THE HABITAT ECOLOGY OF THE POST-METAMORPHIC COASTAL TAILED FROG (Ascaphus truei) by Alexandria L. McEwan HBSc (Biology), Lakehead University, 2007 THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE IN NATURAL RESOURCES AND ENVIRONMENTAL STUDIES (BIOLOGY) UNIVERSITY OF NORTHERN BRITISH COLUMBIA December 2014 © Alexandria L. McEwan, 2014 UMI Number: 1526497 All rights reserved INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted. In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. Di!ss0?t&iori P iiblist’Mlg UMI 1526497 Published by ProQuest LLC 2015. Copyright in the Dissertation held by the Author. Microform Edition © ProQuest LLC. All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code. ProQuest LLC 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106-1346 A b st r a c t The coastal tailed frog (Ascaphus truei) has a biphasic life-history that includes the use o f high-elevation streams and adjacent old growth forests. This species is at risk from a variety o f land-use activities that result in younger and more fragmented terrestrial habitats. I investigated the fine-scale spatial ecology o f post-metamorphic tailed frog populations as revealed by seasonal variation in activity and movement. I used an Information Theoretic Model Comparison approach to relate tailed frog captures in pitfall traps to sets of explanatory factors and infer habitat use by adult tailed frogs located with VHF radio transmitters. Trap data suggested that variation in the relative abundance and distribution was attributed to time o f year, associated with reproductive activity; seasonal trends in local climate; canopy closure; and distance to the larval stream. The most parsimonious logistic regression model for both years o f trapping data demonstrated ‘good’ predictability (Area Under the Curve (AUC) > 0.7). Data from 24 tailed frogs with radio transmitters (8 males, 16 females) fitted to a resource selection function suggested that tailed frogs were positively associated with habitats containing decayed coarse woody debris (CWD), less light, adjacency to the larval stream, wet site types, and cool temperatures (11 -12° C). When compared to males, females were more selective in their choice o f habitat (AUC > 0.7). Overall, the relationship between distribution and activity o f tailed frogs was reflective of reproductive phenology in combination with micro-site or broader climatic conditions. Thus, habitat protections should reflect the behavior and habitat o f the species as it varies across spatiotemporal scales. Table of Contents Chapter 1: General Introduction Species Distribution.............................................................................................................................2 Species Ecology and Biology............................................................................................................ 2 Conservation Issues............................................................................................................................. 3 Conservation M anagement.................................................................................................................4 Previous Research................................................................................................................................ 5 Research Purpose................................................................................................................................. 7 Study Area............................................................................................................................................ 8 Chapter 2: Distribution and Movement Patterns o f Post-Metamorphic Coastal Tailed Frogs Introduction......................................................................................................................................... 19 Chapter Predictions...............................................................................................................21 Methods............................................................................................................................................... 22 Study area............................................................................................................................... 22 Data collection...................................................................................................................... 24 Statistical methods................................................................................................................25 Data analysis............................................................................................................ 25 Model selection........................................................................................................ 31 Model and variable evaluation............................................................................... 31 Results..................................................................................................................................................32 Capture statistics................................................................................................................... 32 Variation in microclimate....................................................................................................32 Orientation and movement...................................................................................................35 Distance o f capture from stream......................................................................................... 35 Forest retention treatment....................................................................................................40 Temporal pattern o f movement........................................................................................... 42 Statistical models.................................................................................................................. 42 Discussion........................................................................................................................................... 44 Conclusion.......................................................................................................................................... 53 Recommendations.............................................................................................................................. 54 Managing forests for the tailed frog....................................................................................54 Chapter 3: Habitat Selection by the Coastal Tailed Frog o f Northwestern BC Introduction......................................................................................................................................... 59 Chapter Predictions...............................................................................................................62 Methods................................................................................................................................................63 Study area............................................................................................................................... 63 Data collection...................................................................................................................... 65 Statistical methods................................................................................................................67 Data analysis.............................................................................................................67 Spatial scales.............................................................................................................68 Model selection and evaluation.............................................................................71 Patterns o f movement............................................................................................. 73 Results..................................................................................................................................................73 Radio-telemetered animals...................................................................................................76 Gender differences in movement and space used............................................................ 74 Forest retention treatment....................................................................................................74 Temporal movement............................................................................................................ 78 Resource selection................................................................................................................82 Discussion........................................................................................................................................... 85 Conclusion..........................................................................................................................................93 Recommendations Forest management and the tailed frog..............................................................................93 Chapter 4: General Summary........................................................................................................... 97 Literature Cited.................................................................................................................................104 L is t o f F i g u r e s Figure 1: Distribution o f the Coastal Western Hemlock Biogeoclimatic Zone in British Columbia. Figure from Meidinger and Pojar (1991). Grey star indicates location o f research site...................................................................................................................... 9 Figure 2: Distribution o f the Interior Cedar-Hemlock Biogeoclimatic Zone in British Columbia. Figure from Meidinger and Pojar (1991). Grey star indicates location o f research area....................................................................................................................11 Figure 3: Location o f the study watersheds, Gosling Creek (GO), Ascaphus Creek (AS) and Kleanza Creek, (KL) and treatments (1- old growth, 2- forest retention buffer and 3- clearcut) near Terrace, BC............................................................................................................ 14 Figure 4: Trap arrays for tailed frogs applied to 3 forest retention treatments found within two watersheds (A). Each array (X) contained 4 traps centered in 2 perpendicular arms o f 1Om o f drift fencing for a total o f 288 traps. Each trap within the array was used to monitor direction o f tailed frog movement in relation to the known larval stream, towards (1), down (2),away (3) and up (4) stream (B)..........................26 Figure 5: Weight (grams), snout-vent-length (mm) and shank lengths (mm) including 95% confidence intervals o f adult tailed frogs in 2 watersheds located east o f Terrace, BC, during 2012 and 2013................................................................................................................34 Figure 6: Microclimate arrays demonstrating the temperature variations between 4 forest retention treatments (old growth, forest retention buffer, regeneration and clearcut) located in the Gosling (A) and Kleanza (B) watershed, east o f Terrace, BC, during 2012......................................................................................................................................... 36 Figure 7: Microclimate arrays demonstrating the temperature variations between 4 forest retention treatments (old growth, forest retention buffer, regeneration and clearcut) located in the Gosling (A) and Kleanza (B) watershed, east o f Terrace, BC, during 2013......................................................................................................................................... 37 Figure 8: Proportions o f tailed frogs captured for each age class relative to the direction o f movement within each forest retention treatment (old growth, forest retention buffer, regeneration, and clearcut) for two watersheds located east o f Terrace, BC, during 2012 and 2013. In the old growth a total o f 59 tailed frogs were captured (48 adults, 7 juveniles and 4 metamorphs). The forest retention buffer had 53 tailed frogs (38 adults, 11 juveniles and 4 metamorphs), the regeneration treatment had 26 tailed frogs captured (18 adults, and 8 juveniles), and the clearcut had 22 individuals captured (7 adults, 4 juveniles and 11 metamorphs).................................................38 Figure 9: Proportions o f tailed frogs captured within each forest retention treatment by age class and the distance from stream o f trap array for 2 watersheds east o f Terrace, BC during 2012 (A) and 2013 (B). In 2012, 52 tailed frogs captured in the old growth, 39 individuals in the forest retention buffer, 19 individuals in the regeneration treatment and 19 individuals in the clearcut....................................................................................................39 Figure 10: Total number o f reproductive and pre-reproductive tailed frogs captured by month during 2012 in 2 watersheds located east o f Terrace, BC. Numbers in brackets represent total captures within a m onth.......................................................................... 41 Figure 11: VHF radio transmitter (2) with belly-belt attachment (1) and antenna (3) used to relocate coastal tailed frogs.......................................................................................................... 