
have a negative effect on tree growth (ring width) in the absence of sufficient availability of
water in the soil. My results suggest that warming without a concurrent increase in
precipitation might negatively influence white spruce growth growing in mixedwood stands.
The interactions between air temperature and SWP during the spring and summer in
pure spruce stands was surprising since I expected a similar interaction between these
climates variables in both stand types. It is unclear why the increase in air temperature
positively increases growth in pure spruce stands when SWP is low and negative influences
growth when SWP is high.
Another interesting result is the positive relationship found between SWP in August
and spruce growth in the MW, even though August was the month with the lowest SWP in
the MW. A possible explanation is that drought stress might cause deciduous trees to lose
their leaves prematurely, allowing white spruce to benefit from the water available in the soil.
In aging boreal mixedwoods, canopy openings resulting from senescence of broadleaved
trees allow for the establishment and release of shade tolerant slower-growing conifers
(Brassard and Chen 2006).
To the best of my knowledge, there is a lack of studies on the influence of monthly
climate variables on annual spruce growth to support my findings regarding the interaction
44

between air temperature and SWP in May and August on spruce growth in each treatment
type. Analysis using the same timeframe for both response and predictor variables would be
helpful to understand spruce growth responses to climate variables throughout the year. I also
argue that we can better understand how trees respond to climate by evaluating the
interactions between climate variables instead of the direct effects of individual climate
variables. In both pure and mixedwood stands, there is an interplay between the amount of
water available in the soil and air temperature to influence annual tree white spruce growth.
My work agreed with others that the growth of spruce is dependent on within
seasonal variability in climate variables (Cortini et al. 2011, Lloyd et al. 2013). However, my
results also indicate that the order of importance of the influence of microclimate variables
on annual spruce growth differs between the two stands. The best annual and inter-annual
microclimate predictors differ between pure and mixedwood. Moreover, my study suggests
that spruce growth responses to microclimate variability depend on whether it is growing in
pure or mixewood stands. A key finding of my work is that stand composition and structure
are important determinants of how annual white spruce growth responds to yearly
fluctuations in seasonal air temperature and SWP variables, and how annual spruce growth
will respond to projected future climate scenarios. My results indicate that spruce wood
production in both stand types could decrease in northeastern British Columbia with a
combination of warmer temperatures (that is, increased evapotranspiration) and drought
during summer (that is, decreased water supply). Spruce growth in mixedwood stands might
be particularly negatively affected by drought stress due to competition with deciduous trees
for water.

45

2.5. Conclusion
The response of white spruce growth to microclimate variability depends on stand
composition and structure. Drought stress is likely to limit boreal white spruce productivity
under warmer future climates in northeastern British Columbia. Spruce growth in
mixedwood stands might be more sensitive to drought stress than in pure spruce stands due to
the higher competition for limiting resources. Seasonal and monthly climate predictor
variables are more suitable to understand annual tree growth responses to climate than annual
climate variables. More studies on intra-annual tree growth and climate relationships are
needed to inform better how trees will respond to future climatic shifts.

46

3. Chapter 3: White spruce sap flow sensitivity to climate variability in pure and
mixedwood stands

Abstract: Drought-induced water stress is one of the main contributors of widespread tree
mortality and growth decline in the western boreal forests of Canada. Using data from 2007
to 2018, I evaluated the importance and the influence of microclimate variables on sap flow
of white spruce (Picea glauca (Moench) Voss) trees throughout the entire growing season in
pure and mixedwood stands located at the Inga Lake site, north-eastern British Columbia.
Using a model selection framework, I evaluated which climatic variables were the most
important drivers of sap flow in pure and mixedwood stands. Sap flow responses to climate
variables within the growing season differed between pure stand and mixedwood stands. My
analysis also indicated that interactions between climate variables, primarily moisture
availability and air temperature, were important for sap flow. The response of white spruce
sap flow to microclimate variability depended on stand composition and structure, and
changed throughout the growing season. Early in the growing season, sap flow is primarily
limited by air temperature, while by mid-summer, drought stress is the key limiting variable.
Drought stress is likely to limit boreal white spruce growth under warmer, future climates in
pure and mixedwood stands in northeastern British Columbia.
Keywords: Sap flow, climate variability, white spruce, drought.

3.1. Introduction
Drought-induced water stress is one of the main contributors of widespread tree
mortality and growth decline in the western boreal forests of Canada (Hogg et al. 2008, Ma et
al. 2012). Previous studies (McDowell 2011) have shown that drought related tree mortality
47

primarily occurs due to hydraulic failure or carbon starvation (Galvez et al. 2011, Plaut et al.
2012, Kulaç et al. 2012, Barigah et al. 2013, Sevanto et al. 2014, Mitchell et al. 2014).
Hydraulic failure is widely recognized as the major cause of woody plant mortality during
drought (McDowell 2011, Choat 2013), while carbon starvation is expected to occur in the
late stages of prolonged drought (McDowell 2011, Kulaç et al. 2012). Low water potential
occurs due to a decrease in soil water content or increased transpiration rate. Low water
potential can impede long-distance water transport and induce cavitation embolism or
dehydration (Sperry 2000, Vilagrosa et al. 2012, Pangle et al. 2015). High levels of embolism
tend to impair water supply to the foliage and lead to tissue desiccation (Dietrich et al. 2019).
As a defense against hydraulic failure, the tree can enter carbon starvation through stomatal
closure, reducing photosynthesis and eventually resulting in a shortage of carbohydrate
metabolites in different tree tissues during prolonged drought (Pangle et al. 2015, Dietrich et
al. 2019).
How trees respond to drought stress depends on species, age, size, competition, and
site conditions (Lloret et al. 2011, Pretzsch and Dieler 2011, Zang et al. 2012). Norway
spruce (Picea abies) and Swiss pine (Pinus cembra) reduce sap flux during drought periods
in an attempt to conserve water (Anfodillo et al. 1998). In contrast, European larch (Larix
decidua) can sustain a relatively high sap flux during drought because of its high water
uptake capacity (Anfodillo et al. 1998). European beech (Fagus sylvatica) can maintain
transpiration under higher soil moisture tension (Pretzsch et al. 2013). However, tree
responses to water stress in pure and mixedwood stands in boreal forests are not well
understood.
There is a lack of information on how mixing of species modifies tree growth under
drought stress compared with their performance in a monospecific environment (Pretzsch et
48

al. 2010, Richards et al. 2010). Past studies have often not included physiological or
hydrological measurements at the individual tree level in pure and mixed stands (Pretzsch et
al. 2013), but have evaluated drought stress in a coarse scale, often using annual tree ring
measurements (Zang et al. 2012, Pretzsch et al. 2013). Assessing sap flow responses to
climate variability across different time scales while accounting for stand composition and
structure can provide the information needed to further increase our understanding in this
area.
Direct measures of sap flow allow a more direct evaluation of tree growth response
and are an effective method for testing biological and hydrological questions (Muñoz-Villers
et al. 2012, Steppe et al. 2015, Berry et al. 2017). Measurement of xylem sap flow via
thermal dissipation probes is one of the most commonly used methods for estimation of
whole-tree water use on forests (Swanson 1994, Granier et al. 1996, Ping et al. 2004).
Observed over time, the amount of sap flow upward through the stem equates with
transpiration at the leaves (Swanson 1994), and many studies have used sap flow
measurements to estimate transpiration (Alarcón et al. 2000, Wullschleger et al. 2001). Tree
transpiration indicates that stomata are open, which is an essential condition for gas exchange
during photosynthesis (Swanson 1994). In other words, water loss can be thought of as the
“price” the plant pays to keep its stomata open. Since transpiration and sap flow are very
closely related, most climate variables that affect transpiration are assumed to also affect sap
flow.
Previous studies clearly showed that climate influences sap flow rates, but there are
still some questions regarding which climate variables are most important in determining sap
flow across different time scales. Sap flow rates have been found to be positively correlated
with air temperature, soil temperature and soil moisture in the early growing season in a
49

warming experiment (Juice et al. 2016). A soil warming study found elevated spring time sap
flow in Norway spruce trees was induced by elevated soil temperatures (Bergh and Linder
1999), while other researchers found the rate of sap flow is positively correlated with air
temperature (Juhász et al. 2013, Chang et al. 2014, Juice et al. 2016). Other studies have
shown stronger relationships between daily sap flow rates and daily vapor pressure deficit
(VPD) compared to daily air temperature (Yin et al. 2004). However, most of the studies
assessed only the effect of individual climate variables on sap flow, in many cases assessing
averaged daily sap flow across the entire growing season.
Interaction models that account for how the climate variables interact with each other
to influence sap flow can be better predictors of sap flow rates across time. For example, the
positive influence of air temperature on sap flow can be magnified when there is greater soil
moisture content. A controlled experiment on potted plants, Camellia japonica L. and
Ligustrom japonica Thumb., showed that stomatal responses to light increases with air
temperature, and that the combination of high air temperature and high vapour pressure
deficit (VPD) limited stomatal opening (Wilson 1948). Sap velocity and VPD have been
found to be linked when soil moisture was high during the early growing season at lower
elevations or throughout the entire growing season at higher elevations (Looker et al. 2018).
However, when soil moisture decreased, VPD and sap velocity became decoupled, most
likely due to decreased stomatal conductance as a water conservation strategy (Looker et al.
2018). Many other studies explicitly highlighted the need for models that account for
interactions between climate variables to influence sap flow (Small and McConnell 2008).
Sap flow responses to climate variables largely depend on the plant species being
considered and the site conditions. Most of the variance in sap flow for a northern red oak
stand in Massachusetts was explained by air and soil temperature, with lesser amounts
50

explained by photosynthetically active radiation (PAR) and VPD (Juice et al. 2016). In
northern New Mexico, ponderosa pine sap flow was found to be correlated with soil
moisture, whereas at a higher elevation, Engelmann spruce sap flow was not clearly
correlated with soil moisture (Small and McConnell 2008). Sap flow for maritime pine stands
in southwestern France was found to be correlated with soil moisture (Delzon and Loustau
2005). Another interesting result was the finding that sap flow and transpiration decreased
with the age of stand and that younger stands were more affected by drought than older
stands.
While many studies assess the relationship between sap flow and individual climate
variables at a coarse time scale, such as between growing seasons, we know little about how
sap flow responds to within-year climate variability. Furthermore, it is still unclear if sap
flow responses to climate variability vary with stand composition and structure. I am not
aware of any studies examining the effects of microclimate variables on sap flow by young
white spruce trees (Picea glauca (Moench) Voss) in pure and mixed boreal forests in western
Canada.
Previous studies have concluded that spruce has isohydric characteristics and low
drought resistance; it reduces water consumption and growth in the early phase of drought
stress through stomata closure (Zang et al. 2012, Pretzsch et al. 2013, Sullivan et al. 2017).
Therefore, studying white spruce trees sap flow responses to climate within pure and mixed
stands is beneficial for interpreting their survival and growth. The objectives of this study
were to:
1. Identify and compare the important microclimate variables to predict spruce sap flow
in pure and mixedwood stands together and separated throughout the entirety of the
growing season.
51

2. Examine and compare the relationships between microclimate variables and sap flow
in pure and mixedwood stands together and separated throughout the entirety of the
growing season.
3. Identify and compare the important microclimate variables to predict spruce sap flow
in spring, early-summer and late-summer for each stand type.
4. Examine and compare the relationships between microclimate variables and spruce
sap flow in spring, early-summer and late-summer for each stand type.

By assessing individual white spruce sap flow responses to climate in the pure and
mixedwood stands, I provide information that will be useful in modeling and managing these
stands across western Canada in both current and future climate conditions.

3.2. Materials and methods
3.2.1. Study area description
(See section 2.2.1)
In this Chapter, I investigated sap flow in two experimental treatments units of the
Inga Lake research site: plot A2 in a pure white spruce treatment (PS), and plot A3 in a
mixedwood treatment (MW).

3.2.2. Microclimate data collection
Microclimate data were obtained from an on-site climate station installed at Inga
Lake. A variety of microclimate variables (Table 3.1; Appendix 9, Tables 9.1 and 9.3) were
measured from 2007 to 2018 every hour (standard time) at plots A2 and A3 and at the
climate station opening, and recorded on a data logger (models CR10X and CR10, Campbell
52

Scientific). At the climate station opening, microclimate variables included solar radiation
and rainfall. At PS (plot A2) and MW (plot A3), microclimate variables included air
temperature, soil temperature, and soil water potential.

Table 3.1 Description of equipment used to measure the microclimate variables.

a

Variable

Position (cm) b

Sensor make/model

Sensor type

Solar radiation a

+300

Li-Cor/LI200S

Silicone pyranometer

Rainfall a

+60 or +80

Sierra Misco/2501 or TE525m

Tipping bucket

Air temperature

+300

Home/fine wire 36AWG

Cu-Co thermocouple

Soil temperature

-15

Home built/twisted soldered wire

Cu-Co thermocouple

Soil water potential

-15

Campbell Sci/model 223

Gypsum Block

Variables measured in climate station opening west of plot A3 (mixedwood treatment). Area

maintained in open condition by annual brushing; size is approximately 20 × 20 m. The other
variables were measured within plots A2 (pure spruce treatment) and A3 (mixedwood treatment).
b

Values for height (+) indicate height (cm) above the ground surface (regardless of whether mineral

soil or organic material). Values for soil depth (-) indicate depth from the mineral soil forest floor
interface.

3.2.3. Sap flow measurements
At each treatment (PS and MW), three white spruce trees were selected to install the
sap flow sensor (model TDP-30, Dynamax). In 2018, the breast height of the three trees
selected in each treatment averaged 17.3 cm in the PS treatment, and 8.3 cm in the MW
treatment. Sap flow velocities were measured by the heat dissipation approach proposed by
Granier (1985). The Thermal Dissipation Probe (TDP) is a heat dissipation sensor that
measures the temperature of a line heat source implanted in the sapwood of a tree, referenced
to the sapwood temperature at a location below the heated probe.

53

Each sensor consists of two 1.2 mm diameter, 30 mm long stainless steel probes (i.e.
needles), a heated element and thermocouple junction above and only a thermocouple
junction below. I peeled off one piece of bark, and drilled two holes about 4 cm vertically
apart from each other at breast height (1.30 m) to insert the probes into the sapwood. The
heating wire in the upper probe was supplied with direct current of 3 V (0.15 to 0.2 W). The
sensor insertion site on each tree was protected with styrofoam eggs and wrapped with
aluminized bubble wrap to avoid thermal influences from solar radiation. Sap flow velocity
was measured on the north side of each spruce tree from 2007 to 2016, and on the south side
of each spruce tree in 2017 and 2018.
Daily maximum sap flow velocity (SFV) was based on the instantaneous
measurements taken every third hour (standard time, total of eight measurements a day) from
2007 to 2018 (April to September). However, SFV was not continually recorded every three
hours for all the trees for the entire study period, there were some gaps in the data due to
sensor damage in specific periods within each study year. The sap flow power was turned on
at the beginning of the hour, heating the upper probe, and was on for the entire hour, then a
measurement of the temperature difference between the two probes was taken at the end of
the hour and power was turned off (coming on again in 2 hours). The temperature differences
between the two probes were recorded by a data logger (model CR10X, Campbell Scientific
Inc.) with a multiplexer (AM16/32A, Campbell Scientific Inc.).
I calculated sap flow velocity with the following formula (Granier 1985):

= 0.119 [(∆

− ∆ )]/∆ ] .

(1)

54

Where

is the daily maximum sap flow velocity (mm/s), ∆

maximum temperature difference when

is the daily

is near 0 (i.e. no-transpiration state, typically at

night) and ∆ is the measured minimum temperature difference of eight instantaneous
measurement between the two probes (i.e. ∆ is the minimum value of the eight
measurements).

3.2.4. Statistical analysis
I summarized the hourly microclimate measurements from 2007 to 2018 into daily
means (Table 3.2).

Table 3.2. Abbreviation and description of the microclimate predictor variables used for the
analyses.
Variable (abbreviation)

Descriptionb

SolRad a
Mean of solar radiation at a height of 3.0 m (KW/m2)
a
Rain
Sum of precipitation at a height of 0.6 m or 0.8 m (mm)
AirTmp
Mean of air temperature at a height of 3.0 m (°C)
SoilTmp
Mean of soil temperature at a depth of 15 cm (°C)
Mean of soil water potential at a depth of 15 cm (MPa)
SWP
a
Variables measured in climate station opening west of plot A3 (mixedwood treatment). Area
maintained in open condition by annual brushing; size is approximately 20 × 20 m. The other
variables were measured within plots A2 (pure spruce treatment) and A3 (mixedwood treatment).
b

Values for height indicate height above the ground surface (regardless of whether mineral soil or

organic material). Values for soil depth indicate depth from the mineral soil forest floor interface.

I analyzed SFV and climate relationships in different time scales using data from
2007 to 2018. I removed the outliers (e.g. sensor failures) from the SFV and climate data
before the analyses (Appendix 9, Table 9.2). First, over the growing season (April to
September) by using the full SFV and climate data set (i.e. all data, PS and MW together)
55

and across the treatments (i.e. PS and MW separated). Second, seasonally by using the SFV
and climate data across the treatments divided in three seasons: spring (April and May),
early-summer (June and July), and late-summer (August and September) (Figure 3.1). I
analyzed the data in these three seasons based on a visual inspection of SFV trend over the
growing season for year.
To select the best model to predict SFV in PS and MW together and separated
throughout the growing season, and in each season for each treatment type, I used a model
selection framework. In other words, all potential models that could be generated using the
selected set of explanatory variables were considered. I created a full mixed effect model
without interaction terms (model 1) and a full mixed effect model with two-way interaction
terms (model 7), and by using a leave-one-climate-variable-out framework, I generated in
total 12 models (Table 3.3).
Then, I fitted SFV to the climate predictor variables using these 12 linear mixed
effects models. I fitted these linear mixed effects models using the R package nlme (Pinheiro
et al. 2019). For the full data set (PS and MW together) over the growing season, I included
individual trees (n= 3) nested within treatment (r= 2) in all 12 models as a random factor. For
each data set of PS and MW separated in each time scale, I included individual tree ID (n= 3)
as a random factor. I selected the best models by evaluating and ranking according to their
Akaike information criterion (AIC). The number of observations in the full data set was on
average 8000, with 4000 observations in each treatment. I did not include model 6 for my
analyses of the best model because this model excluded SWP, a variable with great
ecological relevance to sap flow.
Then, among these best models, I choose one model with and one model without twoway interaction terms to examine and compare the relationships between SFV and climate in
56

PS and MW together and separated throughout the growing season, and between the seasons
in each treatment type.
I analyzed individual (i.e. direct) climate effects on spruce SFV by examining
whether coefficients generated by the mixed effect models were positive or negative. To
examine interactions among climate variables affecting SFV, I plotted the marginal means
interaction from the models (R package emmeans: Lenth et al. 2020). My interpretation was
based on the direction and strength (slope) of the interaction.

57

Figure 3.1 Sap flow velocity of tree T5 loca
ted in the pure spruce treatment, and tree T1
located in
the mixedwood treatment over the 2007 grow
ing season. Vertical lines separate the seas
ons used in
the study: Spring (April and May), early -sum
mer (June and July), and late-summer (Aug
ust and
September).

58

Table 3.3 List of the models used for the model selection. Models 1 to 6 are the models without
interaction. Models 7 to 12 are the models with interaction terms.

Model
number

Variables

1

SolRad + Rain + SoilTmp + AirTmp +SWP

2

Rain + SoilTmp + AirTmp + SWP

3

SolRad + SoilTmp + AirTmp+ SWP

4

SolRad + Rain + AirTmp + SWP

5

SolRad + Rain + SoilTmp + SWP

6

SolRad + Rain + SoilTmp + AirTmp

7

SolRad + Rain + SoilTmp + AirTmp + SWP + SolRad*Rain +
AirTmp*SWP + SoilTmp*SWP + SolRad*SWP

8

SolRad + Rain + SoilTmp + AirTmp + SWP + AirTmp*SWP +
SoilTmp*SWP + SolRad*SWP

9

SolRad + Rain + SoilTmp + AirTmp + SWP + SolRad*Rain +
SoilTmp*SWP + SolRad*SWP

10

SolRad + Rain + SoilTmp + AirTmp + SWP + SolRad*Rain +
AirTmp*SWP + SolRad*SWP

11

SolRad + Rain + SoilTmp + AirTmp + SWP + SolRad*Rain +
AirTmp*SWP + SoilTmp*SWP

12

SolRad + Rain + SoilTmp + AirTmp + SWP + AirTmp*SWP +
SolRad*SWP

3.3. Results
3.3.1. Microclimate and sap flow comparison between the treatments
Visually comparing SFV for each year, I found PS trees typically had higher SFV
than MW trees on the same date. The possible reason for this is that spruce trees in MW are
small and overtopped by broadleaves (Figure 3.1).
59

In both treatments, SFV changed throughout the year. In general, SFV declined in
mid-summer (anytime from mid July to early August), and recovered in some seasons due to
rains (e.g. year 2007) but remained low in other seasons (e.g. year 2012). In year 2007, for
example, we can see a trend where sap flow appeared to decline due to low soil water
potential in mid-summer, and recovered again due to the increase in soil water potential
(Figure 3.2). Similar trends were also observed between air temperature and SFV (e.g. Figure
3.3). There are some seasons when rainfall and soil moisture appeared adequate and we do
not see the mid-summer decline (e.g. year 2013).
Daily microclimate also differed between months for each treatment, and I noticed
some similarities and differences between the treatments (Figure 3.4; Appendix 8.1, Figures
8.1 and 8.2). For instance, MW showed lower SWP compared to PS over the growing season,
with August and September being the months with the lowest SWP in MW (Figure 3.4).

60

Figure 3.2 Sap flow velocity of tree T5 located in the pure spruce treatment (PS) with soil water
potential (SWP) in the PS, and tree T1 located in the mixedwood treatment (MW) with SWP in the
MW over the 2007 growing season.

61

Figure 3.3 Sap flow velocity of tree T5 located in the pure spruce treatment (PS) with air temperature
in the PS, and tree T1 located in the mixedwood treatment (MW) with air temperature in the MW
over the 2007 growing season.

62

Figure 3.4 Comparison of daily mean of soil water potential (SWP) between pure spruce treatment
(in red) and mixedwood treatment (in green) from 2007 to 2018 using box plots. Soil water potential
ranges from 0 MPa to -1.5 MPa. Soil water potential of 0 MPa indicates that the soil is in a state of
saturation, increasingly negative values occur as the soil becomes drier and water less available for
the trees. The box plot visually shows the distribution of the data and skewness through displaying the
interquartile range (box), median (horizontal line), whiskers (vertical lines) and outlines (circles).

3.3.2. Model selection of daily sap flow throughout the growing season
The best models to predict SFV in the growing season using all data and across the
treatments were the models that included two-way interaction terms between variables that
represented light or temperature, and moisture availability (Table 3.4). The full model with
all two-way interactions (model 7) was the best model in PS. The model that excluded the
interaction between solar radiation and rainfall (model 8) was the best model when using all

63

data and in MW.
When comparing the AIC values across the models without interactions, the full
model that included all the climate variables (model 1) was the best model across the
treatments. In contrast, the model that excluded rainfall (model 2) was the best model when
using all data, but with a very small AIC difference from model 1.
I selected models 1 and 7 to examine and compare the relationship between SFV and
climate, using all data and across the treatments. These two models were chosen because
model 1 allowed the analysis of the direct effect of the climate variables in the SFV, and
model 7 the analysis of the effect of the interaction between the climate variables in SFV.
However, I focused more on the interpretation of model 7 since it was selected as the best
overall model and included all the interactions.

Table 3.4 Akaike information criterion (AIC) values of each of the 12 mixed effect models fitted
using the full data set and across the pure spruce and mixedwood treatments throughout the growing
season. The best models of SFV using the full data set and across the treatments are highlighted in
gray. The best models without interaction are highlighted in bold.

All
Model

Pure
Spruce

Mixedwood

AIC

AIC

AIC

1

-44498.19

-21028.38

-24374.46

2

-44498.90

-21026.23

3

-44369.80

4

All
Model

Pure
Spruce

Mixedwood

AIC

AIC

AIC

7

-44522.37

-21053.33

-24405.39

-24367.73

8

-44522.81

-21047.92

-24406.56

-20926.29

-24347.57

9

-44502.71

-21047.65

-24400.75

-44265.51

-20895.13

-24225.16

10

-44497.13

-21051.54

-24403.99

5

-43747.12

-20477.48

-24149.60

11

-44520.91

-21053.31

-24371.50

6

-58248.85*

-28336.10* -30796.23*

12

-44497.34

-21045.86

-24405.19

number

number

*Models excluded from my analyses of the best model because this model excluded SWP, known to have great
ecological relevance to sap flow.

64

3.3.3. Model selection of daily sap flow by season
The best models to predict SFV differed between the spring, early-summer and latesummer for each treatment (Table 3.5). In PS, the best model to predict SFV in spring was
model 1, which was the model without interactions that included all the variables (i.e. full
model). In contrast, the best model to predict SFV in early-summer was the model 11 that
excluded the interaction between solar radiation and SWP. In late-summer, the best model to
predict SFV was model 8 that excluded the interaction between solar radiation and rainfall. In
MW, models 1 and 8 were the best models to predict SFV in spring and early-summer,
respectively. The best model to predict SFV in late-summer was model 12 that excluded solar
radiation and rainfall.
To examine and compare the relationships between SFV and climate between seasons
in each treatment, I also selected models 1 and 7. However, here I also focused on the
interpretation of model 7. Selecting models 1 and 7 also allowed us to analyze the
relationships between SFV and climate when analyzing between seasons versus over the
growing season for each treatment.

65

-9382.48
-9324.23
-9337.87
-9140.58
-10805.76*
-9928.17
-9920.21
-9917.38
-9757.73
-9807.74
-11836.30*

-6745.79
-6712.60
-6743.71
-6570.02
-9279.72*
-10235.25
-10208.16
-10200.44
-10235.42
-10170.41
-10681.77*

2

3

4

5

6

Mixedwood 1

2

3

4

5

6

-8459.02*

-4463.00

-4464.75

-4463.04

-4458.31

-4462.77

-8629.95*

-4909.08

-4949.45

-4943.30

-4929.45

-4948.66

12

11

10

9

8

7

12

11

10

9

8

7

-10232.66

-10233.99

-10233.02

-10234.89

-10233.12

-10233.37

-6745.91

-6743.97

-6744.09

-6744.35

-6744.18

-6742.36

-9955.20

-9958.72

-9953.23

-9973.32

-9973.34

-9971.37

-9389.64

-9406.47

-9397.31

-9393.19

-9398.88

-9406.17

AIC

-4468.91

-4457.38

-4467.28

-4467.67

-4467.41

-4465.85

-4949.59

-4950.01

-4947.73

-4951.16

-4952.53

-4950.59

AIC

* Models excluded from my analyses of the best model because this model excluded SWP, known to have great ecological relevance to sap flow.

-9384.93

-6749.47

Pure Spruce 1

AIC

number AIC

AIC

Early-summer Late-summer

Model with interaction
Spring

number AIC

Early-summer Late-summer
Model

Spring

Model without interaction

Model

but excludes SWP.

66

and mixedwood treatments. The best SFV models for each season across the treatments are highlighted in gray. Asterisk (*) selected as best model

Table 3.5 Akaike information criterion (AIC) values of each of the 12 mixed effect models fitted using the full data set and across the pure spruce

3.3.4. Mixed effect models of sap flow
3.3.4.1.

Climate and sap flow relationships throughout the growing season

The mixed effect models without interaction showed that the variables in some cases
differed regarding having a positive or negative relationship with SFV using all data and
between PS and MW (Table 3.6). Higher mean daily rainfall and soil temperature decreased
SFV, whereas higher air temperature increased SFV. Solar radiation and SWP were the only
variables that differed between the two treatment types regarding their negative or positive
influence on SFV. Higher solar radiation and SWP showed a positive relationship with SFV
in the PS, whereas these variables in MW showed a negative relationship with SFV.
The mixed effect models with interaction showed that the interactions between the
variables in some cases differed regarding their influence on SFV (Appendix 8.2, Figure 8.58.8). The two interactions that involved solar radiation variables were the only interactions
whose direction differed between the two treatments. In PS, the increase in solar radiation
slightly decreased SFV when rainfall was low and strongly decreased SFV when rainfall was
high. In contrast, in MW, the increase in solar radiation increased SFV when rainfall was low
or high, with a stronger increase when rainfall was high. Moreover, the increase in solar
radiation in PS decreased SFV when SWP was low or high, with a stronger decrease when
SWP was high. In MW, the increase in solar radiation increased SFV when SWP was low or
high, with a stronger increase when SWP was low.
The interactions of air temperature and soil temperature with SWP differed between
the two treatment types only with respect to the strength of interaction influences on SFV:
with interaction strength being higher in PS in both cases. In both PS and MW, increases in
air temperature positively influenced SFV when SWP was low or high, with stronger positive
67

influences when SWP was high. Again in both treatments, increases in soil temperature
negatively impacted SFV when SWP was low or high, with stronger negative influences
when SWP was high.