66 Figure 12: Illustration o f the nested sampling design for recording habitat characteristics at used and available locations for tailed frogs at the micro- (l-m 2plot) and macro­ habitat (25-m2 plot) scales. Arrows represent the 5-m coarse woody debris transect lines......................................................................................................................................................69 Figure 13. Comparison o f the mean daily distance (meters) moved (95% confidence intervals) for 24 adult tailed frogs relocated in 3 forest retention treatments (old growth, forest retention buffer, and clearcut) for 3 watersheds east o f Terrace, BC, in 2011 and 2012. Numbers in brackets indicate total number (n) of frogs relocated within each forest retention treatment............................................................................................. 76 Figure 14: Comparison o f the mean distance (meters) traveled from the origin o f capture (95% confidence intervals) for 24 adult tailed frogs relocated in 3 forest retention treatments (old growth, forest retention buffer and clearcut) for 3 watersheds east o f Terrace, BC, in 2011 and 2012. Numbers in brackets indicate total number (n) o f tailed frogs relocated within each forest retention treatment............................................ 77 Figure 15: Mean distance (m) moved (95% confidence intervals) from the stream edge for 24 adult tailed frogs relocated in 3 forest retention treatments (old growth, forest retention buffer, and clearcut) for 3 watersheds located east o f Terrace, BC, during 2011 and 2012.....................................................................................................................................79 L is t o f T a b l e s Table 1: History o f forest harvesting that has occurred within the watersheds for all forest retention treatments (old growth (OG), forest retention buffer (BF), regeneration (RG) and clearcut (CC)) east o f Terrace, BC, where coastal tailed frog populations were monitored...................................................................................................................................15 Table 2: Trap session dates and observed seasonal life cycle o f the tailed frog for 2 watersheds east o f Terrace, BC, surveyed during 2012 and 2013.............................................. 27 Table 3: Independent variables used to model the occurrence and abundance o f tailed frogs captured in pitfall traps in 2 watersheds east o f Terrace, BC, during 2012 and 2013....................................................................................................................................................30 Table 4: Total number of tailed frogs captured and the catch per unit effort (CPUE) for each age class and 4 forest retention treatments (old growth, forest retention buffer, regeneration, and clearcut) for 2 watersheds located east o f Terrace, BC, during 2012 and 2013. CPUE represents the number of individuals/100 trap nights................................... 33 Table 5: Most parsimonious logistic regression models (AICc) for tailed frog capture data collected with pitfall arrays at 6 treatments located east o f Terrace, BC, in 2012 and 2013. Top ranked models represented 95% o f the AICc weight........................................43 Table 6: Coefficients and measure o f statistical significance, including 95% confidence intervals (Cl), for covariates from the most parsimonious logistic regression model (Table 5) for tailed frog capture data collected in 2012 and 2013 from pitfall arrays east o f Terrace, BC...........................................................................................45 Table 7: Independent variables used to derive RSF models for the coastal tailed frog in northwestern BC at the micro- (Mi) and macro-habitat (Ma) scales......................................72 Table 8: Telemetry data for 24 adult tailed frogs identifying the average number o f days followed, weight (grams) and the percent body weight o f the tag for each forest retention treatment in 3 watersheds located east o f Terrace, BC, during 2011 and 2012. Numbers in brackets are standard errors (SE)................................................................... 75 Table 9: Radio telemetry data for 24 adult tailed frogs representing the proportions o f relocations associated with 4 site moisture types (dry, mesic, subhygric or hygric) for each gender located in 3 forest retention types (old growth, forest retention buffer, and clearcut) for 3 watersheds east o f Terrace, BC, during 2011 and 2012. Numbers in brackets represent sample sizes...................................................................................................80 Table 10. Mean distance traveled (m) per day for 24 adult tailed frogs relocated during 2011 and 2012 in 3 watersheds east o f Terrace, BC, within 3 forest retention treatments (old growth, forest retention and clearcut). Numbers in brackets represent standard errors (SE)........................................................................................................................................... 81 Table 11: Seasonal daily movement rates for 24 adult tailed frogs located in 3 forest retention treatments (old growth, forest retention buffer and clearcut) for 3 watersheds east o f Terrace, BC, during 2011 and 2012. Dates associated with season are located in Table 2. Numbers in brackets represent standard errors and sample sizes (n).................... 81 Table 12: Most parsimonious logistic regression models (AICc) for relocation data o f 24 tailed frogs in 3 watersheds east o f Terrace, BC, in 2011 and 2012. Top ranked models represented > 95% o f the AlCc weight (AICc w); the area under the curve (AUC, Standard Error) represents the measure o f predictability for each m odel..................... 83 Table 13: Coefficients and measure o f statistical significance, including 95% confidence intervals (Cl), for covariates from the most parsimonious logistic regression models (Table 13) generated using relocation data for tailed frogs collected in 2011 and 2012 from 3 watersheds east o f Terrace, B C ........................................................... 84 iv L is t o f A p p e n d ic e s Appendix I- Diagram demonstrating the pitfall trap with plastic insert and escape rope. During the 2012 spring session, the escape rope did not contain knotted sections. After discussion with M. Todd, knotted sections were implemented for the remainder o f the 2012 trap session and all o f 2013 sessions...................................................................... 114 Appendix II- Logistic regression models used for the 2012 trap data conducted in 2 watersheds located east o f Terrace, BC, identifying the distribution o f tailed frogs within forest retention treatments.................................................................................................115 Appendix III- Logistic regression models used for the 2013 trap data conducted in 2 watersheds located east o f Terrace, BC, identifying the distribution o f tailed frogs within forest retention treatments................................................................................................117 Appendix IV: Temperature and days since last rain at the 5- and 80-m arrays for 4 forest retention treatments in 2 watersheds located east o f Terrace, BC during 2012..........119 Appendix V: Temperature and days since last rain for 5-m and 80-m arrays in 4 forest retention treatments in 2 watersheds located east o f Terrace, BC during 2013................................................................................................................................... 120 Appendix VI: Graphical representation o f optimal temperature o f used locations by the tailed frog in 3 watersheds located east o f Terrace, BC, during 2011 and 2012. Largest bar represents the optimal temperature for the model based on the Gaussian term (linear and quadratic). The pooled tailed frog model and the female tailed frog model are associated with 11°C, while the male tailed frog data is associated with 12°C.........................................................................................................................................121 Appendix VII: Proportion o f used locations associated with light levels > 103 and < 103 lux for each gender across 3 forest retention treatments (old growth, forest retention buffer and clearcut) for 24 radio telemetered frogs in 3 watersheds east of Terrace, BC during 2011 and 2012................................................................................................122 v A cknow ledgem ents I wish to thank the Habitat Conservation Trust Foundation, the BC Ministry o f Forests, Lands and Natural Resource Operations, BC Ministry o f Environment, the Natural Sciences and Engineering Research Council (NSERC) o f Canada, The Grand River Post-Secondary Education Office, and the University o f Northern British Columbia for providing the funds to support my education and this research. I would like to thank Kelly Houlden and the Safety Officers o f the Coast Mountains (formerly Kalum) Resource District in Terrace for their hours o f safety assurance. I am grateful to Dr. Chris Johnson, for his support, advice and assistance throughout the project. I would like to thank my committee members Melissa Todd, Dr. Pumima Govindarajulu, and Dr. Mark Shrimpton for their many contributions both in the field and in the technical aspects of this project. I am grateful to the countless volunteers and assistants for their time and interest, especially Kristal Johnston (Golob), Ben Millard-Martin and Midoli Bresch for their hard work, positive attitude and comradery in the field. I would like to thank my colleagues in the Natural Resources and Environmental Studies graduate cohort for their support and friendship. I am grateful to my family, Boaz Crees and Merlin for all their love and encouragement at the end of each day. vi C hapter 1 G e n e r a l In t r o d u c t i o n In t r o d u c t i o n Species Distribution The tailed frog genus, Ascaphus, is endemic to the Pacific Northwest o f North America and inhabits the cold-flowing streams associated with those forests (Brown 1975; Wallace and Diller 1998; Bury and Adams 1999; Ritland et al. 2000; Wahbe et al. 2004; Matsuda and Richardson 2005; Hayes et al. 2006). There are 2 species within North America, Ascaphus montanus and A. truei. These species occupy environmental regimes extending from the drier, colder climate found in the interior Rocky Mountains (A. montanus-, Daugherty and Sheldon 1982) to the wet climate along the Pacific coast (A. truei-, Bury 1968). The range o f A. truei extends from northwestern California to the Nass River, north o f Prince Rupert, British Columbia (BC; Dupuis and Bunnell 1997). The range o f A. montanus is more interior from southeastern BC and southeastern Washington into southcentral Idaho and northeastern Oregon (Leonard et al. 1993; Nielson et al. 2001; Stebbins 2003). Species Ecology and Biology Past research on the biology and ecology o f A. truei has focused on the larval life stage (Bury and Adams 1999; Dupuis and Steventon 1999; Wahbe and Bunnell 2001; Hayes et al. 2006; Karracker et al. 2006; Burkholder and Diller 2007). As such, this species is known to be associated with mountainous lotic environments flowing through old-growth stands containing a well-vegetated understory (Green and Campbell 1984). Under laboratory conditions, Brown (1975) reported that for healthy embryonic development streams must maintain temperatures from 5-18.5°C; however, temperatures in more northerly climates may exceed this minimum during the winter. Larvae can take 1-4 years to fully metamorphose, 2 but the variation in timing is reflective o f the geographic range o f this species (Bury and Adams 1999). Ascaphus are the longest-lived anuran (15-20 years) and reach sexual maturity at 8-9 years (Daugherty and Sheldon 1982; Brown 1990). Breeding occurs in early fall (September to October) depending on the geographic location (Green and Campbell 1984). The genus has evolved internal fertilization as a reproductive strategy because breeding occurs in flowing streams or small pools close to larval streams (Green and Campbell 1984; Bull and Carter 1996; Ritland et al. 2000). Clutch sizes for A. truei range from 44-85 eggs deposited in double strands under rocks in cool-flowing streams (Karracker et al. 2006) from late June to early July following the fall breeding. Post-metamorphic A. truei are sensitive to ambient air temperature > 24°C and, relative to many other anurans, are less tolerant o f desiccation, which may limit movement across dry or warm terrestrial environments (Claussen 1973a; Brown 1975; Daugherty and Sheldon 1982). Additionally, multiple studies have demonstrated that tailed frogs are extremely philopatric to reproductive streams (Daugherty and Sheldon 1982; Wahbe et al. 2004; Matsuda and Richardson 2005; Burkholder and Diller 2007). However, juveniles have been identified traveling up to 100 m from larval streams, leading researchers to hypothesize that this age class is responsible for emigration (Daugherty and Sheldon 1982; Wahbe et al. 2004). Conservation Issues Ascaphus have specialized life-stage requirements resulting in an inherent vulnerability to anthropogenic and natural disturbances that may influence the quality or availability o f habitat. As a result, the tailed frog has been designated as a Blue Listed 3 species by the BC Conservation Data Centre (CDC) and has been listed as Special Concern under the federal Species at Risk Act (SARA). Within BC, anthropogenic disturbances that remove forest overstory and reduce the complexity o f downed and standing wood in the understory have the potential to reduce the distribution or productivity o f populations o f the tailed frog (Dupuis and Waterhouse 2001). Some researchers in North America suggest that the decline in amphibian abundance is associated with clearcut or shelterwood logging (Olson et al. 2007; Kluber et al. 2008; Hawkes and Gregory 2012). Factors associated with forest harvesting and the decline of tailed frogs include the loss of complex structural components in maturing or old forests, such as coarse woody debris; reduction o f overstory cover; and changes in hydrological processes (Dupuis and Waterhouse 2001). For example, Hawkes and Gregory (2012) in Washington State, United States o f America (USA), demonstrated that the relative abundance o f tailed frogs was generally higher in habitats upland o f streams prior to logging. Additionally, Wind (2009) suggested southeastern populations o f the tailed frog have been impacted by past dam and water construction around Lillooet BC, which may have interrupted migration pathways. Conservation Management Under the Forest and Range Practices Act (FRPA; formerly Forest Practices Code), the Identified Wildlife Management Strategy (IWMS) recognises two categories o f wildlife that require special management or habitat protection: 1) Species at Risk; and 2) Regionally Important Wildlife. The coastal tailed frog is designated as a Species at Risk because o f habitat requirements that are not protected under the FRPA and the documented effects of timber harvest on tailed frog populations. The tailed frog is associated with structural stages 4 of mature (100-140 years) or old (> 140 years) forests and can be found along S4-S6 streams (small fish or non-fish bearing streams). Currently, these streams have fewer restrictions on anthropogenic activities, such as forestry. As an identified wildlife species (IWMS) some habitat conservation exits in areas designated as ecological reserves, Wildlife Habitat Areas (WHAs), old growth management areas (OGMAs), or through the creation o f special resource management zones for other species such as the grizzly bear. Additionally, certain management guidelines have been created for reserves designed for A. truei. For instance, around areas that contain tailed frogs, OGMAs or WHAs can be established to protect stands in serial stages 6 and 7. Boundaries o f these reserves should contain a 30-m core area and have a 20-m management zone on both sides o f the stream to reduce risk o f windthrow and to maintain microclimate conditions and important structural elements (i.e., coarse woody debris) similar to an unaltered stand. Previous Post-metamorphic Research Past research on Ascaphus has focused on interior rocky mountain and southern populations (Daugherty and Sheldon 1982; Wahbe et al. 2000, 2004; Matsuda and Richardson 2005; Hayes et al. 2006; Burkholder and Diller 2007; Hawkes and Gregory 2012) with less work at their northern extent (Dupuis and Steventon 1999). Given the observed declines of southern populations following anthropogenic disturbances (Hawkins et al. 1988; Bull and Carter 1996; Hawkes and Gregory 2012), there is concern that tailed frogs at the northern extent o f their distribution may also be negatively affected by activities such as forest harvesting, stream diversion, or the creation o f linear corridors. However, the 5 possible impacts are speculative as our knowledge o f habitat requirements for the terrestrial life-stage of the species is limited. Researchers investigating the movement and distribution of post-metamorphic A. truei have relied on data collected using pitfall traps (Wahbe et al. 2004; Matsuda and Richardson 2005; Hawkes and Gregory 2012) or area-constrained searches (Hayes et al. 2006; Burkholder and Diller 2007). However, recaptures are low for A. truei providing limited understanding o f the spatial distribution o f the population or the movement o f individual tailed frogs. A better understanding o f adult tailed frogs’ ecology and biology is necessary to predict, monitor, and prevent or mitigate potential impacts resulting from habitat alteration and disturbance (Deguise and Richardson 2009). Radio telemetry provides information describing animal movement that is not available by other methods (Bartelt and Peterson 2000). Historically, the size and weight o f the transmitter has prevented the use o f radio telemetry for monitoring movements and distribution o f small amphibian species (Forsythe et al. 2004). However, advances in technology, such as the miniaturization o f batteries, have resulted in an opportunity to apply this technique to small cryptic species like the tailed frog. Radio-telemetry locations can be coupled with powerful statistical models to quantify movement dynamics, distribution, and habitat use (Rowley and Alford 2007). Resource selection functions (RSF) are one type o f statistical model that use animal locations to quantify a species’ distribution and habitat requirements (Johnson et al. 2004). A RSF compares the type o f used resources or habitat at an animal location with those resources that are available (Manly et al. 2002). Quantifying and mapping species-habitat relationships can 6 guide reserve design or inform land management activities (Boyce et al. 2002; Araujo et al. 2004; Cabeza et al. 2004; Koper and Manseau 2010). Research Purpose The goal o f my thesis is to quantify the seasonal movement and habitat selection patterns o f a northerly population o f the tailed frog across a range o f forest conditions modified by timber harvesting. These findings will broaden our understanding o f the ecology and life history o f the species within landscapes currently subjected to forestry and other land-use activities. The outcome o f this study will guide the protection, management and restoration of headwater streams associated with tailed frog populations. Within the scope o f my thesis, I addressed 3 research objectives: (1)1 used systematic pitfall data to summarise capture rates by demographic, temporal, environmental and treatment variables suspected to influence abundance and distribution. (2) I used locations from tailed frogs monitored with radio telemetry to develop RSFs that quantified the distribution and habitat selection o f tailed frogs. (3) I employed radio telemetry and systematic pitfall traps to reveal temporal and environmental patterns in the movement o f tailed frogs. My thesis is structured as 4 chapters. In Chapter 1 ,1 provide a general introduction to the study species, research objectives, and the study area. Researchers have observed that southerly populations o f A. truei make seasonal large-scale movements that are dependent on life-history requirements. Furthermore, those movements and ultimately population distribution may be influenced by environmental conditions that act as a physiological or ecological constraint (Matsuda and Richardson 2005; Hayes et al. 2006). Thus, in Chapter 2, 7 I used pitfall capture data and an Information Theoretic Model Comparison (ITMC) approach to test hypotheses predicting the distribution of populations o f tailed frogs within a site relative to temporal, environmental, and treatment variables. In Chapter 3 , 1 used radio-telemetry data to identify the temporal and environmental factors that influence the movement, and resource selection o f individual tailed frogs. Pitfalltrap data have limited utility for addressing such questions as movement and distribution are premised on recaptures, which are low for this species (Maxcy 2000; Wahbe et al. 2004; Matsuda and Richardson 2005; Hayes et al. 2006). This is the first study to use radio telemetry to quantify movement and resource selection o f Ascaphus spp. Lastly, in Chapter 4 I summarize my findings, including a discussion o f the conservation relevance and management applications o f the research. Study Area My research occurred in 3 watersheds in the Hazelton Mountains centered on the Skeena River. These mountains are located on the leeward side o f the Coast Mountains and are typically round-topped or domed ridges with small remnant glaciers that were created during the last ice age. Vegetation communities o f the study area are representative o f the transition between the Coastal Western Hemlock (CWH) and the Interior Cedar-Hemlock (ICH) biogeoclimatic zones. The CWH zone is found along the coast from sea level to 1050 m on the leeward side o f the mountains and has relatively mild temperatures and heavy rainfall (Figure 1). Mean annual temperatures can range from 5.2-10.5°C. 8 Ootribution of the Coastal Western Hemlock Biogeodrmatic Zone Cnputd ku b i t * Columbia Mmsiy of F tn m by C a a b a C a r o p a k ia 1 0 . b b rd i I SOB Figure 1: Distribution o f the Coastal Western Hemlock Biogeoclimatic Zone in British Columbia. Figure from Meidinger and Pojar (1991). Grey star indicates location o f research site. 9 The mean annual precipitation for this region is 2228 mm, approximately 40-50% o f which is snowfall at the northern extent o f the zone (Pojar et al. 1991). Within the CWH, my study sites occurred in the wet submaritime subzone (CWHws). This subzone has some o f the largest conifers in BC and has a dominant overstory consisting o f western hemlock (Tsuga heterophylla), western red cedar (Thuja plicata) and amabilis fir (Abies amabilis). The shrub layer contains Alaskan blueberry ( Vaccinium alaskaense), oval­ leaved blueberry (V. ovalifolium) and false azalea (Menziesiaferruginea). Common species in the herbaceous layer include queen’s cup (Clintonia uniflora), one-sided wintergreen (Orthilia secunda), bunchberry (Cornus canadensis), twinflower (Linnaea borealis) and rattlesnake plantain (Goodyera oblongifolia). Typically, there is a well-developed moss layer that contains such species as red-stemmed feathermoss (Pleurozium schreberi), lanky moss (Rhytidiadelphus loreus), shaggy moss (R. triquetrus) and pipecleaner moss (Rhytidiopsis robusta; Pojar et al. 1991). The ICH zone is located at lower to mid-elevations (100-1000 m), in a small pocket located east o f the Coast Mountains in adjacent parts o f the Hazelton Mountains (Figure 2). The continental climate produces cool wet winters and warm dry summers. The average range of temperatures is 2-8.7°C and annual mean precipitation is 500-1200 mm with approximately 25-50% being snow (Meidinger and Pojar 1991). The most northern subzone is the ICHmc and refers to the ’moist and cold’ climate found from Hazelton to Meziadin Lake. The climax over-story is dominated by conifers including western hemlock, western redcedar, Roche spruce (Picea x lutzii) and subalpine fir (Abies lasiocarpa; Meidinger and Pojar 1991). 10 Distribution of the Interior Cedar-Hemlock Biogeoclimatic Zone too o 200 o parto lor Brush Colirnbrs Mrntsiif o( F om a by Conadron Comgnphics lib. Morcti 1089 Figure 2: Distribution o f the Interior Cedar-Hemlock Biogeoclimatic Zone in British Columbia. Figure from Meidinger and Pojar (1991). Grey star indicates location o f research area. 11 The understory shrub layer consists o f black huckleberry ( Vaccinium membranaceum), devil’s club (Oplopanax horridus) and some skunk cabbage (Lysichiton americanus). Characteristic herbaceous species are bunchberry, twinflower and queen’s cup. The associated non-vascular species include red-stemmed associated non-vascular species include red-stemmed feathermoss, Knight’s plume (Ptilium crista-castrensis) and step moss (Hylocomium splendens; Meidinger and Pojar 1991). Study Sites The study sites were located within 3 watersheds: 1) Ascaphus (AS); 2) Gosling (GO); and 3) Kleanza (KL; Figure 3). Sites within each o f these watersheds were in the CWH zone and had portions that are old growth (i.e., age class 8 or 9; 141-250 years or > 250 years). Each o f the sites was located adjacent to 3rd order (Strahler 1957) larval streams that are positioned in the middle or upper portions o f 3rd order sub-basins (M. Todd Pers. Comm.). I conducted my study across 3 sites representing forest retention treatments: 1) oldgrowth comprised o f undisturbed stands >140 years; 2) stream-side forest retention (i.e., buffer) where harvesting has occurred at a distance o f 30-50 m from the primary stream; and 3) clear-cut harvest o f the site including the area adjacent to the stream edge. As harvesting began in 1987, one site had significant tree growth and was identified as representing forest regeneration (Table 1). All sites were located >50 meters from operational forest service roads. History o f harvest within the watersheds differed based on the time o f road construction. For example, the Kleanza watershed was cut > 30 years and contained regenerating stands that were 25-30 years old. The Ascaphus watershed had 10-15 year old 12 forest retention treatments, while the Gosling watershed was the most recent with harvesting occurring < 10 years in the past. Thus, the harvested stands ranged from age class 1 or 2 (040 years) and represented structural stages o f 1 (immediately post disturbance) to 4 (pole saplings; BCMFR and BCMOE 2010). 