Table 3.6 Mixed effect growth model structure of selected model and fitted coefficient values of
climatic predictor variables using the full data set and across the treatment types over the growing
season. The variables (i.e. direct effect) that differ across the treatment types regarding their negative
or positive influence on SFV is highlighted in gray. The interactions that differ across the treatment
types regarding their directions (i.e. positive or negative) is also highlighted in gray.

All

Pure Spruce

Mixedwood

Model

Variables

Coefficient

Coefficient

Coefficient

Model 1

(Intercept)

0.0052

0.0056

0.0041

SolRad

-0.0009

0.0029

-0.0024

Rain

-0.0002

-0.0002

-0.0001

SoilTmp

-0.0004

-0.0005

-0.0004

AirTmp

0.0006

0.0008

0.0003

SWP

0.0022

0.0046

-0.0002

(Intercept)

0.0052

0.0048

0.0048

SolRad

-0.0011

0.0055

-0.0055

Rain

-0.0001

-0.0001

-0.0001

SoilTmp

-0.0005

-0.0005

-0.0004

AirTmp

0.0006

0.0009

0.0004

SWP

0.0033

0.0006

0.0039

SolRad: Rain

-0.0002

-0.0007

0.0002

AirTmp: SWP

0.0004

0.0004

0.0002

SoilTmp: SWP

-0.0005

-0.0003

-0.0002

SolRad: SWP

-0.0064

0.0081

-0.0219

Model 7

68

3.3.4.2.

Climate and sap flow relationships by season

The results of the mixed effect model without interaction showed that some variables
differed regarding having a positive or negative relationship on SFV between spring, earlysummer and late-summer in each treatment (Table 3.7). In PS, solar radiation was the only
variable where the relationship with SFV differed between the seasons. Higher solar radiation
increased SFV in spring and late-summer, and decreased SFV in early-summer. In all three
seasons in PS, rainfall and soil temperature showed a negative relationship with SFV,
whereas air temperature and SWP showed a positive relationship with SFV.
In MW, most of the variables fluctuated between a positive or a negative relationship
with SFV between the seasons. Higher solar radiation and higher rainfall were both related to
a decrease in SFV in spring and early-summer, and an increase in SFV in late-summer. Soil
temperature and SWP showed a positive relationship with SFV in spring, and a negative
relationship with SFV in early-summer and late-summer. Higher air temperature increased
SFV in all three seasons in the MW.
The results of the mixed effect models with interaction indicated that the interactions
in some cases differed regarding their influence on SFV between seasons in each treatment
(Table 3.7; Appendix 8.3, Figures 8.9-8.15). My results for PS showed that in spring, the
increase in solar radiation positively influenced SFV when rainfall was low or high, with a
stronger positive influence when rainfall was high. Similarly, in late-summer, the increase in
solar radiation also positively influenced SFV when rainfall was low or high, but with a
stronger positive influence when rainfall was low. In the early-summer, the increase in solar
radiation slightly increased SFV when rainfall was low, and strongly decreased SFV when
rainfall was high.

69

In both spring and late-summer in PS, the increase in air temperature showed a
positive influence in SFV when SWP was low or high, with a stronger positive influence
when SWP was high. In early-summer, the increase in air temperature also increased SFV
when SWP was high, but decreased SFV when SWP was low. Moreover, in the spring, the
increase in solar radiation increased SFV when SWP was low or high, with a stronger
increase when SWP was high. In contrast, in the early-summer, the increase in solar radiation
decreased SFV when SWP was low or high, with a stronger decrease when SWP was high. In
late-summer, the increase in solar radiation strongly increased SFV when SWP was high, and
decreased SFV when SWP was low. In all three seasons, the increase in soil temperature
decreased SFV when SWP was high, and increased SFV when SWP was low.
My results for the MW showed that in spring and late-summer, the increase in soil
temperature increased SFV when SWP was high and strongly decreased SFV when SWP was
low. In contrast, in early-summer the increase in soil temperature decreased SFV when SWP
was low or high, with a stronger decrease when SWP was low. In spring and late-summer,
the increase in solar radiation increased SFV when SWP was low or high, with a stronger
increase when SWP was low. In the early-summer, the increase in solar radiation strongly
increased SFV when SWP was low, and decreased SFV when SWP was high.
In all three seasons in MW, the increase in solar radiation increased SFV irrespective
of rainfall, but with a stronger increase when rainfall was high. Furthermore, in all three
seasons, the increase in air temperature positively influenced SFV when SWP was high or
low. However, the stronger positive influence of the air temperature on SFV occurred when
SWP was low in the spring, and high in the late-summer. In early summer, the increase in air
temperature showed a slightly stronger positive influence in SFV when SWP was low.

70

-0.00041
-0.00038
0.00070
0.00244
0.00031
0.00003
-0.00037
0.00598

Rain

SoilTmp

AirTmp

SWP

SolRad: Rain

AirTmp: SWP

SoilTmp: SWP

SolRad: SWP

0.00398

SWP

0.00677

0.00070

AirTmp

SolRad

-0.00034

SoilTmp

0.00539

-0.00037

Rain

(Intercept)

0.00631

SolRad

Model 7

0.00549

(Intercept)

Model 1

Coefficient

Coefficient

Spring

Model

Variables

-0.01729

-0.00189

0.00138

-0.00089

0.00875

0.00128

-0.00102

-0.00004

-0.00329

0.00630

0.00750

0.00112

-0.00080

-0.00020

-0.00427

0.00684

Coefficient

0.02161

-0.00159

0.00031

-0.00027

0.01061

0.00075

-0.00077

-0.00012

0.02526

0.00435

0.00160

0.00062

-0.00026

-0.00017

0.01939

0.00182

Coefficient

Early-summer Late-summer

Pure Spruce

-0.01778

0.00096

-0.00023

0.00076

0.00707

0.00025

0.00016

-0.00035

-0.00839

0.00549

0.00246

0.00026

0.00009

-0.00023

-0.00661

0.00513

Coefficient

Spring

-0.01963

0.00109

-0.00003

0.00003

-0.00609

0.00044

-0.00076

-0.00008

-0.00702

0.00769

-0.00108

0.00047

-0.00109

-0.00006

-0.00414

0.00957

Coefficient

-0.02377

0.00029

0.00006

0.00046

-0.00097

0.00010

0.00014

-0.00006

-0.00213

0.00032

-0.00156

0.00009

-0.00003

0.00006

0.00748

0.00037

Coefficient

Early-summer Late-summer

Mixedwood

71

Table 3.7 Mixed effect growth model structure of selected model and fitted coefficient values of climatic predictor variables across the treatment
for each season. The variables that differed regarding their negative or positive influence on SFV across the seasons in each treatment are
highlighted in gray, with dark gray for the season that differed from the others. The interactions that differ between the seasons in each treatment
regarding their directions (i.e. positive or negative) are also highlighted in gray, with dark gray for the season(s) that the interaction strongly
differed from other(s).

3.4. Discussion
Spruce sap flow responses to climate are dependent on stand composition and
structure. My findings correspond to previous studies indicating that forest composition can
influence the availability of limiting resources such as water and light throughout the year.
For example, canopy openings resulting from the senescence of deciduous trees in a mixed
stand allow for the release of more shade-tolerant, slow-growing conifers (Brassard and Chen
2006). On the other hand, cloudy days could accentuate the shading of spruce growing under
a deciduous canopy, further reducing access to light. The redistribution of soil water by aspen
(Populus tremuloides Michx.) root systems can improve rooting-zone soil moisture conditions
(Brown et al. 2014), which may benefit spruce planted under established aspen (Kabzems et
al. 2016). Mixed species stands composed of trees capable of hydraulical water redistribution,
have the potential to maintain high transpiration rates during periods of water shortage (i.e.
low rainfall periods) (Brown et al. 2014). My findings reinforce the need to study sap flow
taking into consideration stand composition and structure to better understand tree
physiological responses to climate, and better project their responses to future climate
scenarios.
I demonstrate that it is better to assess how trees respond to climate by evaluating the
interactions between climate variables instead of the direct effects of individual climate
variables. In both stands, SFV reflects an interplay between the amount of soil water and both
solar radiation and temperature. For example, a combination of warm temperatures and
abundant soil water leads to higher SFV rates throughout the growing season in both stands.
This finding highlights the challenges associated with trying to interpret how single variables
impact sap flow. For example, the positive relationship between air temperature and SFV
72

found in all time scales across the stands is consistent with previous studies (Juhász et al.
2013, Chang et al. 2014, Juice et al. 2016). However, the positive influence of air temperature
in SFV is highly dependent on the water availability in the soil. A warming experiment in
mature northern red oak (Quercus rubra L.) trees showed that soil moisture declined with
increased temperatures, and that each soil moisture percentage decrease resulted in a decrease
in sap flow of approximately 360 kg H 2O m −2 sapwood area day−1 (Juice et al. 2016). Small
and McConnell (2008) suggest that simple and complex models frequently used to predict
transpiration are not adequate to model the water balance in the spruce forest in northern New
Mexico. They highlight the need for models that account for interactions between soil
moisture and meteorological conditions.
It is better to assess tree physiological responses to climate by analyzing shifts within
the growing season rather than over the entire growing season. My results indicate that SFV
responses to climate over the entire growing season do not correspond to the SFV responses
within the season in each stand. Moreover, my study indicates that by breaking the growing
season into three seasons, different SFV responses between seasons for each stand type can be
identified. Thus, the evaluation of SFV and climate relationships in a finer temporal scale can
better inform how spruce trees within each stand type respond to current and future climate
conditions.
Some climate variables are more important than others for predicting SFV in specific
periods within the growing season in each stand type. My results show that the best model to
predict SFV differed between the seasons in each stand. However, in some cases, there were
minimal differences between the best model and the other models regarding their AIC values.
I argue that breaking down the seasons in a finer scale might allow a better identification of
the most important individual climate variables to predict sap flow, and the possible
73

interaction between these important climate variables.
By analyzing SFV responses to seasonal climate variability in each stand type we can
better understand how trees are responding physiologically to the drought stress. Past studies
have demonstrated that trees have a different mechanism to undergo drought stress (Anfodillo
et al. 1998). However, most of these studies focused on tree responses to drought stress on a
coarse scale (Pretzsch et al. 2013, Sullivan et al. 2017), and many of these studies focused
attention on novel sap flow tree responses using experimental warming (Juice et al. 2016). By
assessing SFV responses to climate on a finer scale, we can better inform when trees in each
stand are likely to respond to drought with processes that can compromise their physiological
integrity.
Trees in each stand also show different SFV responses to the interplay between
seasonal water available in the soil and air temperature. These differences can be attributed to
the availability of the various resources between seasons in each stand type, as well as the
trees exhibiting different strategies in response to drought stress. However, more study is
necessary to understand the reason for such a different response of SFV to the interaction
between air temperature and soil water potential found between seasons in each stand.
The interaction between air temperature and SWP in SFV between the seasons in the
PS suggests that higher air temperature more positively increases SFV when water is available
in the soil during spring and late-summer. However, the interaction between air temperature
and SWP in early-summer suggests that warmer air temperature increases SFV as long as
there is sufficient water available in the soil, and decreases SFV with low water availability in
the soil. The decrease in SFV as a response to higher air temperature and lower soil water
potential suggests a water-saving behavior, where spruce closes stomata to conserve water
during water stress. Spruce species display isohydric behavior, which reduces water
74

consumption and growth in the early phase of drought stress through stomata closure (Zang et
al. 2012, Pretzsch et al. 2013, Sullivan et al. 2017). Norway spruce (Picea abies) and Swiss
pine (Pinus cembra) were found to reduce sap flux rates during drought periods, suggesting
this water-saving behaviour (Anfodillo et al. 1998).
In the mixedwood stand, the interaction between the air temperature and SWP in SFV
between the seasons suggests that the increase in air temperature increases SFV even when
soil water availability is low throughout the growing season. My results confirmed that in the
late-summer, higher air temperature increases SFV as more water is available. However, a
surprising result is the stronger influence of higher air temperature in SFV when there was
low soil water available during spring and early-summer. Since spruce has a isohydric
character, I do not expect that they will continue to transpire under water stress, as would a
species with anisohydric behavior, such as European larch (Larix decidua) and European
beech (Fagus sylvatica).

3.5. Conclusion
My work indicated that we can better predict spruce sap flow responses to climate
throughout the growing season by using the interaction models, and that evaluating sap flow
and climate relationships at a finer temporal scale can improve our understanding. Moreover,
sap flow responses to climate variability depend on whether the tree is growing in the pure or
mixedwood stands. A key finding of my work is that stand composition and structure are
important determinants of how SFV in white spruce responds to fluctuations in climate
variables within the growing season, and how they will respond to projected future climate
scenarios. Spruce sap flow in both stands is likely to increase as the climate warms in
northeastern British Columbia, increasing the demand for soil water. As this resource
75

becomes less available, white spruce in both stands are likely to respond with processes that
can compromise their physiological integrity. Drought stress is likely to limit boreal white
spruce growth under warmer future climates in pure and mixedwood stands in northeastern
British Columbia. More studies on tree sap flow and climate relationships on a finer scale are
needed to inform better how trees will respond to projected drought stress in different stand
composition and structure.