13 Figure 3: Location o f the study watersheds, Gosling Creek (GO), Ascaphus Creek (AS) and Kleanza Creek, (KL) and treatments (1- old growth, 2- forest retention buffer and 3- clearcut) near Terrace, BC. Table 1: History of forest harvesting that has occurred within the watersheds for all forest retention treatments (old growth (OG), forest retention buffer (BF), regeneration (RG) and clearcut (CC)) east of Terrace, BC, where coastal tailed frog populations were monitored. Watershed Forest Elevation Biogeoclimatic Treatment Status Dominant Dominant Harvest Retention (m) Ecosystem Canopy Condition Understory Date Treatment Classification Gosling OG 463 CWHwsl >250 year old Moist pockets of N/A western hemlock devil’s club and Streamside OG. Opposite bank OG. and subalpine fir skunk cabbage with Upstream OG. forest with multiVaccinium species storied canopy. throughout._______ Gosling Streamside WHA BF 417 CWHwsl >250 year old Dominated by 2006 Vaccinium species, western hemlock core area reserve 3050 m OG (average and subalpine fir bracken fern (Ptreidium 40 m). Opposite forest with a bankOG. Upstream recently disturbed aquilinum), fireweed area outside of (Chamerion OG. angustifolium), and stream buffer. logging debris with a few moist pockets of devil’s club and skunk cabbage. Within the buffer only Vaccinium species.___________ Streamside variable Gosling Dominated by 1998. CC 319 CWHws2 Western hemlock 2010 width retention and subalpine fir Vaccinium species (average 30 m; found in streamside with very few moist range 0-50 m) with retention with dbh pockets of devil’s dbh < 30 cm. <30 cm. Recently club. Outside of Opposite bank CC buffer, dominate in disturbed site with with similar pattern no canopy outside fireweed and logging of stream buffer. of streamside debris with very few retention moist pockets of Upstream CC with devil’s club and similar pattern of Vaccinium species. Ascaphus OG 550 CWHwsl Ascaphus BF 582 CWHwsl Ascaphus CC 273 CWHwsl Kleanza OG 801 CWHws2 141-250 year old western hemlock and subalpine fir forest, with western red cedar and Sitka spruce (Picea sitchensis) component. Multistoried canopy. 141-250 year old western hemlock and subalpine fir forest, with western red cedar and Sitka spruce within streamside buffer. No dominant canopy outside of buffer. 141-250 year old western hemlock and subalpine fir within streamside retention. Dense pole/sapling western hemlock and lodgepole pine (Pinus cortata) < 10 m tall outside of buffer. > 250 year old western hemlock, amabilis and subalpine fir with streamside retention. Streamside OG. Opposite bank OG. Upstream OG immediately adjacent, with CC harvesting in headwaters. Moist pockets of devil’s club, skunk cabbage with Vaccinium species and false azalea) throughout. N/A Vaccinium species and a few moist pockets of devil’s club within buffer. Dense Vaccinium spp and false azalea outside of buffer. 19941995 WHA core area reserve 40-50 m OG buffer. Opposite bank OG. Upstream OG immediately adjacent with CC harvesting in headwaters. Vaccinium species within streamside buffer with moist pockets of devil’s club and red osier dogwood (Cornus stolonifera) by the stream. 19921995 Few moist pockets of devil’s club and skunk cabbage with Vaccinium species N/A Streamside variable width retention (average < 25 m) with substantial blowdown. Opposite bank OG. Upstream OG immediately adjacent with CC harvesting in headwaters. Streamside OG. Opposite bank CC in 2009. Upstream OG. Kleanza BF 650 CWHws2 Kleanza RG 528 CWHwsl multi-storied canopy. > 250 year old western hemlock, amabilis and subalpine fir. Multi-storied canopy within streamside buffer. Dense pole/sapling western hemlock and subalpine fir < 15 m tall outside of buffer. Dense pole/sapling western hemlock and subalpine fir < 20 m tall. Vaccinium species with some false azalea and moist pockets of abundant devil’s club and skunk cabbage. 19771979 WHA core area reserve -50 m OG buffer. Opposite bank CC (1991) and salvage logging (1997). Upstream OG. Poorly developed understorey with moist pockets of devil’s club and lady fern. 19861987 Streamside no retention. Opposite bank CC in 1980. Upstream OG. C hapter 2 F in e - S c a l e D is t r ib u t io n a n d M o v e m e n t P a t t e r n s o f P o s t - M e t a m o r p h ic C o a s t a l T a il e d F r o g s In t r o d u c t i o n The distribution and movement patterns o f individual animals can reveal mechanisms that dictate the spatial ecology, population dynamics such as breeding phenology, and evolution of a species (Daugherty and Sheldon 1982). For example, many amphibians have biphasic life cycles requiring long-distance movements to and from breeding or over­ wintering sites. During the summer, short excursions occur within relatively small patches of habitat related to specific climactic conditions and food (Semlitsch 2008). To understand how such movements relate to broader population processes, one must conduct studies that reveal the interactions between the environment and the species’ spatial ecology. Such inferences have both theoretical and applied applications, especially for amphibian species influenced by human development and habitat alteration (Semlitsch 2008). Ascaphus truei, coastal tailed frog, is endemic to the Pacific Northwest. They are habitat specialists with life-history requirements closely associated with forests o f this region. The tailed frog can live between 15-20 years and depending on the geographical location, take 1-5 years to metamorphose (Daugherty and Sheldon 1982). Water temperatures > 18.5°C can affect embryonic development o f larvae and adults are susceptible to ambient air temperatures > 24°C (Claussen 1973b; Brown 1975). This species is also less tolerant to desiccation than other amphibians and requires forest stands with a complex understory (Claussen 1973a, 1973b; Daughtery and Sheldon 1982). The response o f tailed frog larvae to forest practices has been relatively well documented because o f the ease o f working with this life stage (Com and Bury 1991; Wallace and Diller 1998; Wahbe and Bunnell 2001). Conversely, tailed frogs inhabiting terrestrial habitats are difficult to locate because o f cryptic colouring, relatively small size, 19 movement through or occupancy o f downed and rotten wood and nocturnal activity patterns (Metter 1967; Green and Campbell 1984). The removal o f overstory and reduction in structural complexity in the understory can restrict the distribution o f post-metamorphic tailed frogs with implications for survival and reproduction (Hawkes and Gregory 2012). When occupying unmodified habitats, adult tailed frogs are extremely philopatric to the larval stream (Wahbe et al. 2004; Matsuda and Richardson 2005; Burkholder and Diller 2007). Reproductive females o f southerly populations have been observed to alter their distribution depending on the time o f year such as moving upstream after ovipositioning or downstream to breed (Landreth and Ferguson 1967; Brown 1975; Wahbe et al. 2004; Hayes et al. 2006). Burkholder and Diller (2007) also concluded that females moved greater longitudinal distances along the stream edge in comparison to adult males and juveniles. Previous research on southerly populations o f the tailed frog has demonstrated variation in the spatial distribution o f age classes (adult versus juveniles). Wahbe et al. (2000) found that reproductive adults were more common along the stream edge and suggested that this was reflective o f philopatry. Additionally, Wahbe et al. (2000) captured the majority o f newly metamorphosed tailed frogs 100 m from the stream edge suggesting dispersal by younger animals. Alternatively, Burkholder and Diller (2007) noted that both immature and adult tailed frogs exhibited site fidelity. Conversely, Matsuda and Richardson (2005) demonstrated no significant age- or habitat-specific differences in the direction of movement, but found a greater proportion o f individuals moved parallel with the stream. Within BC, the alteration and loss o f terrestrial habitat for tailed frogs can be attributed to clear-cut logging and associated road construction (Deguise and Richardson 2009). Removal o f overstory changes the light, temperature, humidity, and soil moisture 20 within a stand. For example, once the overstory canopy is removed an increase in the amount o f light penetrating the understory increases the temperature and decreases the soil moisture (Brofoske et al. 1995). Because amphibians may undergo long-distance movements for dispersal and demonstrate fidelity to breeding sites, a change in the environment can create barriers that affect population connectivity at multiple spatial scales (deMaynadier and Hunter 1995; Gibbs 1998; Knapp et al. 2003; Wahbe et al. 2004; Hawkes and Gregory 2012). Understanding the processes that influence site fidelity and associated movements not only provides insights into the ecology of tailed frogs, but also increases our understanding o f how habitat change influences the dynamics o f individual populations (Daugherty and Sheldon 1982). C h a p t e r P r e d ic t io n s Drawing on the existing literature, I tested 3 predictions that described and explained the fine-scale distribution and movement o f the tailed frog at their northern extent in western BC. I used systematic pitfall data to summarise capture rates by demographic, temporal, environmental and treatment variables suspected to influence relative abundance and finescale distribution. I monitored the movement o f frogs across 4 forest retention treatments: 1) recent clearcut with variable stream retention; 2) clearcut with regeneration (< 25 year post harvest); 3) forest retention at the larval stream (Wildlife Habitat Area 50-m core reserve); and 4) forest retention across the study site. Demographics and movement - I predicted that the gender and age o f a tailed frog would influence movement and distribution. Juveniles would be located at greater distances from the stream edge in all 4 treatments; this pattern of distribution reflects dispersal behavior observed in other studies. Furthermore, I predicted that the movement patterns o f 21 reproductive-aged tailed frogs would correspond with the ovipositioning and breeding periods. Forest retention treatment and movement - I predicted that tailed frogs in old-growth habitats would be captured at greater distances from the stream edge relative to habitats with forest harvesting. This pattern o f movement was in response to environmental constraints resulting from reduced forest complexity and microclimatic conditions that exceeded the physical tolerance for dry and warm environments. Weather and movement —I predicted that tailed frogs would demonstrate more active and dynamic movements and be located further from the stream edge when weather conditions (i.e., rainfall events, temperature and relative humidity) were more suitable for above-ground terrestrial movements. M ethods Study Area The study area was situated on the leeward side o f the Coastal Mountains in northwestern British Columbia. Vegetation communities included the Coastal Western Hemlock (CWH) and some transitional species associated with the Interior Cedar-Hemlock (ICH) Biogeoclimatic Ecosystem Classification (BEC) zones. The CWH has mild temperatures and heavy rainfall with climax forests o f western hemlock {Tsuga heterophylla) and amabilis fir (Abies amabilis). The ICH, reflective o f a transitional climate between coastal and interior stands, contained climax forests o f western hemlock and subalpine fir (Abies lasiocarpa). The main disturbance across these study sites was timber harvest. See Chapter 1 for a more detailed description o f the study area. 22 Pitfall data were collected during 2012 and 2013 from sites in the Gosling (GO) and Kleanza (KL) watersheds. Sites in each watershed were classified as 1 o f 4 categories representing forest harvest history within the watershed: 1) no history o f harvest; 2) a forest retention buffer (30-50 m) 3) clearcut with minimal (dbh < 30 cm) to no overstory canopy; and 4) regeneration (< 25-year overstory). The GO sites were located within the transition zones between the CWHws and the ICHmc2 (Meidinger and Pojar 1991). Study sites were classified as CWHwsl or CWHws2 and situated at relatively low elevations (319-463 m above sea level (asl)). These sites were dominated by clearcut or old (> 250 years) western hemlock and amabilis fir with moist pockets o f devil’s club (Oplopanax horridus), skunk cabbage (Lysichiton americanus) and abundant Vaccinium species. Most harvesting in this watershed occurred recently (< 5 years) resulting in no tree canopy and substantial growth o f fireweed (Chamerion angustifolium) and bracken fern (Pteridium aquilinum) where logging had occurred. The sites within the KL watershed were also found within two variants o f the Coastal Western Hemlock BEC zone (CWHws2, CW Hwsl). Two sites were located at higher elevations (650 m and 801 m asl; CWHws2). The old growth treatment had > 250 year old western hemlock, amabilis and subalpine fir in a multi-storied canopy; while the forest retention buffer treatment contained 50 m o f > 250 year old western hemlock, amabilis and subalpine fir along the stream. Both contained moist microsites in the understory with abundant devil’s club and skunk cabbage. The third site was lower in elevation (528 m asl; CW H w sl) and had dense pole/sampling regeneration o f western hemlock and lodgepole pine (Pinus contorta) < 20 m tall. This site was harvested in 1987 and had poorly developed understory with moist pockets o f devil’s club and lady fern (Athyrium filix-femina). 23 Data Collection I installed 288 pitfall traps within the 2 watersheds (n = 144 per watershed). Each trap was 38 cm in depth and 15 cm in diameter. A plastic insert was put in place to limit escape o f tailed frogs by jumping. To limit rodent mortality, a length of 40-cm twine was secured to the insert (Karracker 2001); however, after the 2012 trap session, the twine was knotted to better allow escape by rodents (Appendix I). Within each watershed we installed 36 pitfall arrays each containing 4 traps and 2 perpendicular arms o f 10-m lengths o f drift fencing (Matsuda 2001). Arrays were placed systematically at 4 distances (5, 30, 55 and 80 m) from the known larval stream for a total o f 3 arrays (12 traps) at each distance (Figure 4A). Each captured tailed frog was assigned a direction o f movement according to the drift fence orientation relative to the adjacent stream. Tailed frogs were recorded as moving downstream (with the flow o f water), upstream (against the flow o f water), away from the stream, and toward the stream (Figure 4B). Trap session duration - To control for temporal influences on tailed frog behavior, all traps within a watershed were open simultaneously and were visited daily during the months o f June-October for a 6-night period. During 2012, 3 trap sessions were conducted from mid-July until the end o f October resulting in 5,184 trap nights. For 2013, 3 trap sessions were run across 5 months (June-October) resulting in 6,192 trap nights (Table 2). Tailedfrog morphometries- For each captured tailed frog 1 recorded the snout vent length (SVL), shank length (SL), weight and gender (male or female) based on visible sexual characteristics (external cloaca, breeding pads or eggs present). I classified each animal as adult (reproductive); juvenile (pre-reproductive) or metamorph (containing rudimentary tail). The reproductive status o f adults was determined by the visibility o f eggs in females (gravid) 24 or darkened breeding pads in males (breeding). I used visual implant elastomer (VIE; Northwest Marine Technologies, Inc.) to permanently mark each individual with a unique colour combination. If a tailed frog was < 2.0 g a mark indicating only year and watershed was assigned. Statistical Methods Data analysis Presence o f tailed frogs within a trap allowed me to quantify the activity o f each demographic and age class relative to forest retention treatment, distance from stream, season, weather and microclimate data. I applied statistical count models relating the number o f tailed frogs captured per day in each trap to temporal, site, and environmental variables. Count models accommodate both over dispersion and presence o f excess zeros (i.e., traps with no counts), a common occurrence in ecological data (Nielsen et al. 2005). Following extensive testing, I determined that the distribution o f count data was too infrequent for a Poisson or negative binomial distribution. Thus, I compressed the count data to a binary variable (0 = no catch, 1= > 1 catch) and used logistic regression to calculate the likelihood o f capturing a tailed frog in a trap. Combinations o f variables within each logistic regression model served as hypotheses representing ecologically plausible explanations for the capture o f at least one tailed frog (Appendix II and III; Table 3). The number o f tailed frogs captured was not sufficient (n = 129 in 2012 and n = 31 in 2013) to split model sets into adult (reproductive) and juvenile (pre-reproductive) age classes. However, to control for a difference in years, I tested identical model sets for both 2012 and 2013. 25 U pstrea m 30 m 5m I | ■50 m U § K w CA H i i 55 m 80 m 50 m X X X X X X X X X y X = pitfall array D ow n strea m A U pst r ea m A w ay (3 T o w a r d s (1 ) D o w n str ea m Figure 4: Trap arrays for tailed frogs applied to 3 forest retention treatments found within two watersheds (A). Each array (X) contained 4 traps centered in 2 perpendicular arms o f 1Om o f drift fencing for a total o f 288 traps. Each trap within the array was used to monitor direction o f tailed frog movement in relation to the known larval stream, towards (1), down (2), away (3) and up (4) stream (B). 26 Table 2: Trap session dates and observed seasonal life cycle o f the tailed frog for 2 watersheds east o f Terrace, BC, surveyed during 2012 and 2 0 1 3 ._______________ Session Life cycle Y ear W atershed Dates Spring Gosling ovipositioning July 17-22 2012 Spring Kleanza July 22-27 ovipositioning Summer foraging Gosling August 25-30 Summer foraging Kleanza August 19-24 breeding Gosling October 9-14 Fall breeding Fall Kleanza October 1-6 2013 Kleanza Gosling Kleanza Gosling Kleanza Gosling June 2-8 July 3-8 August 7-12 August 16-21 September 10-16 September 27-October 2 Spring Spring Summer Summer Fall Fall Gosling October 3-8 Fall 27 ovipositioning ovipositioning foraging foraging movement to breeding sites movement to breeding sites movement to breeding sites/breeding 1 used variance inflation factors (VIF) to assess multicollinearity among covariates. An individual VIF > 10 or a mean VIF > 1 suggested that a model had high levels of multicollinearity (Chatterjee et al. 2000). Temporal variable - Trapping sessions were assigned a categorical variable to distinguish the month. During 2012, sessions occurred on: 1) July 17th-27th; 2) August 19th30th; and 3) October 1st-14th. In 2013, the number o f sessions and temporal dispersion of effort increased: 1) June 2nd-8th; 2) July 3) August 7th-12th and 16^-21st; 4) September 11th- 16th and September 27th- October 2nd; and 5) October 3rd-8th (Table 2). Site variables - For each pitfall array, I measured habitat variables at a 25-m2 plot (2.82-m radius). I measured the percent coverage o f shrubs and trees < 2 m (Bi) and those > 2 m, but < 10 m (B 2 ; Ministry o f Forests 1998). These layers were used in combination with indicator plants to quantify the site moisture at each trap array. Sites in each watershed were classified as 1 o f 4 canopy covers based on the forest harvest history within the watershed: 1) no history o f harvest (n = 96 traps); 2) a forest retention buffer (30-50 m; n = 96) containing trees > 250 years old and a distinct edge from a previous cut; 3) clearcut with minimal (dbh < 30 cm) to no overstory canopy (n = 48); and 4) regeneration (< 25-year overstory; n = 48). This variable represented the overstory and structural complexity at each site. Trap numbers varied among disturbance categories. Each trap array was classified into 1 of 3 categories according to the distance o f the array from the clearcut (hard) edge: 1) located between the stream edge and > 10 m from the clearcut edge; 2) < 10 m on either side o f the hard edge; and 3) > 10 meters from the hard edge within a clearcut. Additionally, I used a spherical densiometer to quantify the percent canopy closure. 28 I used the Ecosystems o f BC (Meidinger and Pojar 1991) for the Prince Rupert Forest Region and indicator plant species within the 25-m2 plot, to quantify site moisture as: 1) ‘dry’ if no species associated with wet ecosystems were present; 2) ‘mesic’ when devil’s club was < 10% with minimal oak fern (Gymnocarpium dryopteris)\ 3) ‘subhygric’ when < 20% devil’s club was present and dominant in oak fern; and 4) ‘hygric’ when the site had > 20% devil’s club and was dominated by lady fern or trees with affinity for wet soil (e.g., Salix spp.). Within the 2 watersheds, site elevations ranged between 325-801 m asl. I used a Gaussian term to account for the nonlinear influence o f site elevation on the distribution o f tailed frogs. This equation consisted o f both the linear component, which identified the increase in tailed frog captures, and a squared component to identify at what elevation tailed frog presence decreased as a function of increasing elevation. Environmental variables - For each trap day I used microclimate stations and field observations to measure weather: air temperature, relative humidity, and days since last rain. Microclimate stations were deployed at 5, 30, 50, 100 m from the stream edge and consisted o f 1 DS1923-F5 Hygrochron (temperature and relative humidity) iButton data logger (Maxim Integrated) suspended 1 m above the ground under a radiation shield (M. Todd Unpub. Data). From those data I calculated the maximum and minimum temperature and humidity in a 24-hour period (10 a.m .-lO a.m.) as well as the difference between the maximum and minimum. Similar to elevation, the temperature variable was fit as a nonlinear Gaussian function. 29 Table 3: Independent variables used to model the occurrence and abundance o f tailed frogs captured in pitfall traps in 2 watersheds east of Terrace, BC, during 2012 and 2013.__________________________________________________________________________ Variable Variable Description Theme year Temporal Year o f trap sessions: 1) 2012 and 2) 2013 Trap month. 2012: 1) July; 2) August; and 3) October. 2013: 1) June; 2) July; 3) August; 4) trap month September; and 5) October. Overstory tree coverage: 1) old growth; 2) harvested buffer; 3) clearcut; 4) regeneration. Treatment canopy cover densiometer Percent canopy opening at the pitfall array. site moisture Inferred site wetness based on BEC indicator plants in the 25-m2 plot: 1) zonal series (dry); 2) mesic; 3) subhygric; and 4) hygric (page 29). distance to edge Distance of trap arrays from hard edge: 1) stream edge to < 10 m from hard edge; 2) arrays ± 10 m of the hard edge; and 3) arrays > 10 m from a hard edge. distance to stream Straight-line distance from the larval stream edge. orientation Trap arrangement within array indicating direction of tailed frog movement in relation to larval stream: 1) towards; 2) down; 3) away; and 4) up. site elevation* Elevation in meters defining the start point of a site in the forest retention treatment. Maximum temperature over a 24-hour period using two-microclimate transects. Environment tempMAX* Minimum temperature over a 24-hour period using two-microclimate transects. tempMiN* Difference between the maximum and minimum temperatures over a 24-hour period. tempvaribility Maximum humidity over a 24-hour period using two-microclimate transects. humidityMAx Minimum humidity over a 24-hour period using two-microclimate transects. humidityMiN Difference between the maximum and minimum humidity over a 24-hour period. hum idity varibility Number of days since the last precipitation event._____________________________________ days since last rain *Variable had Gaussian term applied to represent a non-linear relationship. Model selection I used Akaike’s Information Criterion for small samples (AICc) to rank and select the most parsimonious logistic regression model. I used AAICc to identify the ‘best’ model o f the set and Akaike weights (AICvv) to quantify model selection uncertainty. If models were nearly equivalent (i.e., AAICc < 2), then I selected the model with the fewest number o f parameters (Johnson et al. 2006). I used AICc to test whether repeated sampling within a treatment, trap array or trap warranted a mixed effects model (i.e., random effect for multiple sampling within fixed treatment). I determined that a full mixed effect model was not warranted. However, I used the cluster option within Stata (ver.12.1, StataCorp LP, 2012) to correct the variance for repeated sampling at a trap (Williams 2000; Nielson et al. 2001). Model and variable evaluation I calculated the Area Under the Curve (AUC) for the Receiver Operating Characteristic (ROC) to test the predictive performance o f the most parsimonious model. The curve evaluates the proportion o f correctly and incorrectly classified predictions over a continuous range o f probability thresholds from 0 to 1.0 (Pearce and Ferrier 2000). Scores of 0.5-0.7 suggest a poor model, 0.7-0.9 a good model, and 0.9-1.0 a highly predictive model (Swets 1988). When performing the ROC test I withheld each trap record sequentially from the model building process and then used that withheld record to generate a predicted probability. Thus, a bootstrapping-like process resulted in an independent evaluation o f the predictive performance o f the logistic regression model. I used 95% confidence intervals to assess the relative strength o f the coefficients within the most parsimonious model. When intervals do not overlap zero, covariates are 31 considered as a significant factor influencing the distribution of tailed frogs. All statistical analysis was conducted using Stata (ver.12.1, StataCorp LP, 2012). R esu lts Capture statistics In 2012, 5,184 trap nights resulted in the capture o f 129 tailed frogs (63 female, 29 male, 37 juveniles or metamorphs) for a catch per unit effort (CPUE) rate o f 2.49 tailed frogs/100 trap nights. An increased trap effort in 2013 resulted in 6,192 trap nights and the capture o f 31 tailed frogs (14 female, 14 male, 3 juveniles) with 0.50 tailed frogs/100 trap nights. Over the 2 years, adults in the old growth had the greatest CPUE with 1.30 individuals/100 trap nights (Table 4). Juveniles had the greatest CPUE in the regeneration treatment (0.45 individuals/100 trap nights) and metamorphs in the clearcut (0.55 individuals/100 trap nights). In both years, 95% confidence intervals revealed that females had a significantly greater average weight (wt), snout-vent length (SVL) and shank length (SL) when compared to males (Figure 5). Variation in microclimate During 2012, the GO watershed had more rain events closer together in the summer and fall compared to the spring. In the KL watershed more rain events occurred during the summer trapping session compared to the fall or summer (Appendix IV). In 2013, the fall trapping sessions had far more rain events in the GO compared the KL watershed (Appendix V). 