76

4. Chapter 4: Conclusions

The key finding of my thesis is that white spruce growth responses to inter and intraannual climate variability depends on the stand composition and structure. Assessing tree
growth responses to climate variables in different time scales and the influence of stand
composition and structure provide information that is essential for understanding, managing,
and forecasting of forest stands. Previous studies have predominantly focused on tree growth
and climate relationships at a coarse scale (i.e. yearly), mainly using a dendrochronological
approach, and not taking into consideration the stand composition and structure. Moreover,
many of these studies used climate variables from nearby climate stations instead of
microclimate data measured from an on-site climate station. Studies (Barber et al. 2000,
Cortini et al. 2011, Lloyd et al. 2013) demonstrated that white spruce (Picea glauca (Moench)
Voss) is climate sensitive and has low resistance to drought, but there are very few studies of
their responses to climate variability in a pure and mixedwood stands.
To bridge the knowledge gaps, I studied how individual white spruce trees are
influenced by inter and intra-annual variations in climate variables and how stand
composition and structure influence their responses. Specifically, I had two main objectives:
analysis of annual white spruce growth sensitivity to microclimate variables at different time
scales, and their sensitivity in pure versus mixedwood stands (Chapter 2); analysis of white
spruce sap flow sensitivity to microclimate variables at different time scales, and their
sensitivity in pure versus mixedwood stands (Chapter 3).
In this concluding chapter, I synthesize the main findings from each of the chapters
while discussing the implications for white spruce performance in pure and mixedwood
stands for projected climate shifts, and recommend future research directions.
77

4.1. Main findings and contributions

Chapter 2: White spruce growth sensitivity to climate variability in pure and mixed stands.
I found that the influence of soil water potential (SWP) and air temperature on spruce
growth varies throughout the year. Similar to other studies, I concluded that seasonal and
monthly climate predictor variables are more suitable to understand annual tree growth
responses to climate than annual climate variables. I also argue that we can better understand
how trees respond to climate by evaluating the interactions between climate variables instead
of the direct effects of individual climate variables. In both stands, there is an interplay
between the amount of water available in the soil and air temperature which influences annual
white spruce growth.
Similar to other studies (Barber et al. 2000, Wilmking et al. 2004), I found that warm
springs increase spruce growth, and warm summers decrease spruce growth in both pure and
mixedwood stands. However, here I analyzed the increase in air temperature with no
interaction with soil water available. The increase in air temperature increases the demand for
soil water, which is more available during spring, mainly because of the snowmelt. In the
summer, soil water becomes less accessible, and air temperature increases. The increase in air
temperature combined with low soil water tend to induce stomatal closure and reduces net
photosynthesis. Another way to interpret this is that during the spring, spruce growth seems to
be limited by air temperature, whereas in the summer, soil water appears to be the primary
limiting variable. However, I did not compare the performance between the models with
interaction and the models without interaction, thus my interpretation is based on a
comparison between spruce growth responses to individual climate variables versus their
responses to the interactions between the climate variables.
78

A key finding of Chapter 2 is that stand composition and structure are important
determinants of how annual spruce growth responds to yearly fluctuations in seasonal air
temperature and SWP variables, and how annual spruce growth will respond to projected
future climate scenarios. Annual spruce growth responses to climate variability depends on
whether it is growing in pure or mixedwood stands. An interesting finding is that spruce
growth in pure stands has a positive relationship with SWP during spring and summer, while
spruce growth in mixedwood stands has a negative relationship. I infer that the negative
relationship between spruce growth and SWP during spring and summer in the mixedwood
stands might be consistent with a drought stress mechanism. Spruce trees might not be
benefiting from the water available in the soil, and competing trees might be the main factor
influencing the availability of this resource.
Interaction model result indicates that in mixedwood stands, spruce growth positively
responds to an increase in spring air temperature as long there is sufficient water available in
the soil. Otherwise, spruce growth is negatively influenced by the increase in air temperature
during spring. The negative influence of warm summer temperature to spruce growth in the
mixedwood stand even when soil water levels were good support my hypothesis that drought
stress is the key limiting variable.
It is unclear why air temperature positively increases spruce growth in pure spruce
stands when soil water is low, and negatively influences spruce growth when soil water is
high during spring and summer. Although monthly and seasonal climate variables are more
suitable to predict annual spruce growth, we can derive a better understanding of the
relationship between tree growth and climate variability by evaluating on a finer scale. The
order of importance of climate variables, as well as the best annual and inter-annual climate
predictors of annual spruce growth differ between the two stands. Thus, predictions of annual
79

spruce growth based on climate using annual variables and not taking into consideration stand
composition and structure could lead to erroneous projections.
Chapter 2 revealed that a combination of warmer temperatures and drought during
summer will negatively affect spruce trees in pure and mixedwood stands in the studied
region. White spruce growing in mixedwood stands might be more sensitive to drought stress
than in pure stands due to the higher competition for limiting resources (primarily water). Soil
water potential measurements taken from 1999 to 2017 on-site indicate that mixedwood
stands are drier than pure white spruce stands, with August and September being the months
with the lowest soil water potential in both stands. These results suggest that broadleaves are
not improving soil water conditions with a hydraulic lift mechanism. Instead, my work
suggests that broadleaves are limiting soil water and light access to spruce trees in the
mixedwood stand during summer.
The main question raised in this Chapter is how spruce trees are responding to climate
variability on a finer scale, leading to my study in Chapter 3.

Chapter 3: White spruce sap flow sensitivity to climate variability in pure and mixedwood
stands.
Chapter 3 of this thesis analyzes the impact of a range of microclimate variables on the
sap flow of white spruce trees in pure and mixedwood stands. A key finding is that it is
necessary to assess a tree’s sap flow response to climate at a seasonal or finer temporal grain
rather than over the entire growing season. My work indicates that by breaking the growing
season into three seasons, there are different SFV responses between seasons for each stand
type. Moreover, I also demonstrate that it is better to assess how trees respond to climate by
evaluating the interactions between climate variables instead of the direct effects of individual
80

climate variables. In both stands, there is an interplay between the amount of soil water and
both solar radiation and temperature to influence tree SFV.
A key finding of Chapter 3 is that stand composition and structure are important
determinants of how SFV in white spruce responds to fluctuations in climate variables within
the growing season, and how they will respond to projected future climate scenarios. Spruce
SFV responses to climate variability depends on whether it is growing in pure or mixedwood
stands. Thus, evaluating how spruce sap flow will respond to current and future climate
variability, especially drought stress, using the entire growing season and not taking into
consideration stand composition and structure could lead to erroneous projections.
I found that warm air temperature increases SFV within the growing season in both
stands. However, spruce trees in the two stand types exhibited different SFV responses to the
interplay between seasonal water available in the soil and air temperature. I hypothesize that
the differences are related to the availability of resources between seasons in each stand type,
and possibly to tree life-history strategies to drought stress in specific periods. In pure stands,
higher air temperature positively increases SFV as more water is available in the soil during
spring and late-summer. However, in the early-summer, warmer air temperature increases
SFV as long as there is sufficient water available in the soil, and decreases SFV with low
water availability in the soil, suggesting a water-saving strategy. In mixedwood stands,
increase in air temperature positively influences SFV even when soil water availability is low
throughout the growing season.
The interaction between air temperature and SWP during summer results may suggest
that SFV in the mixedwood stand is limited by low light during the summer, and when air
temperature is high it is because it is sunny and there is more light. I argue that besides water,
light could be another limited variable of SFV in the mixedwood stand during the summer.
81

However, I did not include photosynthetically active radiation (PAR) in my analysis, and I do
not have any evidence that support the relationship of PAR and air temperature. If PAR is a
limited variable of SFV during the summer, this would support even more my hypothesis that
competition for limiting resources (e.g. water and light) is higher in the mixedwood stands
comparing to the pure stands. In the mixedwood stand, PAR is high in the spring but declines
in the summer (Appendix 6, Figure 6.7). Before leaf out, more light is available to spruce
trees but after leaf out, broadleaves block light access to spruce trees beneath their canopy.
Visually comparing SFV for each year for both stands, I found that SFV declines in
mid-summer (anytime from mid-July to early summer), which is the same period where both
stands have low soil water. I can see a trend that early in the growing season, SFV showed
strong responses to the increase in air temperature corresponding to the period of ample soil
availability. However in the mid-summer when SWP becomes limiting, SFV did not show
strong responses to the increase in air temperature. My findings indicated that early in the
growing season, SFV is primarily limited by air temperature, while by mid-summer, drought
stress is the key limiting variable in both stands.
Another interesting finding is that spruce trees in the mixedwood stand have lower
SFV comparing to the spruce trees in the pure stand. The spruce trees in the mixedwood stand
are small, slow-growing and overtopped by broadleaves. The spruce trees in pure stands are
larger, more rapidly growing and their foliage is in full light. Sap flow velocity comparison
between spruces in the stand types indicates that spruce trees in the pure stand are transpiring
at a more rapid rate than spruces in the mixedwood stand. However, my sap flow
measurements are expressed per unit area (mm3/mm2), which allows for differences in the
total quantity of sap flowing up each tree. The soil in the mixedwood stands are drier than the
pure spruce stands. Spruce in plot A3 (mixedwood treatment) suffered substantial mortality
82

and poor growth, while spruce in plot A2 (pure spruce treatment) had little mortality.
Together, these findings support my hypothesis that drought conditions limit SFV and spruce
growth in both stands, but drought might be more severe in the mixedwood stand than in the
pure stand.
Chapter 3 reinforced the main findings from Chapter 2 regarding the importance of
stand composition and structure, interactions between the climate variables, and the need to
study trees response to climate at a finer scale. Chapter 3 also asserts that drought stress is the
main concern for the performance of white spruce trees in the two stands, and the concern of
drought stress is even greater for the mixedwood stands compared to the pure spruce stands
on my study site. Spruce sap flow in both stands is likely to increase as the climate warms,
increasing the demand for soil water. As this resource becomes less available, white spruce in
both stands are likely to respond with processes that can compromise their physiological
integrity.
Especially in ongoing changes in climate, soil water is the key factor to consider when
deciding where to plant spruce trees and which species to mix with them to allow a favorable
microclimate and competitive environment for trees to grow. Declining growth and increased
mortality of spruce in boreal forests have generally been attributed to drought stress, and
drought stress is projected to increase, especially during summer. A management possibility
to reduce the negative impact of drought stress is the appropriate site selection to plant spruce
trees. For example, evaluate the soil moisture regime of the site before planting the spruce
trees and give preference for sites with high soil water holding capacity. In addition, site
preparation to improve the microclimate experienced by the spruce seedlings and the control
competition vegetation (Cortini et al. 2011). Another possibility is the selection of the
appropriate species to mix with spruce trees that benefit the microclimate experienced by the
83

trees within this stand.
Forest managers should also consider how to adjust the competitive relationships that
affect stand structure. For example, under drought prone-conditions, planting spruce trees
under the understory beneath aspen and other broadleaves might negatively influence spruce
growth due to the high competition for water and light. Instead, forest managers should
develop strategies that give the spruce trees a competitive advantage so that they can grow
above or at the same height as associated broadleaved trees. Foresters should also consider
density management, for example, considering spacing (e.g. planting trees in a wider spacing)
and thinning treatments as strategies of reducing moisture competition and maintaining tree
health in mixed and pure stands during drought prone-conditions (Sohn et al. 2016).

4.2. Future research
Future research should study individual tree’s growth response to climate variables in
pure and mixedwood stands, using a framework similar to what I used in my thesis, but on a
finer time scale. For example, sap flow responses to climate variability in weekly or monthly
time periods using hourly sap flow and climate variable measurements. This study might
allows us to better identify the key limiting variables and interactions in each time scale, and
to identify when in the summer soil water is most limited to spruce trees. In addition, since
dendrochronology methods just allow the evaluation of tree growth data over both seasonal
and annual timeframes, high resolution dendrometers can be used to measure intra-annual tree
growth responses to climate since they provide continuous measurements of tree
circumference. High resolution dendrometer data can also be linked to sap flow measurements
to gain insight into the relationship between tree photosynthesis, sap flow, and fluctuations in
bole diameter in response to microsite climate.
84

Future work could also investigate additional climate variables and their interactions.
Theoretically, sap flow rates should be driven by the availability of photosynthetic active
radiation, as well as being impacted by seasonal variables such as snow depth. Future studies
could use a similar framework to my thesis to study other species in the mixedwood stands,
providing answers to how trees within these stands compete or partition resources (especially
water). For instance, many of the questions I raised in my thesis regarding competition for
water could be answered if, for example, sap flow was also measured in the aspen and willow
trees. Moreover, future study could also include a competition index in the model to predict
sap flow and radial growth based on climate.

4.3. Concluding remarks
My thesis advances our understanding of how individual white spruce trees respond to
climate variables in pure and mixedwood stands, provides information on expected changes in
their sap flow and radial growth in relation to climate variability, and demonstrates the
importance of appropriate site selection and management of these stands. This knowledge is
needed for forest managers to develop strategic and operational plans in order to maximize
forest sustainably and productivity in a current and projected changes in climate.

85

5. Bibliography
Alarcón, J. J., R. Domingo, S. R. Green, M. J. Sánchez-Blanco, P. Rodríguez, and A.
Torrecillas. 2000. Sap flow as an indicator of transpiration and the water status of
young apricot trees. Plant and Soil 227:77–85.
Anfodillo, T., S. Rento, V. Carraro, L. Furlanetto, C. Urbinati, and M. Carrer. 1998. Tree
water relations and climatic variations at the alpine timberline: Seasonal changes of
sap flux and xylem water potential in Larix decidua Miller, Picea abies (L.) Karst. and
Pinus cembra L. Annales des Sciences Forestières 55:159–172.
Arnone III, J. A., P. S. J. Verburg, D. W. Johnson, J. D. Larsen, R. L. Jasoni, A. J. Lucchesi,
C. M. Batts, C. von Nagy, W. G. Coulombe, D. E. Schorran, P. E. Buck, B. H.
Braswell, J. S. Coleman, R. A. Sherry, L. L. Wallace, Y. Luo, and D. S. Schimel.
2008. Prolonged suppression of ecosystem carbon dioxide uptake after an
anomalously warm year. Nature 455:383–386.
Balducci, L., A. Deslauriers, A. Giovannelli, M. Beaulieu, S. Delzon, S. Rossi, and C. B. K.
Rathgeber. 2015. How do drought and warming influence survival and wood traits of
Picea mariana saplings? Journal of Experimental Botany 66:377–389.
Barber, V. A., G. P. Juday, and B. P. Finney. 2000. Reduced growth of Alaskan white spruce
in the twentieth century from temperature-induced drought stress. Nature 405:668673.
Barigah, T. S., O. Charrier, M. Douris, M. Bonhomme, S. Herbette, T. Améglio, R. Fichot, F.
Brignolas, and H. Cochard. 2013. Water stress-induced xylem hydraulic failure is a
causal factor of tree mortality in beech and poplar. Annals of Botany 112:1431–1437.
Bartoń, K. 2019. Multi-model inference. R package version 1.43.6. https://cran.rproject.org/web/packages/MuMIn/index.html.
Bergh, J., and S. Linder. 1999. Effects of soil warming during spring on photosynthetic
recovery in boreal Norway spruce stands. Global Change Biology 5:245–253.

86

Berry, Z. C., N. Looker, F. Holwerda, G. Aguilar, L. Rodrigo, P. Ortiz Colin, T. González
Martínez, and H. Asbjornsen. 2017. Why size matters: The interactive influences of
tree diameter distribution and sap flow parameters on upscaled transpiration. Tree
Physiology 13:1–13.
Bonan, G. B. 2002. Ecological climatology: Concepts and applications. Cambridge University
Press, New York:31.
Brassard, B. W., and H. Y. H. Chen. 2006. Stand structural dynamics of North American
boreal forests. Critical Reviews in Plant Sciences 25:115–137.
Brown, S. M., R. M. Petrone, L. Chasmer, C. Mendoza, M. S. Lazerjan, S. M. Landhäusser,
U. Silins, J. Leach, and K. J. Devito. 2014. Atmospheric and soil moisture controls on
evapotranspiration from above and within a Western Boreal Plain aspen forest.
Hydrological Processes 28:4449–4462.
Bunn, A. 2008. A dendrochronology program library in R (dplR). Dendrochronologia, 25:
115-124.
Burnham, K. P., and D. R. Anderson. 2002. Model selection and multimodel inference: a
practical information-theoretic approach. Second edition. Springer-Verlag, New York.
Callaway, R. M., and L. R. Walker. 1997. Competition and facilitation: A synthetic approach
to interactions in plant communities. Ecology 78:1958–1965.
Chang, X., W. Zhao, and Z. He. 2014. Radial pattern of sap flow and response to
microclimate and soil moisture in Qinghai spruce (Picea crassifolia) in the upper
Heihe River Basin of arid northwestern China. Agricultural and Forest Meteorology
187:14–21.
Choat, B. 2013. Predicting thresholds of drought-induced mortality in woody plant species.
Tree Physiology 33:669–671.

87

Clark, J. S., D. M. Bell, M. Kwit, A. Stine, B. Vierra, and K. Zhu. 2012. Individual-scale
inference to anticipate climate-change vulnerability of biodiversity. Philosophical
Transactions of the Royal Society B: Biological Sciences 367:236–246.
Cleland, E. E., I. Chuine, A. Menzel, H. A. Mooney, and M. D. Schwartz. 2007. Shifting plant
phenology in response to global change. Trends in Ecology & Evolution 22:357–365.
Coomes, D. A., and R. B. Allen. 2007. Effects of size, competition and altitude on tree
growth. Journal of Ecology 95:1084–1097.
Cooper, R. N., J. T. Houghton, J. J. McCarthy, and B. Metz. 2002. Climate change 2001: The
scientific basis. Foreign Affairs 81:208.
Cortini, F., P. G. Comeau, J. O. Boateng, L. Bedford, J. McClarnon, and A. Powelson. 2011.
Effects of climate on growth of lodgepole pine and white spruce following site
preparation and its implications in a changing climate. Canadian Journal of Forest
Research 41:180–194.
Cortini, F., P. G. Comeau, and M. Bokalo. 2012. Trembling aspen competition and climate
effects on white spruce growth in boreal mixtures of Western Canada. Forest Ecology
and Management 277:67–73.
D’Arrigo, R. D., R. K. Kaufmann, N. Davi, G. C. Jacoby, C. Laskowski, R. B. Myneni, and P.
Cherubini. 2004. Thresholds for warming-induced growth decline at elevational tree
line in the Yukon Territory. Global Biogeochemical Cycles 18:1–7.
Delong, S. C., A. Banner, W. H. MacKenzie, B. J. Rogers, and B. Kaytor. 2011. A field guide
to ecosystem identification for the Boreal White and Black Spruce Zone of British
Columbia. Pages 1–250. BC Ministry of Forests and Range, Forest Science Program,
Land Management Handbook No. 65. 250 p.
Delzon, S., and D. Loustau. 2005. Age-related decline in stand water use: sap flow and
transpiration in a pine forest chronosequence. Agricultural and Forest Meteorology
129:105–119.

88

Deslauriers, A., J.-G. Huang, L. Balducci, M. Beaulieu, and S. Rossi. 2016. The contribution
of carbon and water in modulating wood formation in black spruce saplings. Plant
Physiology 170:2072–2084.
Dietrich, L., S. Delzon, G. Hoch, and A. Kahmen. 2019. No role for xylem embolism or
carbohydrate shortage in temperate trees during the severe 2015 drought. Journal of
Ecology 107:334–349.
Elliott, K. J., C. F. Miniat, N. Pederson, and S. H. Laseter. 2015. Forest tree growth response
to hydroclimate variability in the southern Appalachians. Global Change Biology
21:4627–4641.
FAO, editor. 2016. Forests and agriculture: land-use challenges and opportunities. FAO,
Rome.
Farrar, J. L. 1995. Trees in Canada. Natural Resources Canada, Canadian Forest Service,
Headquarters, Ottawa, Copublished by Fitzhenry and Whiteside Limited, Markham,
Ontario.
Fernández-de-Uña, L., I. Cañellas, and G. Gea-Izquierdo. 2015. Stand competition determines
how different tree species will cope with a warming climate. PLOS ONE
10:e0122255.
Flato, G., N. Gillett, A. Cannon, and J. Anstey. 2019. Modelling future climate change;
Chapter 3 in Canada’s changing climate report, (ed.) E. Bush and D.S. Lemmen. Pages
74–111. Government of Canada, Ottawa, Ontario,.
Foord, V. 2016. Climate patterns, trends, and projections for the Omineca, Skeena, and
northeast natural resource regions, British Columbia. Technical Report, Ministry of
Forests, Lands, Natural Resource Operations and Rural Development, Victoria,
Province of British Columbia.
Galvez, D. A., S. M. Landhäusser, and M. T. Tyree. 2011. Root carbon reserve dynamics in
aspen seedlings: Does simulated drought induce reserve limitation? Tree Physiology
31:250–257.
89

Goldblum, D., and L. S. Rigg. 2005. Tree growth response to climate change at the deciduous
boreal forest ecotone, Ontario, Canada. Canadian Journal of Forest Research 35:2709–
2718.
Granier, A. 1985. Une nouvelle méthode pour la mesure du flux de sève brute dans le tronc
des arbres. Annales des Sciences Forestières 42:193–200.
Granier, A., P. Biron, N. Bréda, J.-Y. Pontailler, and B. Saugier. 1996. Transpiration of trees
and forest stands: Short and long-term monitoring using sapflow methods. Global
Change Biology 2:265–274.
Grissino-Mayer, H. 2001. Evaluating crossdating accuracy: A manual and tutorial for the
computer program COFECHA. Tree-Ring Research, 57, 205–221.
Haeussler, S., R. Kabzems, L. Bedford, and J. McClarnon. submitted. Thirty one years of
structural and successional change in a boreal mixedwood plant community after
silvicultural disturbances. Submitted to Canadian Journal of Forest Research (Jan
2020).
Hamann, A., and T. Wang. 2006. Potential effects of climate change on ecosystem and tree
species distribution in British Columbia. Ecology 87:2773–2786.
Harley, G. L., H. D. Grissino-Mayer, and S. P. Horn. 2011. The dendrochronology of Pinus
elliottii in the Lower Florida Keys: Chronology development and climate response.
Tree-Ring Research 67:39–50.
Herrmann, V., S. M. McMahon, M. Detto, J. A. Lutz, S. J. Davies, C.-H. Chang-Yang, and K.
J. Anderson-Teixeira. 2016. Tree circumference dynamics in four forests characterized
using automated dendrometer bands. PLOS ONE 11:e0169020.
Hogg, E. H. (Ted), J. P. Brandt, and M. Michaelian. 2008. Impacts of a regional drought on
the productivity, dieback, and biomass of western Canadian aspen forests. Canadian
Journal of Forest Research 38:1373–1384.

90

Holmes, R. L. 1983. Computer-assisted quality control in tree-ring dating and measurements.
Tree-Ring Bulletin 44:69–75.
IPCC. 2013. Climate Change 2013: The physical science basis. Contribution of working
group I to the fifth assessment report of the Intergovernmental Panel on Climate
Change. [Stocker, T.F., D. Qin, G.-K. Plattner, M. Tignor, S.K. Allen, J. Boschung, A.
Nauels, Y. Xia, V. Bex and P.M. Midgley (eds.)]. Cambridge University Press,
Cambridge, United Kingdom and New York, NY, USA, 1535 pp.
Jiang, X., J.-G. Huang, K. J. Stadt, P. G. Comeau, and H. Y. H. Chen. 2016. Spatial climatedependent growth response of boreal mixedwood forest in western Canada. Global
and Planetary Change 139:141–150.
Johnston, M., and T. Williamson. 2005. Climate change implications for stand yields and soil
expectation values: a northern Saskatchewan case study. The Forestry Chronicle
81:683–690.
Juhász, Á., P. Sepsi, Z. Nagy, L. Tőkei, and K. Hrotkó. 2013. Water consumption of sweet
cherry trees estimated by sap flow measurement. Scientia Horticulturae 164:41–49.
Juice, S. M., P. H. Templer, N. G. Phillips, A. M. Ellison, and S. L. Pelini. 2016. Ecosystem
warming increases sap flow rates of northern red oak trees. Ecosphere 7:e01221.
Kabzems, R., P. G. Comeau, C. N. Filipescu, B. Rogers, and A. F. Linnell Nemec. 2016.
Creating boreal mixedwoods by planting spruce under aspen: Successful establishment
in uncertain future climates. Canadian Journal of Forest Research 46:1217–1223.
Kelty, M. J. 1992. Comparative productivity of monocultures and mixed-species stands.
Pages 125–141 in M. J. Kelty, B. C. Larson, and C. D. Oliver, editors. The ecology
and silviculture of mixed-species forests: a festschrift for David M. Smith. Springer
Netherlands, Dordrecht.
Konar, M., R. Muneepeerakul, S. Azaele, E. Bertuzzo, A. Rinaldo, and I. Rodriguez-Iturbe.
2010. Potential impacts of precipitation change on large-scale patterns of tree
diversity. Water Resources Research 46:W11515.
91

Kozlowski, T. T., and S. G. Pallardy. 2002. Acclimation and adaptive responses of woody
plants to environmental stresses. The Botanical Review 68:270–334.
Kulaç, Ş., P. Nzokou, D. Guney, B. M. Cregg, and I. Turna. 2012. Growth and physiological
response of Fraser Fir [Abies fraseri (Pursh) Poir.] seedlings to water stress: Seasonal
and diurnal variations in photosynthetic pigments and carbohydrate concentration.
HortScience 47:1512–1519.
Kunstler, G., C. H. Albert, B. Courbaud, S. Lavergne, W. Thuiller, G. Vieilledent, N. E.
Zimmermann, and D. A. Coomes. 2011. Effects of competition on tree radial-growth
vary in importance but not in intensity along climatic gradients. Journal of Ecology
99:300–312.
Lautner, S. 2013. Wood formation under drought stress and salinity. Pages 187–202 Cellular
aspects of wood formation. Springer, Berlin, Heidelberg.
Le Quéré, C., M. R. Raupach, J. G. Canadell, G. Marland, L. Bopp, P. Ciais, T. J. Conway, S.
C. Doney, R. A. Feely, P. Foster, P. Friedlingstein, K. Gurney, R. A. Houghton, J. I.
House, C. Huntingford, P. E. Levy, M. R. Lomas, J. Majkut, N. Metzl, J. P. Ometto,
G. P. Peters, I. C. Prentice, J. T. Randerson, S. W. Running, J. L. Sarmiento, U.
Schuster, S. Sitch, T. Takahashi, N. Viovy, G. R. van der Werf, and F. I. Woodward.
2009. Trends in the sources and sinks of carbon dioxide. Nature Geoscience 2:831–
836.
Lebourgeois, F., P. Eberle, P. Merian, and I. Seynave. 2014. Social status-mediated tree-ring
responses to climate of Abies alba and Fagus sylvatica shift in importance with
increasing stand basal area. Forest Ecology and Management 328:209–218.
Legendre-Fixx, M., L. D. L. Anderegg, A. K. Ettinger, and J. HilleRisLambers. 2017. Siteand species-specific influences on sub-alpine conifer growth in Mt. Rainier National
Park, USA. Forests 9:1–16.

92

Lenth, R., H. Singmann, L. Jonathon, P. Buerkner, and H. Maxime. 2020. emmeans:
Estimated marginal means. R package version 3.1-140. https://cran.rproject.org/web/packages/emmeans/index.html.
Liaw, A., and M. Wiener. 2002. Classification and regression by randomForest. R news 2:18–
22.
Linderholm, H. W., and K. Linderholm. 2014. Age-dependent climate sensitivity of Pinus
sylvestris L. in the central Scandinavian Mountains. Boreal Environment Research
9:307–318.
Lloret, F., E. G. Keeling, and A. Sala. 2011. Components of tree resilience: Effects of
successive low-growth episodes in old ponderosa pine forests. Oikos 120:1909–1920.
Lloyd, A. H., P. A. Duffy, and D. H. Mann. 2013. Nonlinear responses of white spruce
growth to climate variability in interior Alaska. Canadian Journal of Forest Research
43:331–343.
Lloyd, A. H., and C. L. Fastie. 2003. Recent changes in treeline forest distribution and
structure in interior Alaska. Écoscience 10:176–185.
Lloyd, A. H., T. S. Rupp, C. L. Fastie, and A. M. Starfield. 2002. Patterns and dynamics of
treeline advance on the Seward Peninsula, Alaska. Journal of Geophysical Research:
Atmospheres 107:1–15.
Lo, Y.-H., J. A. Blanco, B. Seely, C. Welham, and J. P. (Hamish) Kimmins. 2010.
Relationships between climate and tree radial growth in interior British Columbia,
Canada. Forest Ecology and Management 259:932–942.
Looker, N., J. Martin, Z. Hoylman, K. Jencso, and J. Hu. 2018. Diurnal and seasonal coupling
of conifer sap flow and vapour pressure deficit across topoclimatic gradients in a
subalpine catchment. Ecohydrology 11:e1994.