32 Table 4: Total number of tailed frogs captured and the catch per unit effort (CPUE) for each age class and 4 forest retention treatments (old growth, forest retention buffer, regeneration, and clearcut) for 2 watersheds located east of Terrace, BC, during 2012 and 2013. CPUE represents the number o f individuals/100 trap nights._______________________________________________________________ M etam orphs CPUE Adults CPUE Juveniles CPUE Forest retention treatm ent N um ber of trap nights Both vears Old growth Forest retention buffer Regeneration Clearcut 3792 3792 1776 2016 48 38 18 7 1.30 1.00 1.00 0.35 7 11 8 4 0.18 0.32 0.45 0.20 4 4 0 11 0.11 0.11 0 0.55 2012 Old growth Forest retention buffer Regeneration Clearcut 1728 1728 864 864 43 24 11 6 2.49 1.39 1.27 0.69 5 12 8 2 2.89 0.69 0.93 0.23 4 4 0 11 0.23 0.23 0 1.27 2013 Old growth Forest retention buffer Regeneration Clearcut 2064 2064 912 1152 5 14 7 1 0.24 0.68 0.77 0 2 0 0 2 0 0 0 0.17 0 0 0 0 0 0 0 0 - 12.00 10.00 t s o - 8.00 X DO - 6.00 X * Vi - 4.00 | b Snout-vent length (mm) □ Shank length (mm) ■ Weight (g) C 0 >1 3 O c 00 - 2.00 - 0.00 female male female 2012 male 2013 Gender and year Figure 5: Weight (grams), snout-vent-length (mm) and shank lengths (mm) including 95% confidence intervals o f adult tailed frogs in 2 watersheds located east o f Terrace, BC, during 2012 and 2013. 34 Mircoclimate arrays in each watershed during 2012 demonstrated that temperatures for the 5-m arrays were comparable to the respective 80-m arrays for the old growth and clearcut treatments. However, the 5-m temperature arrays in the forest retention buffer treatments were similar to the old growth, while the 80-m arrays were more reflective of arrays at the same distance located in the clearcut or regeneration treatments (Figure 6). Additionally, microclimate arrays in 2013 demonstrated greater variability in weather patterns compared to those observed in 2012 (Figure 7). Orientation o f movement The greatest proportion o f tailed frogs captured in this study were found moving perpendicular (away or towards; 69.4%) to the stream rather than parallel (upstream or downstream; 30.6%). The orientation o f movement was largely influenced by adults for 3 o f the treatments (old growth, forest retention buffer, and regeneration). For example, in the old growth, tailed frogs had the greatest proportion o f captures moving away from the stream (59%; Figure 8); however, this was largely attributed to post-ovipositioning females. Conversely, tailed frogs within the clearcut were evenly distributed between perpendicular and parallel movement. Fewer adults were present in these treatments which resulted in metamorphs influencing the observed orientation o f movement (50% o f captures) (Figure 8). Distance o f capture from stream O f the 160 individuals captured, 56.9% were located in the 5-m trap arrays (n = 80 in 2012 and n = 11 in 2013). As the distance from stream increased, the number o f individuals captured decreased with a total o f 44 tailed frogs captured in arrays > 55 m (Figure 9). Across all treatments, adults were most likely captured in arrays < 30 m from the larval stream (59.5% o f adult captures). 34 55 % 25 ftl 20 ni ll I d i U i 15 rti 10 5 OtdjpiMtk I Fonasl Ckaruut (Hdgnwdi Forest retention buffer retention bufftx Spring Clearcut O U pw H i F m tid a tk i nh Clearcut Fall CQ 3 5 30 30 25 25 20 15 15 10 5 O id I d I d OM-ftronrtlk Forest rctc&tioa Regencntioa buffer Old growth Forest rctmtioiL Rcgc&cratiot buOcr Old growth Forest retention Regeneration buffer Spnn« Summer Fall B Harvest treatment and season H 5-m array G 80-m array Figure 6: Microclimate arrays demonstrating the temperature variations between 4 forest retention treatments (old growth, forest retention buffer, regeneration and clearcut) located in the Gosling (A) and Kleanza (B) watershed, east of Terrace, BC, during 2012. 35 35 35 35 30 30 30 30 25 25 25 25 20 20 20 ID IU I 15 15 15 rh 10 b 0 A 5 0 Old growth Forest retention Clearcut i II 10 Old growth Forest Ckacnt retention buffer b u tler Old growth Summer ___________ Spring 10 5 n Forest Cleaicat retention buffer 11 Forest Clesiciit Old growli rotation buflfer F all(l) Fall (2) 35 35 35 JO 30 JO 25 25 25 20 20 15 10 5 0 OWgrowth Forest I rfl 15 10 Regeneration Old growth Old growth buffer Forest Regeoentiai retention buffer Spring F orest Regeneration retention buffer FaU B Harvest treatment and season B 5-m array Q 80-m array Figure 7: Microclimate arrays demonstrating the temperature variations between 4 forest retention treatments (old growth, forest retention buffer, regeneration and clearcut) located in the Gosling (A) and Kleanza (B) watershed, east of Terrace, BC, during 2013. I towards down away towards down away ■Adults 50 m retention buffer Old growth '© jl.fl ■Juveniles I OMetamorphs 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.2 i 0.1 0 towards down towards away down away up Clearcut Regeneration Orientation Figure 8: Proportions of tailed frogs captured for each age class relative to the direction of movement within each forest retention treatment (old growth, forest retention buffer, regeneration, and clearcut) for two watersheds located east of Terrace, BC, during 2012 and 2013. In the old growth a total o f 59 tailed frogs were captured (48 adults, 7 juveniles and 4 metamorphs). The forest retention buffer had 53 tailed frogs (38 adults, 11 juveniles and 4 metamorphs), the regeneration treatment had 26 tailed frogs captured (18 adults, and 8 juveniles), and the clearcut had 22 individuals captured (7 adults, 4 juveniles and 11 metamorphs). 1 0.8 u.6 0.4 02 0 in . 5 0.8 06 0.4 j j 30 55 i a2 80 i 5 ________ O k ! g ro w th 30 55 80 F o rest re te n tio n b u ffe r 08 70% of their captures in arrays > 5 5 m from the larval stream (71.5% in old growth and 75% in clearcut). Conversely, juveniles in the forest retention buffer and regeneration treatments had < 30% o f captures in arrays > 55 m from the larval stream (27.3% 50-m forest retention and 12.5% in regeneration). Forest retention treatment O f the 160 tailed frogs captured over the 2 years, 59 individuals were captured in the old growth (n = 2 sites), 53 were located in the forest retention buffer treatments (n = 2), 26 were found in the regeneration treatment (n = 1), and 22 were found in the clearcut treatment (n = 1; Table 2). However, capture rates for the three age classes o f tailed frogs differed among forest retention treatments. The greatest CPUE for adults was in the old growth treatments (1.3 adults/100 trap nights) with more juveniles (0.45 individuals/100 nights) and metamorphs (0.55 individuals/100 trap nights; Table 2) being captured in the regeneration and clearcut treatments, respectively. Additionally, the number o f tailed frogs captured in each o f the forest retention treatments varied across the two years o f sampling. In 2012, the greatest proportion o f captures was in the old growth treatments (3.0 tailed frogs/100 trap nights) whereas in 2013 the greatest capture rates was in the regeneration treatment (0.77 tailed frogs/100 trap nights). 39 no 30 ■ Female ■ Male mJuvenile □ Metamorph July (57) August (26) Month October (38) Figure 10: Total number o f reproductive and pre-reproductive tailed frogs captured by month during 2012 in 2 watersheds located east o f Terrace, BC. Numbers in brackets represent total captures within a month. 40 Temporal pattern o f movement Adult tailed frogs demonstrated differences in the rate o f capture depending on the time of year and gender (Figure 10). For example, females were found in traps during months associated with the species’ breeding phenology (i.e., spring and fall) and were distributed closest to the stream at this time. During July, females were either gravid or showed indications o f post-ovipositioning while in October captured tailed frogs displayed breeding characteristics such as darkened breeding pads in males and visible eggs in females. Consistent with the known chronology o f larval development, metamorphs began to emerge in August, as the oldest larvae metamorphose with their greatest occurrence in traps during October (Figure 10). Statistical models There was little model selection uncertainty for the 2012 capture year as two logistic regression models were responsible for the majority o f the AICc weight (Table 5). The highest ranked model contained temporal, treatment, and environmental variables. However, the second highest ranked model (distance to stream + trap month + site elevation2 + days since last rain) was only slightly more parsimonious (k = 8) and therefore selected as the ‘best’ model of the set. The ROC test based on independent data suggested that ‘good’ predictive performance for both o f the top ranked models (AUC = 0.753, SE = 0.026; AUC = 0.754, SE= 0.026) models. None o f the model variables had excessive multicollinearity (VIF > 10). 41 Table 5: Most parsimonious logistic regression models (AICc) for tailed frog capture data collected with pitfall arrays at 6 treatments located east of Terrace, BC, in 2012 and 2013. Top ranked models represented 95% of the AICc weight.__________________ Model R ank k AICc AAICc AlCcW 2012 distance to stream+densiometer+trap month+site elevation2+days since last rain* 1 9 953.54 0.0 0.506 distance to stream+trap month+site elevation2+days since last rain*t 2 8 953.61 0.1 0.491 days since last rain+site elevation2*f 1 4 366.15 0.0 0.926 distance to stream+densiometer+trap month+site elevation2+days since last rain* 2 9 373.54 7.4 0.023 orientation+days since last rain 3 6 374.29 8.1 0.016 2013 *Both the linear and quadratic terms were applied to models. fM ost parsimonious model. During 2013, 3 models explained 96.5% o f the AICc weight (Table 5). The top ranked model was by far the most parsimonious (AICcw = 0.926) consisting o f covariates for days since last rain and a quadratic term for site elevation. The second ranked model was the same as the top ranked model for the 2012 capture data (Table 5). The most parsimonious model for 2013 was ‘good’ at predicting the occurrence o f tailed frogs (AUC = 0.722 SE = 0.05). None o f the variables had excessive multicollinearity (VIF > 10). In 2012, not all coefficients within the top ranked model were statistically significant. The presence o f tailed frogs in traps was negatively correlated with distance to stream, days since the last rain event and the month o f August (p < 0.05; Table 6). Conversely, tailed frog presence was positively correlated with the time periods o f July and October; however, these coefficients were not statistically significant (p > 0.05; Table 6). The nonlinear elevation term (linear and quadratic) was positively correlated with the capture o f tailed frogs in sites between 440 to 460 m elevations. All coefficients within the top ranked model for the 2013 capture data were statistically significant (p < 0.05; Table 4). Similar to 2012, days since the last rain event was negatively correlated with tailed frog capture. The nonlinear elevation term (linear and quadratic) increased in 2013 and was positively correlated with the capture o f tailed frogs between 600 and 620 m. D is c u s s io n Declines in amphibian populations o f the Pacific Northwest may be related to timber harvest in streamside uplands. Understanding the distribution o f this Blue Listed species relative to timber harvest, can identify new and evaluate current Best Management Practices (BMP) designed to maintain tailed frogs across human-modified landscapes. 43 Table 6: Coefficients and measure of statistical significance, including 95% confidence intervals (Cl), for covariates from the most parsimonious logistic regression model (Table 5) for tailed frog capture data collected in 2012 and 2013 from pitfall arrays east of Terrace, BC.___________________________________________________________________________________________________ Variable Coefficient SE z Lower 95% Cl Upper 95% Cl P 2012 distance to stream -0.029 0.006 -4.841 <0.001 -0.041 -0.017 July 0.287 0.153 1.883 0.06 -0.012 0.586 August -0.438 0.148 -2.955 0.003 -0.728 -0.147 October 0.151 0.14 1.075 0.282 -0.124 0.425 site elevation 0.014 0.006 2.443 0.015 0.003 0.025 site elevation2 <-0.001 <0.001 -2.755 0.006 <-0.001 <-0.001 days since last rain -0.186 0.041 -4.555 <0.001 -0.266 -0.106 constant -5.409 1.389 -3.895 <0.001 -8.131 -2.687 days since last rain -0.435 0.119 -3.654 <0.001 -0.669 -0.202 site elevation 0.034 0.015 2.207 0.027 0.004 0.064 site elevation2 <-0.001 <0.001 -2.151 0.031 <-0.001 <-0.001 constant -14.497 4.462 -3.249 <0.001 -23.243 -5.750 2013 However, the effects o f these forest practices often are not evident until years after harvest (deMaynadier and Hunter 1995; Karracker and Welsh 2006; Petranka et al. 1993; Kluber et al. 2008; Hawkes and Gregory 2012). Information theoretic approaches for model selection have been relatively underused in the field o f herpetology (Mazerolle 2006). Although this approach is robust to bias, the models I tested only represent the data collected at the northern distribution o f the species. Those data may be limited by sampling bias that resulted from non-continuous trapping sessions that were most successful when tailed frogs were moving across the study sites. For example, females are highly mobile during the breeding season. Relatively infrequent and discontinuous 6-night trapping sessions could fail to represent that activity in any one year. Similarly, results suggest that a correlation between the timing o f trapping session and dryhot weather would reduce capture rate. Both factors might partially explain the small sample of captured tailed frogs in 2013 Despite the potential for sampling bias, the most parsimonious logistic regression models for both trap years were good predictors o f frog activity and distribution. Furthermore, both models contained variables representing temporal, treatment and environmental factors: 1) distance to stream; 2) trap month; 3) site elevation; and 4) days since the last rain event. By exploring the mechanisms that dictate patterns in frog activity, movement, and fine-scale distribution, we can identify land management