93

Lord, T. M., and A. J. Green. 1973. Soils of the Rose Prairie-Blueberry River area, map sheet
94A/N ½, Peace River District, British Columbia. Report No. 17 of the British
Columbia Soil Survey, Agriculture Canada, Vancouver, B.C. (map only).
Lu, P., W. H. Parker, M. Cherry, S. Colombo, W. C. Parker, R. Man, and N. Roubal. 2014.
Survival and growth patterns of white spruce (Picea glauca [Moench] Voss)
rangewide provenances and their implications for climate change adaptation. Ecology
and Evolution 4:2360–2374.
Luo, Y., and H. Y. H. Chen. 2015. Climate change-associated tree mortality increases without
decreasing water availability. Ecology Letters 18:1207–1215.
Ma, Z., C. Peng, Q. Zhu, H. Chen, G. Yu, W. Li, X. Zhou, W. Wang, and W. Zhang. 2012.
Regional drought-induced reduction in the biomass carbon sink of Canada’s boreal
forests. Proceedings of the National Academy of Sciences 109:2423–2427.
Madrigal-González, J., and M. A. Zavala. 2014. Competition and tree age modulated last
century pine growth responses to high frequency of dry years in a water limited forest
ecosystem. Agricultural and Forest Meteorology 192–193:18–26.
McCarthy, J. J., G. d’experts intergouvernemental sur l’évolution du climat, and I. P. on C. C.
W. G. II. 2001. Climate Change 2001: Impacts, adaptation, and vulnerability:
contribution of working group II to the third assessment report of the
Intergovernmental Panel on Climate Change. Cambridge University Press.
McDowell, N. G. 2011. Mechanisms linking drought, hydraulics, carbon metabolism, and
vegetation mortality. Plant Physiology 155:1051–1059.
Messaoud, Y., and H. Y. H. Chen. 2011. The influence of recent climate change on tree
height growth differs with species and spatial environment. PLoS ONE 6:1–8.
Metsaranta, J. M., and V. J. Lieffers. 2008. Inequality of size and size increment in Pinus
banksiana in relation to stand dynamics and annual growth rate. Annals of Botany
101:561–571.

94

Mitchell, P. J., A. P. O’Grady, K. R. Hayes, and E. A. Pinkard. 2014. Exposure of trees to
drought-induced die-off is defined by a common climatic threshold across different
vegetation types. Ecology and Evolution 4:1088–1101.
Miyamoto, Y., H. P. Griesbauer, and D. S. Green. 2010. Growth responses of three coexisting
conifer species to climate across wide geographic and climate ranges in Yukon and
British Columbia. Forest Ecology and Management 259:514–523.
Morin, X., L. Fahse, M. Scherer-Lorenzen, and H. Bugmann. 2011. Tree species richness
promotes productivity in temperate forests through strong complementarity between
species. Ecology Letters 14:1211–1219.
Muñoz-Villers, L. E., F. Holwerda, M. Gómez-Cárdenas, M. Equihua, H. Asbjornsen, L. A.
Bruijnzeel, B. E. Marín-Castro, and C. Tobón. 2012. Water balances of old-growth
and regenerating montane cloud forests in central Veracruz, Mexico. Journal of
Hydrology 462–463:53–66.
Nienstaedt, H., and J. C. Zasada. 1990. Picea glauca (Moench) Voss. white spruce. Pages
204–226 Silvics of North America. United States Department of Agriculture.
Agriculture Handbook 654.
Pan, Y., J. M. Chen, R. Birdsey, K. McCullough, L. He, and F. Deng. 2011. Age structure and
disturbance legacy of North American forests. Biogeosciences 8:715–732.
Pangle, R. E., J.-M. Limousin, J. A. Plaut, E. A. Yepez, P. J. Hudson, A. L. Boutz, N. Gehres,
W. T. Pockman, and N. G. McDowell. 2015. Prolonged experimental drought reduces
plant hydraulic conductance and transpiration and increases mortality in a piñonjuniper woodland. Ecology and Evolution 5:1618–1638.
Piao, S., P. Ciais, P. Friedlingstein, P. Peylin, M. Reichstein, S. Luyssaert, H. Margolis, J.
Fang, A. Barr, A. Chen, A. Grelle, D. Y. Hollinger, T. Laurila, A. Lindroth, A. D.
Richardson, and T. Vesala. 2008. Net carbon dioxide losses of northern ecosystems in
response to autumn warming. Nature 451:49–52.

95

Ping, L., U. Laurent, and P. Zhao. 2004. Granier’s thermal dissipation probe (TDP) method
for measuring sap flow in trees: Theory and practice. Acta Botanica Sinica-English
Edition 46:631–646.
Pinheiro, J., S. DebRoy, D. Sarkar, and R Core Team. 2019. nlme: Linear and nonlinear
mixed effects models. R package version 3.1-140. https://CRAN.Rproject.org/package=nlme.
Plaut, J. A., E. A. Yepez, J. Hill, R. Pangle, J. S. Sperry, W. T. Pockman, and N. G.
McDowell. 2012. Hydraulic limits preceding mortality in a piñon-juniper woodland
under experimental drought. Plant, Cell & Environment 35:1601–1617.
Powelson, A., J. Heineman, L. Bedford, J. McClarnon, A. F. L. Nemec, and J. OtchereBoateng. 2016. White spruce responses to mechanical site preparation, chemical site
preparation, and post-planting vegetation control over three decades in the Boreal
White and Black Spruce Zone of British Columbia. Pages 1–65. Technical Report 103,
Province of British Columbia, Victoria, British Columbia.
Pretzsch, H., J. Block, J. Dieler, P. H. Dong, U. Kohnle, J. Nagel, H. Spellmann, and A.
Zingg. 2010. Comparison between the productivity of pure and mixed stands of
Norway spruce and European beech along an ecological gradient. Annals of Forest
Science 67:712–712.
Pretzsch, H., and J. Dieler. 2011. The dependency of the size-growth relationship of Norway
spruce (Picea abies [L.] Karst.) and European beech (Fagus sylvatica [L.]) in forest
stands on long-term site conditions, drought events, and ozone stress. Trees 25:355–
369.
Pretzsch, H., G. Schütze, and E. Uhl. 2013. Resistance of European tree species to drought
stress in mixed versus pure forests: Evidence of stress release by inter-specific
facilitation. Plant Biology 15:483–495.
Price, D. T., R. I. Alfaro, K. J. Brown, M. D. Flannigan, R. A. Fleming, E. H. Hogg, M. P.
Girardin, T. Lakusta, M. Johnston, D. W. McKenney, J. H. Pedlar, T. Stratton, R. N.
96

Sturrock, I. D. Thompson, J. A. Trofymow, and L. A. Venier. 2013. Anticipating the
consequences of climate change for Canada’s boreal forest ecosystems. Environmental
Reviews 21:322–365.
R Development Core Team. 2011. R: A language and environment for statistical computing.
R Foundation for Statistical Computing, Vienna. http://www.R-project.org.
Rehfeldt, G. E., C. C. Ying, D. L. Spittlehouse, and D. A. Hamilton. 1999. Genetic responses
to climate in Pinus contorta: niche breadth, climate change, and reforestation.
Ecological Monographs 69:375–407.
Richards, A. E., D. I. Forrester, J. Bauhus, and M. Scherer-Lorenzen. 2010. The influence of
mixed tree plantations on the nutrition of individual species: A review. Tree
Physiology 30:1192–1208.
Rohner, B., H. Bugmann, and C. Bigler. 2013. Estimating the age–diameter relationship of
oak species in Switzerland using nonlinear mixed-effects models. European Journal of
Forest Research 132:751–764.
Ruiz-Benito, P., E. R. Lines, L. Gómez-Aparicio, M. A. Zavala, and D. A. Coomes. 2013.
Patterns and drivers of tree mortality in Iberian Forests: Climatic effects are modified
by competition. PLoS ONE 8:1–10.
Sánchez-Salguero, R., J. C. Linares, J. J. Camarero, J. Madrigal-González, A. Hevia, Á.
Sánchez-Miranda, J. A. Ballesteros-Cánovas, R. Alfaro-Sánchez, A. I. GarcíaCervigón, C. Bigler, and A. Rigling. 2015. Disentangling the effects of competition
and climate on individual tree growth: A retrospective and dynamic approach in Scots
pine. Forest Ecology and Management 358:12–25.
Sevanto, S., N. G. Mcdowell, L. T. Dickman, R. Pangle, and W. T. Pockman. 2014. How do
trees die? A test of the hydraulic failure and carbon starvation hypotheses. Plant, Cell
& Environment 37:153–161.
Small, E. E., and J. R. McConnell. 2008. Comparison of soil moisture and meteorological
controls on pine and spruce transpiration. Ecohydrology 1:205–214.
97

Sohn, J. A., S. Saha, and J. Bauhus. 2016. Potential of forest thinning to mitigate drought
stress: A meta-analysis. Forest Ecology and Management 380:261–273.
Sperry, J. S. 2000. Hydraulic constraints on plant gas exchange. Agricultural and Forest
Meteorology 104:13–23.
Spittlehouse, D. L. 2008. Climate change, impacts and adaptation scenarios: Climate change
and forest and range management in British Columbia. Pages 1–38. Technical Report
045, British Columbia, Ministry of Forests and Range, Forest Science Program,
Victoria, British Columbia.
Steppe, K., M. W. Vandegehuchte, R. Tognetti, and M. Mencuccini. 2015. Sap flow as a key
trait in the understanding of plant hydraulic functioning. Tree Physiology 35:341–345.
Stokes, M. A., and T. L. Smiley. 1968. Introduction to tree-ring dating. University of Chicago
Press, Chicago.
Sullivan, P. F., R. R. Pattison, A. H. Brownlee, S. M. P. Cahoon, and T. N. Hollingsworth.
2016. Effect of tree-ring detrending method on apparent growth trends of black and
white spruce in interior Alaska. Environmental Research Letters 11:114007.
Sullivan, P. F., R. R. Pattison, A. H. Brownlee, S. M. P. Cahoon, and T. N. Hollingsworth.
2017. Limited evidence of declining growth among moisture-limited black and white
spruce in interior Alaska. Scientific Reports 7:1–14.
Swanson, R. H. 1994. Significant historical developments in thermal methods for measuring
sap flow in trees. Agricultural and Forest Meteorology 72:113–132.
Thomson, A. M., and W. H. Parker. 2008. Boreal forest provenance tests used to predict
optimal growth and response to climate change. 1. Jack pine. Canadian Journal of
Forest Research 38:157–170.
Trouvé, R., J.-D. Bontemps, I. Seynave, C. Collet, and F. Lebourgeois. 2015. Stand density,
tree social status and water stress influence allocation in height and diameter growth of
Quercus petraea (Liebl.). Tree Physiology 35:1035–1046.
98

van Mantgem, P. J., N. L. Stephenson, J. C. Byrne, L. D. Daniels, J. F. Franklin, P. Z. Fulé,
M. E. Harmon, A. J. Larson, J. M. Smith, A. H. Taylor, and T. T. Veblen. 2009.
Widespread increase of tree mortality rates in the western United States. Science (New
York, N.Y.) 323:521–524.
Vilagrosa, A., E. Chirino, J. J. Peguero-Pina, T. S. Barigah, H. Cochard, and E. Gil-Pelegrín.
2012. Xylem cavitation and embolism in plants living in water-limited ecosystems.
Pages 63–109 in R. Aroca, editor. Plant responses to drought stress: From
morphological to molecular features. Springer, Berlin, Heidelberg.
Wang, T., A. Hamann, D. L. Spittlehouse, and T. Q. Murdock. 2011. ClimateWNA—Highresolution spatial climate data for Western North America. Journal of Applied
Meteorology and Climatology 51:16–29.
Weber, P., H. Bugmann, P. Fonti, and A. Rigling. 2008. Using a retrospective dynamic
competition index to reconstruct forest succession. Forest Ecology and Management
254:96–106.
Wilmking, M., G. P. Juday, V. A. Barber, and H. S. J. Zald. 2004. Recent climate warming
forces contrasting growth responses of white spruce at treeline in Alaska through
temperature thresholds. Global Change Biology. 10:1724-1736.
Wilson, C. C. 1948. The effect of some environmental factors on the movements of guard
cells. Plant Physiology 23:5–37.
Woods, A., D. C. Coates, and A. Hamann. 2005. Is an unprecedented Dothistroma needle
blight epidemic related to climate change? BioScience 55:761–769.
Wullschleger, S. D., P. J. Hanson, and D. E. Todd. 2001. Transpiration from a multi-species
deciduous forest as estimated by xylem sap flow techniques. Forest Ecology and
Management 143:205–213.
Yamaguchi, D. K. 1991. A simple method for cross‐dating increment cores from living trees.
Canadian Journal of Forest Research. 21:414–416.

99

Yin, G., G. Zhou, J. Morris, Z. Huang, G. Chu, and G. Zhou. 2004. Sap flow response of
Eucaylyptus (Eucalyptus urophylla) to environmental stress in South China. Journal of
Zhejiang University- Science 5:1218–1225.
Young-Robertson, J. M., W. R. Bolton, U. S. Bhatt, J. Cristóbal, and R. Thoman. 2016.
Deciduous trees are a large and overlooked sink for snowmelt water in the boreal
forest. Scientific Reports 6:1–10.
Zang, C., H. Pretzsch, and A. Rothe. 2012. Size-dependent responses to summer drought in
Scots pine, Norway spruce and common oak. Trees 26:557–569.
Zolkos, S. G., P. Jantz, T. Cormier, L. R. Iverson, D. W. McKenney, and S. J. Goetz. 2015.
Projected tree species redistribution under climate change: Implications for ecosystem
vulnerability across protected areas in the eastern United States. Ecosystems 18:202–
220.

100

6. Appendix 1: Study area

Figure 6.1 Inga Overview Map. The treatments that I used in this study are the untreated (mixedwood
treatment, MW) with the plots A3, B7, C8, D1 and E7; and the herbicide (pure white spruce PS
treatment) with the plots A2, B1, C5, D4 and E1. Map source: Powelson et al. (2016).

101

Figure 6.2 Inga Map. Mixedwood treatment with the plots A3, B7, C8, D1 and E7; and pure white
spruce treatment with the plots A2, B1, C5, D4 and E1. Climate stations are located in the open area
(west of plot A3), plot A3 and plot A2. Map image source: Vivid - Canada, DigitalGlobe (2014).

102

a)

b)

Figure 6.3 Illustration of pure spruce treatment- plot A2 (a) and mixedwood treatment- plot A3 (b).
Date of the pictures: September 23, 2018 (fall).

103

Table 6.1 Total basal area (m2/ha) of all species, spruces and deciduous in circular neighborhood plots
(n= 11) centered on selected spruce trees (t= 11) in plot A2 (pure spruce treatment) and in circular
neighborhood plots (n= 4) centered on selected spruce trees (t= 4) in plot A3 (mixedwood treatment).
Basal area calculated using the diameter at breast heigh (DBH) of all trees located in each circular
neighborhood plot with radius of 5.98 m (0.011 ha). Tree identification (ID) refers to the selected
spruce trees (i.e. focal trees) that circular neighborhood plots were centered.