practices that will aid in the conservation o f coastal tailed frogs. Orientation o f movement Although orientation was not a significant predictor for tailed frog presence in traps, this variable identifies increased activity associated with the specific life stage or 45 requirements o f the species. For example, Daugherty and Sheldon (1982) concluded that populations o f A. montanus were philopatric to streams and that pre-reproductive individuals were the main dispersers. This was because adults remained relatively sedentary (within 40 m o f capture), while juveniles were found to make more frequent movements along the stream. For populations o f A. truei in southern BC, a greater parallel (up and down) movement in relation to the stream edge was observed by Wahbe et al. (2000) and Matsuda and Richardson (2005). Upstream travel was greatest in adults while juveniles were captured moving perpendicular to the stream suggesting dispersal. Populations o f A. truei in northwestern BC demonstrated a greater tendency to move perpendicular (towards or away) to the larval stream. In 2012, the greatest proportions of tailed frogs were found moving away from the stream (41%). Increased activity resulting in the away orientation was largely attributed to the 31 reproductive females captured in the spring (84% o f total captures at that time). Conversely, in 2013 tailed frogs had the greatest proportion o f their movement towards the stream (56%). These were males trapped during summer in the forest retention buffer. This movement coincided with a prolonged period without rain (Appendix V). Wahbe and Bunnell (2001) documented significant downstream movement by larval tailed frogs. In order for populations to remain stable when one life stage is in a lotic habitat, a ‘colonization cycle’ (Muller 1974) must occur. During 2012, the CPUE for newly metamorphosed individuals was 1.6 times greater for upstream movement post-emergence than any other orientation. This upstream pattern o f movement may be instinctual in order to counter the downstream movement by larvae. 46 Distance o f capture from stream In 2012, distance from stream edge was an important predictor of tailed frog presence within traps. For all age classes, 57% o f captures were located in the 5-m streamside arrays. In addition, 82% o f adults captured in the forest retention buffer sites were located < 30 m from the stream edge. Similarly, studies in southern BC reported that adult tailed frogs had a greater affinity for areas within 20 m of the stream in clearcuts and mature forests (Matsuda and Richardson 2000; Wahbe et al. 2000). This suggests that the fine-scale distribution o f tailed frogs in terrestrial habitats can be explained by the location o f the maternal stream. Brosofske et al. (1997) reported that the microclimate remains relatively constant up to 31-62 m from the stream edge in forest stands o f the Pacific Northwest. This lack o f gradient in air temperature, soil temperature, and relative humidity might explain the variation in distribution of tailed frogs observed across my study sites. For example, during 2012 the average temperature in the old growth at the 5-m array was similar to the 80-m array during the spring trapping session (Figure 6). Capture rates for tailed frogs have been reported to vary according to time o f year and sex (Landreth and Ferguson 1967; Brown 1975; Wahbe et al. 2004; Hayes et al. 2006). Females in northern BC had a greater probability o f capture (64% o f adults) and were located < 30 m from the stream. This near-stream distribution may be indicative o f movements associated with reproduction; females captured at this time were either gravid or postovipositioning. Daugherty and Sheldon (1982) surmised that pre-reproductive individuals o f A. montanus undergo dispersal to locate food, shelter or new breeding grounds. Individuals dispersing from degraded or high-density areas can potentially find higher quality or more 47 available habitat (deMaynadier and Hunter 1999), but may experience greater mortality than sedentary individuals (Rappole et al. 1989). Similar to southern populations o f A. truei (Matsuda and Richardson 2005; Wahbe et al. 2004), a greater proportion o f juveniles were found in traps at the 55-m arrays for old growth and clearcut treatments. Bury and Com (1991) suggested that tailed frogs emerge in the fall. In northwestern BC, metamorphs were often trapped in streamside arrays during October (Figure 10). Furthermore, o f the 11 metamorphs captured in the clearcut, 7 individuals were newly metamorphosed individuals captured in arrays < 30 m from the streamside (Figure 9). Newly metamorphosed individuals may be more sensitive to ecological and physiological constraints relative to other life-stages (deMaynadier and Hunter 1999). Limited perpendicular movement from the stream edge is likely caused by a more suitable near­ stream microclimate (Brosofske et al 1997). Forest retention treatment Similar to studies on southern populations o f tailed frogs (Wahbe et al. 2004, Matsuda and Richardson 2005; Hawkes and Gregory 2012), adults were more abundant (43% o f total captures) in mature stands and decreased as canopy overstory was reduced (Figure 9). Although the treatment type was not significant in the model predictions, the variable identifying canopy closure at the trap array (densitometer), suggests that overstory cover is important in predicting tailed frog presence (Table 5). Previous studies for southerly populations o f A. truei have reported that larvae in clearcuts undergo earlier metamorphosis (3-4 years) and are found at higher densities (Richardson and Neill 1998; Wahbe and Bunnell 2001; Wahbe et al. 2004). However, this habitat type could act as an ecological sink, as newly metamorphosed individuals may be 48 more sensitive to ecological and physical constraints relative to other life-stages (deMaynadier and Hunter 1999). Although newly metamorphosed individuals existed in the clearcut treatment, few were captured in arrays further than streamside (Figure 9), identifying the importance of the near-stream microclimate for ameliorating the physiological constraints o f this life stage. Additionally, a greater number o f metamorphs, but fewer adults within clearcuts could be the result of either low recruitment o f metamorphs to adulthood or greater post-metamorphic dispersal from this habitat type (Wahbe et al. 2004). This may explain why fewer adults and a greater CPUE for metamorphs existed within this forest retention type compared to the old growth, forest retention buffer or regeneration treatments. Temporal patterns o f movement Quantifying seasonal variation in the movement o f terrestrial tailed frogs can increase our understanding o f dispersal and population connectivity as well as seasonal habitat requirements. Where studied, the seasonal movements o f Ascaphus have varied among species and populations. For A. montanus, Daugherty and Sheldon (1982b) found no difference in the direction o f movements across seasons. However, Matsuda and Richardson (2000) noted that the greatest rate o f movement for coastal populations o f A. truei in southern BC occurred during spring and fall. Others have proposed that breeding results in annual migration (Landreth and Ferguson 1967; Brown 1975; Wahbe and Bunnell 2001; Wahbe et al. 2004; Hayes et al. 2006); however, little or no data has been provided to support this prediction. Trapping data from this study suggests that increased seasonal movements correlate with the species breeding phenology. Specifically, movement to ovipositioning habitats were observed during June and July. Gravid females were caught moving towards or up the larval 49 stream in early June 2013 in the Kleanza sites. Furthermore, the majority o f adults traveling away from the stream in July 2012 were females that had deposited eggs. O f the 23 adult frogs captured in October, 2012, 9 females were gravid and 9 males had darkened breeding pads. In 2013, 6 adult frogs were captured in October with 4 gravid females and 1 male in breeding condition. These data suggest that in northwestern BC breeding begins in the late fall - something confirmed by direct observation (McEwan Pers. Obs.). Similar to Matsuda and Richardson (2000), I observed a negative correlation between tailed frog captures and the summer trapping session. In 2013, however, captures increased by a magnitude of 1.56 during the summer compared to the spring. This movement and the subsequent captures may have been a response to a prolonged period o f no rain (> 13 days; Appendix V). The majority o f captures (64%) were adults within the forest retention buffer and the regeneration sites moving towards the stream. With the onset o f fall, the number o f tailed frogs captured in pitfall arrays increased. Cool temperatures (< 11°C) and moderate to heavy rain events < 1 day apart increased tailed frog movement. For example, 8 metamorphs in 2012 were moving upstream (Figure 8) potentially to overwintering habitat while 15 individuals were located at arrays > 30 m (Figure 9). Metamorphosis o f populations in Washington and Oregon occurred during a brief time period in late summer (Bury and Adams 1999). Data from this study documented emergence from the larval stream beginning in late August with the majority o f captures occurring in mid-October (Figure 10). This would suggest that metamorphic emergence at the species’ northern range extends into the fall. 50 Elevation Trapping success, thus, the activity and movement o f tailed frogs, was correlated with the elevation o f the study sites. Amphibians found at high elevations have adapted to a colder thermal regime, but have physiological constraints that limit activity to a relatively narrow range o f temperatures (McCaffery and Maxell 2010). In this study area, elevation is strongly correlated with the onset o f the snow-free period and associated warmer air temperature. Thus, the timing o f ovipositioning, metamorphosis and breeding in addition to post-winter emergence and activity is related to the elevation at which the population (i.e., study site) is located. As o f June 1, 2012, there were still significant snow packs (195% o f normal) within watersheds of the Skeena River Basin (BC River Forecast Centre 2012). Sites below 528 m were snow free by the end o f May and had earlier emergence from overwintering locations than sites above 650 m where presence was not detected until mid-July. The majority of captures during 2012 were at 463 m. Conversely, by the beginning o f June 2013, snow indices were well below normal, suggesting that melt was 1-2 weeks earlier than normal (BC River Forecast Centre 2013). The first tailed frogs were observed by the end o f May when sites were accessible. The majority o f captures in 2013 occurred at 650 m. Capture data suggests that the timing o f emergence, ovipositioning and breeding is dependent on the start o f the snow-free period. In BC, temperatures are expected to increase (0.5°C per decade) with the northern portion warming faster. An increase in winter precipitation as rainfall and a decline in summer stream flow for snow-dominated water systems are also predicted (BCMWLAP 2002; Hamann and Wang 2006). This alteration o f 51 temperature regime and stream flow may affect the behavioural patterns and quality of terrestrial and aquatic habitat for the tailed frog in northern BC. C o n c l u s io n Drawing on the existing literature, I tested 3 predictions to describe and explain the fine-scale distribution and movement o f the tailed frog at their northern extent in western BC. Juveniles were more commonly located in arrays > 55 m from the stream edge in both the old growth and clearcut harvested sites, suggesting that juveniles are dispersing from high density or degraded habitats. Additionally, movement patterns o f reproductive aged tailed frogs did correspond with key reproductive periods. For example, during the spring, 34 females were captured displaying recent signs o f breeding (i.e., gravid or postovipositioning) and 59.5% o f adult females were captured in 5-m arrays next to the larval stream. This supports my prediction that movement patterns and the probability o f capturing females are related to reproductive events. Tailed frogs were more commonly located in traps following rain events (< 1 day) and when temperatures remained below the thermal maxima (26.5°C). Thus, pitfall traps are best at quantifying population distribution and relative abundance when tailed frogs are active and demonstrating above-ground terrestrial movements. Similar to studies in southern BC, the populations o f tailed frog I studied were most likely to be captured < 20 m from the stream edge regardless o f forest retention type. This refutes my prediction that tailed frogs in old growth habitats would be captured at greater distances from the stream edge. However, this result may simply reflect a behavioural preference for stream-side habitats by ovipositioning females and emerging metamorphs, regardless o f canopy closure. The capture rate o f adults did decrease as overstory was 52 reduced suggesting that forest harvesting does alter the distribution o f this age class within a treatment. Although pitfall traps can be useful for assessing temporal patterns o f behaviour, capture rates are likely sensitive to the trapping schedule. For example, had trapping occurred earlier in 2012, the emergence o f post-ovipositioning females from the larval stream may have been lower as was the case in 2013. For this technique to fully represent the seasonal ecology o f tailed frog, a more continuous duration o f trapping is necessary. Furthermore, more continuous trapping would address temporal effects associated with weather (e.g., rain, heat wave, or drought) that influence frog movement and capture success. Because o f time constraints, I only monitored the traps at each site for one 6-night period at 3 times during the growing season (May-October). However, when comparing the capture success o f my study to previous works on southerly populations, total numbers were similar. For example, over 3 years o f trapping, Wahbe (2003) captured 254 tailed frogs; while Matsuda (2001) captured 175 tailed frogs over 2 years. R e c o m m e n d a t io n s Managing forests fo r the tailedfrog Tailed frog research has focused on the impacts o f forest practices relative to larval development and aquatic habitat (Com and Bury 1991; Wallace and Diller 1998; Wahbe and Bunnell 2001) with few recent studies considering the post-metamorphic life stages (Wahbe et al. 2004; Matsuda and Richardson 2005; Burkholder and Diller 2007; Hawkes and Gregory 2012). Compared to most other anurans, tailed frogs are less tolerant o f warm temperatures and prone to desiccation. Thus, the species requires cold mountain streams and complex vertical and horizontal wood structure that is typically found in old forests (Aubry 53 and Hall 1991; Com and Bury 1991; Dupuis and Waterhouse 2001; K lubere/a/. 