Treatment
Pure Spruce

Plot Tree ID
A2 7
15
18
21
23
34
35
36
37
44
46

Mixedwood

A3

3
4
11
859

Basal (DBH) area (m2/ha)
Total (all species) Spruces Alder
Aspen
35.668
28.225 0
1.252
38.109
37.108 0
0
37.324
36.484 0
0
34.458
34.220 0
0
29.439
29.258 0
0
37.819
37.819 0
0
43.405
43.224 0
0
38.529
38.273 0.075
0
26.062
24.895 0
0.665
33.550
32.383 0
0
40.204
40.204 0
0
38.311
13.438
41.281
46.353

1.174
0.189
1.628
0.593

0
0
0
0

22.555
4.651
13.836
36.402

Willow
6.191
1.001
0.840
0.238
0.181
0
0.181
0.181
0.503
1.167
0
14.581
8.599
25.817
9.359

Figure 6.4 Illustration of the climate station opening west of plot A3 (mixedwood treatment). Area
maintained in open condition by annual brushing; size is approximately 20 × 20 m. Date of the picture:
August 24, 2018 (summer).
104

a)

b)

Figure 6.5 Illustration of the view of plot A3 (mixedwood treatment) from the area maintained in a
open condition in the May (a) and September (b), 2019.

105

Figure 6.6 Illustration of tree core extraction from a spruce tree in pure spruce treatment- plot A2 (a)
and a spruce tree in mixedwood treatment- plot A3 (b). Date of the pictures: August 24, 2018.

Figure 6.7 Daily average photosynthetically active radiation (PAR) for the 2007 year at a height of 3.0
m in the open area (west of plot A3) and at a height of 3.0 under the broadleaf canopy in the
mixedwood treatment plot A3.
106

7. Appendix 2: Results for Chapter 2
7.1. Random Forest (Liaw and Wiener 2002) results

Figure 7.1 Relative importance of the 17 annual climatic predictor variables using the full data set.
Each variable’s mean minimum depth refers to the average minimal depth that the variable within the
Random Forest regression tree. Variable with lower mean minimum depth indicates a split closer to
the root of the tree and an increased importance of the variable to spruce growth. The word “tree”
refers to the regression trees. The number of trees means the number of runs of regression (10000
runs). NA values indicate the variable was not used in an individual regression tree.

107

Figure 7.2 Relative importance of the 12 annual climatic predictor variables in the pure spruce
treatment.

108

Figure 7.3 Relative importance of the 12 annual climatic predictor variables in the mixedwood
treatment.

109

Figure 7.4 Relative importance of the 20 out of 46 seasonal climatic predictor variables in the pure
spruce treatment.

110

Figure 7.5 Relative importance of the 20 out of 46 seasonal climatic predictor variables in the
mixedwood treatment.

111

Figure 7.6 Relative importance of the 20 out of 138 monthly climatic predictor variables in the pure
spruce treatment.

112

Figure 7.7 Relative importance of the 20 out of 138 monthly climatic predictor variables in the
mixedwood treatment.

Table 7.1 Random Forest (Liaw and Wiener 2002) selection and ranking of the 5 most important
annual predictor variables for tree growth using the full data set.

Rank
1
2
3
4
5

Full data set
Annual variables
SoilTmp_50cm_Annual
ndays_AirTmp5_1.3m_Annual
AirTmp_3.0m_Annual
SolRad_3.0m_open_Annual
SWP_15cm_Annual

113

Table 7.2 Random Forest selection and ranking of the 5 most important annual, seasonal and monthly
predictor variables for tree growth across the treatment types.

Rank
1
2
3
4
5

Pure Spruce
Annual
Rain_open_Annual
ndays_AirTmp5_1.3m_Annual
SoilTmp_50cm_Annual
SolRad_3.0m_open_Annual
AirTmp_3.0m_Annual

Mixedwood
Annual
SolRad_3.0m_open_Annual
SWP_15cm_Annual
AirTmp_3.0m_Annual
SoilTmp_2.5cm_Annual
Rain_open_Annual

Rank
1
2
3
4
5

Seasonal
ndays_AirTmp5_1.3m_Summer
AirTmp_3.0m_Summer
SoilTmp_2.5cm_Fall
AirTmp_1.3m_Fall
SoilTmp_2.5cm_Winter

Seasonal
AirTmp_1.3m_Winter
SolRad_3.0_open_Winter
ndays_AirTmp0_1.3m_Fall
ndays_AirTmp5_1.3m_Fall
SWP_50cm_Spring

Rank
1
2
3
4
5

Monthly
SoilTmp_2.5cm_Jan
SolRad_3.0m_open_Feb
SolRad_3.0m_open_Aug
SolRad_3.0m_open_May
SWP_15cm_Oct

Monthly
ndays_AirTmp5_1.3m_Nov
ndays_AirTmp5_1.3m_Sep
Rain_open_Sep
SolRad_3.0_open_Dec
SolRad_3.0_open_Aug

114

Coefficient
1.3968
0.0257
-0.0026
-0.0764
0.0039
-0.3308

2.0994
0.1733
-0.0125
0.2722
-0.0387
-6.5465
-0.0045
0.0328
0.0082
0.0004
-0.0067
0.8765
0.0643

Full data set
Best model without interaction
Fixed effects
(Intercept)
AirTmp_3.0m_Annual

ndays_AirTmp5_1.3m_Annual

SoilTmp_50cm_Annual
SolRad_3.0m_open_Annual
SWP_15cm_Annual

Best model with interaction
(Intercept)
AirTmp_3.0m_Annual

ndays_AirTmp5_1.3m_Annual

SoilTmp_50cm_Annual
SolRad_3.0m_open_Annual
SWP_15cm_Annual
AirTmp_3.0m_Annual:ndays_AirTmp5_1.3m_Annual
AirTmp_3.0m_Annual:SoilTmp_50cm_Annual
AirTmp_3.0m_Annual:SolRad_3.0m_open_Annual
ndays_AirTmp5_1.3m_Annual:SolRad_3.0m_open_Annual
SoilTmp_50cm_Annual:SolRad_3.0m_open_Annual
SoilTmp_50cm_Annual:SWP_15cm_Annual
SolRad_3.0m_open_Annual:SWP_15cm_Annual

0.0706
0.0180
0.8879
0.0013
0.0088
0.0012
0.0001
0.0011
0.1159
0.0118

0.0058

0.8651
0.1744

0.0096
0.0007
0.0657

0.0009

Std.Error
0.1249
0.0058

1050
1050
1050
1050
1050
1050
1050
1050
1050
1050

1050

1050
1050

1057
1057
1057

1057

DF
1057
1057

3.8544
-2.1494
-7.3732
-3.5026
3.7243
6.7588
3.3411
-5.9104
7.5616
5.4302

-2.1498

2.4269
0.9940

-7.9301
5.5190
-5.0353

-3.0837

t-value
11.1801
4.4396

0.0001
0.0318
0.0000
0.0005
0.0002
0.0000
0.0009
0.0000
0.0000
0.0000

0.0318

0.0154
0.3205

0.0000
0.0000
0.0000

0.0021

p-value
0.0000
0.0000

Treatment
Plot in Treatment
Tree_ID in Plot in
Treatment
Residual

Random effects
Treatment
Plot in Treatment
Tree_ID in Plot in
Treatment
Residual

0.1300

2.30E-06

0.0392
2.94E-07

0.1394

2.46E-06

0.0201
3.08E-07

115

-1351.58

AIC
-1212.52

Table 7.3 Details of the best mixed effect growth models with and without two-way interaction terms using annual climatic predictor variables for
the full data set.

7.2. Mixed effect models: Full data – Annual

Figure 7.8 Interaction between annual air temperature and annual sum of number of days where air
temperature is above 5 °C on spruce growth using the full data set.

Figure 7.9 Interaction between annual air temperature and annual soil temperature on spruce growth
using the full data set.
116

Figure 7.10 Interaction between annual air temperature and annual solar radiation on spruce growth
using the full data set.

Figure 7.11 Interaction between annual sum of number of days where air temperature is above 5 °C
and solar radiation on spruce growth using the full data set.

117

Figure 7.12 Interaction between annual soil temperature and annual solar radiation on spruce growth
using the full data set.

Figure 7.13 Interaction between soil temperature and soil water potential on spruce growth using the
full data set.

118

Figure 7.14 Interaction between annual solar radiation and soil water potential on spruce growth
using the full data set.

119

0.8158
0.0251
0.0006
-0.0844
0.0028
-0.9732

(Intercept)

AirTmp_3.0m_Annual

Rain_open_Annual

SoilTmp_50cm_Annual

SolRad_3.0m_open_Annual

SWP_15cm_Annual

-6.5251
0.7194
0.0095
1.7126
0.0444
-20.9467
-0.0004
0.0788
-0.0193
-0.7322
-0.0030
-0.01194
8.569631
-0.05792

(Intercept)

AirTmp_3.0m_Annual

Rain_open_Annual

SoilTmp_50cm_Annual

SolRad_3.0m_open_Annual

SWP_15cm_Annual

AirTmp_3.0m_Annual:Rain_open_Annual

AirTmp_3.0m_Annual:SoilTmp_50cm_Annual

AirTmp_3.0m_Annual:SolRad_3.0m_open_Annual

AirTmp_3.0m_Annual:SWP_15cm_Annual

Rain_open_Annual:SoilTmp_50cm_Annual

Rain_open_Annual:SWP_15cm_Annual

SoilTmp_50cm_Annual:SWP_15cm_Annual

SolRad_3.0m_open_Annual:SWP_15cm_Annual

Best model with interaction

Coefficient

Fixed effects

Best model without interaction

Pure Spruce

0.031185

1.313505

0.001829

0.0003

0.2873

0.0033

0.0192

0.0001

4.3197

0.0083

0.2953

0.0011

0.1970

1.2683

0.1165

0.0007

0.0141

0.0001

0.0081

0.0662

Std.Error

526

526

526

526

526

526

526

526

526

526

526

526

526

526

534

534

534

534

534

534

DF

-1.8572

6.5242

-6.5281

-8.5980

-2.5487

-5.8276

4.1043

-4.4859

-4.8491

5.3549

5.7991

8.9556

3.6512

-5.1450

-8.3506

4.0369

-5.9740

10.1967

3.0870

12.3235

t-value

0.0638

0.0000

0.0000

0.0000

0.0111

0.0000

0.0000

0.0000

0.0000

0.0000

0.0000

0.0000

0.0003

0.0000

0.0000

0.0001

0.0000

0.0000

0.0021

0.0000

p-value

Residual

Tree_ID in Plot

Plot

Residual

Tree_ID in Plot

Plot

Random effects

0.1040

1.15E-06

1.13E-06

0.1304

1.42E-06

1.40E-06

-927.08

-685.46

AIC

120

Table 7.4 Details of the best mixed effect growth models with and without two-way interaction terms using annual climatic predictor variables for
the pure spruce treatment.

7.3. Mixed effect models: Across the treatments – Annual

1.2427
0.0304
-0.0002
-0.0918
0.0020

(Intercept)

AirTmp_3.0m_Annual

Rain_open_Annual

SoilTmp_50cm_Annual

SolRad_3.0m_open_Annual

0.9850
-2.5458
0.0048
-0.7511
0.1165
-1.4004
0.0014
0.3420
-6.1797
0.0004
-0.0002
-0.0051
-1.8687
0.3574

(Intercept)

AirTmp_3.0m_Annual

Rain_open_Annual

SoilTmp_50cm_Annual

SolRad_3.0m_open_Annual

SWP_15cm_Annual

AirTmp_3.0m_Annual:Rain_open_Annual

AirTmp_3.0m_Annual:SoilTmp_50cm_Annual

AirTmp_3.0m_Annual:SWP_15cm_Annual

Rain_open_Annual:SoilTmp_50cm_Annual

Rain_open_Annual:SolRad_3.0m_open_Annual

SoilTmp_50cm_Annual:SolRad_3.0m_open_Annual

SoilTmp_50cm_Annual:SWP_15cm_Annual

SolRad_3.0m_open_Annual:SWP_15cm_Annual

Best model with interaction

Coefficient

Fixed effects

Best model without interaction

Mixedwood

0.0359

0.5077

0.0025

0.0000

0.0001

0.6965

0.0385

0.0003

1.7537

0.0134

0.1899

0.0007

0.2737

0.8783

0.0007

0.0131

0.0001

0.0085

0.0590

Std.Error

510

510

510

510

510

510

510

510

510

510

510

510

510

510

519

519

519

519

519

DF

9.9661

-3.6807

-1.9957

-11.0174

3.1917

-8.8730

8.8921

4.6572

-0.7985

8.6659

-3.9553

7.4150

-9.3007

1.1215

3.1247

-7.0127

-3.5242

3.5911

21.0472

t-value

0.0000

0.0003

0.0465

0.0000

0.0015

0.0000

0.0000

0.0000

0.4249

0.0000

0.0001

0.0000

0.0000

0.2626

0.0019

0.0000

0.0005

0.0004

0.0000

p-value

Residual

Tree_ID in Plot

Plot

Residual

Tree_ID in Plot

Plot

Random effects

0.1084

1.24E-06

1.26E-06

0.135753

1.52E-06

1.53E-06

121

-854.49

-623.24

AIC

Table 7.5 Details of the best mixed effect growth models with and without two-way interaction terms using annual climatic predictor variables for
the mixedwood treatment.

Figure 7.15 Interaction between annual air temperature and annual rainfall on spruce growth in the
pure spruce treatment.

Figure 7.16 Interaction between annual air temperature and annual soil temperature on spruce growth
in the pure spruce treatment.
122

Figure 7.17 Interaction between annual air temperature and annual solar radiation on spruce growth
in the pure spruce treatment.

Figure 7.18 Interaction between annual air temperature and annual soil water potential on spruce
growth in the pure spruce treatment.

123

Figure 7.19 Interaction between annual rainfall and annual soil temperature on spruce growth in the
pure spruce treatment.

Figure 7.20 Interaction between annual rainfall and soil water potential on spruce growth in the pure
spruce treatment.

124

Figure 7.21 Interaction between annual soil temperature and annual soil water potential on spruce in
the pure spruce treatment.

Figure 7.22 Interaction between annual solar radiation and annual soil water potential on spruce
growth in the pure spruce treatment.

125

Figure 7.23 Interaction between annual air temperature and annual rainfall on spruce growth in the
mixedwood treatment.

Figure 7.24 Interaction between annual air temperature and annual soil temperature on spruce growth
in the mixedwood treatment.

126

Figure 7.25 Interaction between annual air temperature and annual soil water potential on spruce
growth in the mixedwood treatment.

Figure 7.26 Interaction between annual rainfall and annual soil temperature on spruce growth in the
mixedwood treatment.

127

Figure 7.27 Interaction between annual rainfall and solar radiation on spruce growth in the
mixedwood treatment.

Figure 7.28 Interaction between annual soil temperature and annual solar radiation on spruce growth
in the mixedwood treatment.

128

Figure 7.29 Interaction between annual soil temperature and annual soil water potential on spruce
growth in the mixedwood treatment.

Figure 7.30 Interaction between annual solar radiation and annual soil water potential on spruce
growth in the mixedwood treatment.

129

Coefficient
1.0105
0.0147
0.1013

1.0740
-0.0028
0.3975
0.1128
-0.0964

Coefficient
1.3854
-0.0331
-0.6713

1.2374
0.2180
-0.0660
-5.6627
-3.5513
1.8833
0.2510

Pure Spruce
Best model without interaction
Fixed effects
(Intercept)
AirTmp_3.0m_Spring
SWP_15cm_Summer

Best model with interaction
(Intercept)
AirTmp_3.0m_Spring
SWP_15cm_Spring
SWP_15cm_Summer
AirTmp_3.0m_Spring:SWP_15cm_Spring

Mixedwood
Best model without interaction
Fixed effects
(Intercept)
AirTmp_3.0m_Summer
SWP_15cm_Spring

Best model with interaction
(Intercept)
AirTmp_3.0m_Spring
AirTmp_3.0m_Summer
SWP_15cm_Spring
SWP_15cm_Summer
AirTmp_3.0m_Spring:SWP_15cm_Spring
AirTmp_3.0m_Summer:SWP_15cm_Summer

0.3521
0.0223
0.0254
0.5240
1.0034
0.1907
0.0719

Std.Error
0.1356
0.0100
0.1329

0.0227
0.0070
0.1202
0.0321
0.0304

Std.Error
0.0122
0.0038
0.0314

348
348
348
348
348
348
348

DF
352
352
352

446
446
446
446
446

DF
448
448
448

3.5143
9.7813
-2.6042
-10.8067
-3.5393
9.8758
3.4903

t-value
10.2172
-3.2959
-5.0507

47.2908
-0.3985
3.3078
3.5165
-3.1679

t-value
82.8069
3.8931
3.2219

0.0005
0.0000
0.0096
0.0000
0.0005
0.0000
0.0005

p-value
0.0000
0.0011
0.0000

0.0000
0.6905
0.0010
0.0005
0.0016

p-value
0.0000
0.0001
0.0014

Plot
Tree_ID in Plot
Residual

Random effects
Plot
Tree_ID in Plot
Residual

Plot
Tree_ID in Plot
Residual

Random effects
Plot
Tree_ID in Plot
Residual

1.51E-06
1.65E-06
0.1326

1.59E-06
1.71E-06
0.1490

2.09E-06
1.83E-06
0.1473

2.10E-06
1.84E-06
0.1490

130

-442.21

AIC
-360.39

-460.61

AIC
-453.54

Table 7.6 Details of the best mixed effect growth models with and without two-way interaction terms using seasonal climatic predictor variables
for the pure spruce and mixedwood treatments.

7.4. Mixed effect models: Across the treatments – Seasonal

Figure 7.31 Interaction between air temperature and soil water potential during spring and summer
on spruce growth in the pure spruce treatment.

131

Figure 7.32 Interaction between air temperature and soil water potential during spring and summer
on spruce growth in the mixedwood treatment.

132

Coefficient
0.9776
-0.0108
0.0168
-0.1265

1.4605
-0.0519
0.0252
1.2048
-0.1302
-0.0859

Coefficient
1.1042
-0.0134
0.0392
-1.4246

1.9536
-0.0746
1.4040
-1.2312
-0.1003

Fixed effects

Variable
(Intercept)
AirTmp_3.0m_Aug
AirTmp_3.0m_May
SWP_15cm_May

Best model with interaction
(Intercept)
AirTmp_3.0m_Aug
AirTmp_3.0m_May
SWP_15cm_Aug
SWP_15cm_May
AirTmp_3.0m_Aug:SWP_15cm_Aug

Mixedwood
Best model without interaction
Fixed effects
(Intercept)
AirTmp_3.0m_Aug
SWP_15cm_Aug
SWP_15cm_May

Best model with interaction
(Intercept)
AirTmp_3.0m_Aug
SWP_15cm_Aug
SWP_15cm_May
AirTmp_3.0m_Aug:SWP_15cm_Aug

Pure Spruce
Best model without interaction

0.2743
0.0199
0.4183
0.1962
0.0307

Std.Error
0.0890
0.0069
0.0259
0.1893

0.0848
0.0079
0.0049
0.1844
0.0426
0.0130

0.0450
0.0050
0.0049
0.0407

Std.Error

370
370
370
370
370

DF
371
371
371
371

474
474
474
474
474
474

476
476
476
476

DF

7.1211
-3.7430
3.3562
-6.2744
-3.2686

t-value
12.4088
-1.9367
1.5143
-7.5239

17.2183
-6.5617
5.1156
6.5348
-3.0540
-6.5981

21.7423
-2.1741
3.4155
-3.1086

t-value

0.0000
0.0002
0.0009
0.0000
0.0012

p-value
0.0000
0.0535
0.1308
0.0000

0.0000
0.0000
0.0000
0.0000
0.0024
0.0000

0.0000
0.0302
0.0007
0.0020

p-value

Plot
Tree_ID in Plot
Residual

Random effects
Plot
Tree_ID in Plot
Residual

Plot
Tree_ID in Plot
Residual

Plot
Tree_ID in Plot
Residual

Random effects

1.47E-06
1.54E-06
0.1401

1.51E-06
1.57E-06
0.1420

1.84E-06
1.71E-06
0.1421

1.89E-06
1.77E-06
0.1481

-425.48

AIC
-416.81

AIC
-523.92

-485.45

AIC

133

Table 7.7 Details of the best mixed effect growth models with and without two-way interaction terms using monthly climatic predictor variables
for the pure spruce and mixedwood treatments.

7.5. Mixed effect models: Across the treatments – Monthly

Figure 7.33 Interaction between air temperature and soil water potential in May and August on
spruce growth in the pure spruce treatment.

134

Figure 7.34 Interaction between air temperature and soil water potential in May and August on
spruce growth in the mixedwood treatment.

135

7.6. Microclimate comparison between pure spruce and mixedwood treatments by
month using daily data from 1999 to 2017

Figure 7.35 Comparison of daily means of air temperature at height of 3.0 m between pure spruce and
mixedwood treatments by month from 1999 to 2017. The box plot visually shows the distribution of the
data and skewness through displaying the interquartile range (box), median (horizontal line), whiskers
(vertical lines) and outlines (circles).

136

Figure 7.36 Comparison of daily means of soil temperature at a depth of 15 cm between pure spruce and
mixedwood treatments by month from 1999 to 2017.

Figure 7.37 Comparison of daily mean of soil water potential (SWP) at a depth of 15 cm between pure
spruce and mixedwood treatments by month from 1999 to 2017.

137

8. Appendix 3: Results for Chapter 3

8.1. Microclimate comparison between pure spruce and mixedwood treatments by
month using daily data from 2007 to 2018

Figure 8.1 Comparison of daily mean of air temperature height of 3.0 m between pure spruce and
mixedwood treatments by month from 2007 to 2018. The box plot visually shows the distribution of the
data and skewness through displaying the interquartile range (box), median (horizontal line), whiskers
(vertical lines) and outlines (circles).

138

Figure 8.2 Comparison of daily mean of soil temperature at a depth of 15 cm between pure spruce and
mixed by month from 2007 to 2018.

Figure 8.3 Daily sum of precipitation in the open area by month from 2007 to 2018.

139

Figure 8.4 Daily mean of solar radiation in the open area by month from 2007 to 2018. The box plot
visually shows the distribution of the data and skewness through displaying the interquartile range (box),
median (horizontal line), whiskers (vertical lines) and outlines (circles).

140

8.2. Mixed effect models: throughout the growing season

Figure 8.5 Interaction between solar radiation and rainfall in the pure spruce treatment (a) and
mixedwood treatment (b).

141

Figure 8.6 Interaction between air temperature and SWP in the pure spruce treatment (a) and mixedwood
treatment (b).

142

Figure 8.7 Interaction between soil temperature and SWP in the pure spruce treatment (a) and
mixedwood treatment (b).

143

Figure 8.8 Interaction between solar radiation and SWP in the pure spruce treatment (a) and mixedwood
treatment (b).

144

8.3. Mixed effect models: between seasons

Figure 8.9 Interaction between solar radiation and rainfall in spring (a), early-summer (b), and latesummer (c) in the pure spruce treatment.

145

Figure 8.10 Interaction between air temperature and SWP in spring (a), early-summer (b), and latesummer (c) in the pure spruce treatment.

146

Figure 8.11 Interaction between soil temperature and SWP in spring (a), early-summer (b), and latesummer (c) in the pure spruce treatment.

147

Figure 8.12 Interaction between solar radiation and SWP in spring (a), early-summer (b), and latesummer (c) in the pure spruce treatment.

148

Figure 8.13 Interaction between solar radiation and rainfall in spring (a), early-summer (b), and latesummer (c) in the mixedwood treatment.

149

Figure 8.14 Interaction between air temperature and SWP in spring (a), early-summer (b), and latesummer (c) in the mixedwood treatment.

150

Figure 8.15 Interaction between soil temperature and SWP in spring (a), early-summer (b), and latesummer (c) in the mixedwood treatment.

151

Figure 8.16 Interaction between solar radiation and SWP in spring (a), early-summer (b), and latesummer (c) in the mixedwood treatment.

152

Plot A2 (pure
spruce treatment)

Plot A3
(mixedwood
treatment)

20×20 m opening
west of plot A3

Location

Soil water potential

Soil temperature

Air temperature

Sapflow c

Soil water potential

Silicon pyranometer

Cu-Co thermocouple

Gypsum block

Cu-Co thermocouple

Home built/ fine wire 36
Cu-Co thermocouple
AWG

2
2
3
3
3
3
3

Home built/twisted
soldered wire

Dynamax/TDP-30

Constant power; Delta T

Gypsum block
Campbell Scientific/223
sensors manufactured by
Gypsum block
Delmhorst
Gypsum block
Gypsum block

Home built/twisted
soldered wire

Home built/ fine wire 36
Cu-Co thermocouple
AWG

Li-Cor/LI200S

3

2
2
3
3
3
3
3
3
3
2
2

+130
+300
-2.5
-15
-50
-2.5
-15
-50
-15
-15
-15
Spruce tree
(N or S
aspect) c
+130
+300
-2.5
-15
-50
-2.5
-15

Air temperature

Soil temperature

2

+300

Solar radiation

1

Home built/ fine wire 36
Cu-Co thermocouple
AWG
Sierra Misco/2501 or
Tipping bucket
TE525m

2
2

Sensor type

Sensor make/model

Reps

+60 or +80

Position
(cm) a
+130
+300

Rainfall b

Air temperature

Measurement
variable

Table 9.1 Equipment descriptions for 1999-2018 climate stations.

9. Appendix 4: Metadata for Chapter 2 and Chapter 3

Unshielded

new 2011
new 2012
new 2018

Unshielded

Serial # 22636 &
22634

Unshielded

Other info

153

2013
2014
2016
2017
2018

2012

2011

Year
2008
2009
2010

Day and Month
9 April to 19 May
7 June to 19 September
10 June to 24 September for tree #2 of plot A3 (mixedwood treatment)
All year for tree #2 of plot A3 (mixedwood treatment)
From 28 August
31 May to 22 September for tree #2 of plot A3 (mixedwood treatment)
7 to 30 May
18 April to 21 September for tree #2 of plot A3 (mixedwood treatment)
13 May to 11 September for tree #2 of plot A3 (mixedwood treatment)
From 21 July for tree #2 of plot A3 (mixedwood treatment)
From June 11
16 July to 22 September for tree #1 of plot A3 (mixedwood treatment)

Table 9.2 Sap flow velocity data removed from the data set prior to analysis.
Rationale
Bad data unexplained reason
Battery power to sensors failed due to broken wire
Bad data wiring panel problem
Bad data wiring panel problem
Power for sapflow sensors ran out
Bad data wiring panel problem
No battery power due to bad connection with solar panel
Sensor not working
Sensor not working
Sensor not working
Bad data unexplained reason
Sensor damaged by a bear

154

-50
3
Campbell Scientific/223
-15
3
Gypsum block
new 2011
sensors manufactured by
-15
2
Gypsum block
new 2012
Delmhorst
-15
2
Gypsum block
new 2018
Spruce tree
Sapflow c
(N or S
3
Dynamax/TDP-30
Constant power; Delta T
aspect) c
a
Values for height (+) indicate height (cm) above the ground surface (regardless of whether mineral soil or organic material). Values for soil depth
(-) indicate depth from the mineral soil forest floor interface.
b
Raingauge was not recording all tips from the tipping bucket during 2015 growing season. It was removed on 22 Sep 2015 and replaced with a
TE 525m raingauge on 26 May 2016.
c
Sap flow velocity was measured on the north side of each spruce tree from 2007 to 2016, and on the south side of each spruce tree in 2017 and
2018.

1999

x

x
x
x
x

Air temperature

Soil temperature

Soil water potential (1999 gypsum blocks)

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

x

x

x

Soil water potential (2018 gypsum blocks)

Soil water potential (2012 gypsum blocks)

Soil water potential (2011 gypsum blocks)

x

x
x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x x
x

x
x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

Soil water potential (1999 gypsum blocks)

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

Soil temperature

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

Air temperature

x

x

x

x

x

x

x

x
x

x

x

x

x

x

x

x
x

x

x

x

x

x

x

Sapflow (N or S aspect) a
x

x

x

x

x

x

x

x

x

x

x

x

x

x

x

155

Sapflow (N or S aspect) a
x x x x x x x x x x x x
a
Sap flow velocity was measured on the north side of each spruce tree from 2007 to 2016, and on the south side of each spruce tree in 2017 and
2018.

Plot A2 (pure
spruce
treatment)

2015

Soil water potential (2018 gypsum blocks)

Soil water potential (2012 gypsum blocks)

Soil water potential (2011 gypsum blocks)

x

x

x

x

x

Plot A3
(mixedwood
treatment)

2016

Year post-planting 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

Measurement year

2000

20×20 m
Air temperature
opening
Rainfall
adjacent to plot
Solar radiation
A3

Climate station
location

2017

Table 9.3 Years that measurements were included in the data set (x). Years that are not checked with “x” were either not measured in the year or
presented measurement error over the entire year due to equipment problems and were not included in the data set. Hourly measurement errors
were removed from the data set. Colours correspond to climate station locations.
2018