2008; Hawkes and Gregory 2012). When managing forests for this species, it is important to understand how post-metamorphic tailed frogs are distributed in relation to life-stage and the corresponding requisite environmental conditions. During 2013, temperatures were warmer and snow pack melted faster, triggering earlier emergence from over-wintering sites and ovipositioning. This was evident from fewer gravid or post-ovipositioning females captured during the spring. The majority o f tailed frogs captured over the two years were in the 5-m arrays, closer to rain events (< 1 day) and when air temperatures were < 26.5°C. Furthermore, tailed frogs were captured within each timber forest retention treatment, but as overstory declined, the number o f captures decreased with an increasing distance from the stream. These results confirm past work noting the relationship between the distribution and activity o f tailed frogs relative to micro-site or broader climatic conditions, and the need for intact overstory and complex forest structure (Daughtery and Sheldon 1982; Matsuda and Richardson 1999). Wahbe et al. (2004), for example, reported that relative to clearcuts, tailed frogs in old growth forests were distributed farther from the larval stream. Furthermore, results o f my study confirmed the prediction that tailed frogs will demonstrate more active and dynamic movement and distribution when weather conditions (e.g., rainfall events) are suitable for above ground terrestrial movements. One approach for maintaining post-metamorphic tailed frogs across managed forest stands is the designation o f riparian reserves (buffers; Bull and Carter 1996; Dupuis and Steventon 1999). Within this study, treatments with approximately 50-m retention buffer had similar abundance o f tailed frogs (adults, juveniles and metamorphs) as the old-growth 54 treatments. In addition, the distribution o f captured tailed frogs across all treatments was greatest within an area < 30 m from the edge o f the larval stream. These data suggest that a buffer o f insufficient size (< 50 m) will potentially exclude habitat typically used by tailed frogs. The Forest Practices Code Act o f BC identifies two riparian zones within the Riparian Management Area: 1) riparian reserve zone where harvesting is not permitted and 2) riparian management zone where constraints to harvesting exist. These zones now serve as guidance under the Forest and Range Practices Act. However, there is no regulatory requirement for a riparian reserve zone around smaller non-fish bearing streams typical o f the headwater systems where breeding populations o f tailed frogs are found. Within these streams, bank vegetation is left undisturbed if it is < 10 m from the stream edge and deemed to be nonmerchantable (BC Ministry o f Forests and Ministry o f Environment 1995). In addition, the riparian management zone that may exist on these smaller streams is 20-30 m from the stream edge and places minimal constraints on timber harvest practices. The results o f this study suggests that a < 20 m riparian management zone with minimal constraints to harvesting and a lack of a riparian reserve zone are not suitable for populations o f tailed frogs in northwestern BC. There is no consensus on the amount o f standing timber that should be retained as a buffer for stream obligate amphibians. Olson and Rugger (2007) suggest that a 6-76 m harvest buffer would retain amphibian populations found in small streams adjacent to stands in the Pacific Northwest that were harvested using moderate thinning practices. Stoddard and Hayes (2005) suggest that riparian reserve zones around small streams adjacent to clearcuts should be > 46 m to retain stream amphibians. Similarly, Perkins and Hunter 55 (2006) report that in Maine, buffers between 23-35 m are not adequate for maintaining riparian salamanders at an abundance comparable to non-harvested sites. Furthermore, Young (2000) suggests that a riparian reserve o f 70-90 m would allow for riparian-stream linkages with downed wood, litter, bank stability and suitable microclimate. Marsh and Trenham (2001) suggest that the maximum observed distance traveled by an amphibian species from the stream should serve as the minimum width o f a riparian reserve zone. The maximum distance traveled accounts for the nature o f incomplete sampling in most long­ distance movement studies. Riparian reserves should be established for tailed frogs around important streams and breeding areas (BCMWLAP 2004). The following 3 criteria would ensure that the reserves accommodate both the larval and terrestrial life stages: 1) the reserves should include the area near streams most likely to be used by tailed frogs on a seasonal basis and be at least 20 m; and 2) a riparian management zone should extend past the reserve zone and be large enough to ameliorate edge effects; and 3) management zones should be used to maintain connectivity between reserves (BCMWLAP 2004; Spears and Strofer 2008). Maintaining functional riparian reserves and managing upland forest as habitat for tailed frogs will ensure the continuance o f upstream movements by metamorphs and females to headwaters, increasing dispersal and gene flow. Furthermore, the retention o f riparian areas will increase the total area o f habitat for resident tailed frogs and act as a corridor for within site movement and dispersal. 56 C hapter 3 H a b i t a t S e l e c t i o n b y t h e C o a s t a l T a i l e d F r o g o f N o r t h w e s t e r n BC In t r o d u c t i o n Historically, conservation studies and efforts have focused on vascular plants, birds mammals or fish, but not on cryptic or nonmarket species (Griffiths and Dos Santos 2012). In North America, the US Endangered Species (1973) and the Canadian Species at Risk (2004) Acts, have resulted in the formal consideration o f nongame species (Semlitsch 2002). Unfortunately, a bias still exists for a number o f taxonomic groups, specifically amphibians, who receive relatively little attention (Griffiths and Dos Santos 2012). Current conservation strategies for terrestrial mammals and birds are insufficient for amphibians, a taxon that is declining globally (Houlahan et al. 2000; Dodd 2009). Because amphibians require both aquatic and terrestrial environments (Semlitsch 2002), conservation efforts and practices must target their biphasic life history. Although scientists agree that a number o f factors are contributing to the extirpation or extinction o f amphibian species, the most wide-spread cause is believed to be habitat alteration (Houlahan et al. 2000; Semlitsch 2002; Cushman 2006; Dodd 2009). Habitat is lost or altered by draining and filling wetlands, channelizing streams, creating impoundments, and removing native vegetation. Fortunately, these causal processes can be identified, minimized, and even reversed. Given the scope o f threats to amphibian species, habitat alteration is relatively easy to address through land-use planning and the protection of habitat (Semlitsch 2002). Such strategies first require a better understanding o f the interactions between land use and the ecological and physiological requirements o f the focal amphibian species. Information describing distribution and movement is necessary to understand the ecology o f a species. For example, many amphibians use aquatic sites for breeding and as a 58 developmental environment for larvae whereas adults are found in terrestrial habitats. Thus, distributional patterns can reflect age-specific variation in ecological requirements (Daugherty and Sheldon 1982). However, our knowledge o f the spatial ecology o f many amphibian species during the terrestrial life stage is limited (Rowley and Alford 2007). Terrestrial habitats may be critical for maturation and ultimately reproduction by adults, but are overlooked during conservation planning. There is a well-established empirical relationship between the area and ecological characteristics o f forest cover and the occurrence o f many amphibian species (Dupuis and Steventon 1999; Ascaphus Consulting 2003; Wahbe et al. 2004; Matsuda and Richardson 2005). For example, Nuzzo and Mierzwa (2000) found that the greatest abundance and diversity o f amphibians were in forests that retained ‘natural’ characteristics including overstory, ground level vegetation, and coarse woody debris. Others have reported a negative response o f amphibian populations to anthropogenic activities such as roads (Farhig et al. 1995; Gibbs 1998) or timber harvest (Welsh and Ollivier 1998; deMaynadier and Hunter 1998, 1999; Dupuis et al. 2000; Chan-McLeod 2003; Hawkes and Gregory 2012). As an example, Johnston (1998) found that timber harvest restricted the movement and altered the use o f habitat by the Pacific giant salamander (Dicamptodon tenebrosus). Previous research in the Pacific Northwest has established that amphibians are sensitive to forest management practices (Perkins and Hunter 2006). Additionally, amphibian abundance has shown a positive correlation with the structure and complexity o f forest stands related to harvest history (Aubry and Hall 1991; Com and Bury 1991; Wahbe et al. 2004). For example, Maxcy and Richardson (1999) concluded that 4 salamander species responded negatively to forest harvesting because o f a loss o f stable climatic conditions (e.g., 59 temperature and humidity) that old growth stands provide. Amphibians require habitats with high moisture and a narrow range o f temperatures (Palis 1997, 1998; deMaynadier and Hunter 1998; Chan-McLeod 2003; Baldwin et al. 2006; Perkins and Hunter 2006). In the case of southerly populations o f the coastal tailed frog (Ascaphus truei), Matsuda and Richardson (2005) concluded that adults were less abundant in clearcuts; however, movements or the distance traveled from stream was not influenced by forest harvest. Additionally, Hawkes and Gregory (2012) concluded that the abundance o f tailed frogs was negatively affected by upland logging 10 years post-harvest. In BC, many populations o f amphibians are at risk because o f anthropogenic activities, including urban development, environmental contamination, logging, and water impoundment. These threats have resulted in ~ 25% o f the amphibian species in the province to be listed as Endangered or Threatened (BC Environment 2002). The coastal tailed frog is on the provincial Blue List and as a species o f Special Concern under the Federal Species at Risk Act (SARA). This species is susceptible to threats from urban development, timber harvest, and independent power projects (IPP). These activities remove canopy overstory, reduce coarse woody debris, and can impact water quality and flow within larval streams. The majority o f research and inventory studies o f the coastal tailed frog have considered the distribution and habitat requirements o f the larva (Wahbe 1996; Wallace and Diller 1998; Bury and Adams 1999; Dupuis et al. 2000; Ritland et al. 2000; Wahbe and Bunnell 2001). Until recently, few studies have focused on the ecology o f the terrestrial stage o f the species (Wahbe et al. 2004; Matsuda and Richardson 2005; Hayes et al. 2006; Burkholder and Diller 2007) and none have investigated the ecology o f this species at the northern extent o f its range. 60 C h a p t e r P r e d ic t io n s In this chapter, I investigated environmental factors that influenced resource selection, movement, and distribution o f adult tailed frogs across a range o f terrestrial habitats affected by timber harvest. I used site and environmental data collected at the patch (micro-habitat) and forest stand (macro-habitat) scales to quantify resource selection o f adult frogs monitored with radio telemetry. I tested 3 predictions that were supported by existing literature and our general understanding o f the habitat ecology o f forest-dependent amphibians. Forest structure and resource use - I predicted that the distribution o f tailed frogs would be positively associated with habitat features that ameliorate variability in climate, increase availability o f food resources and provide protection from predators; this reflects previous studies identifying the importance o f downed wood and other subterranean structures as habitat for amphibians. Temporal, climate and resource use - I predicted that during warmer and drier portions o f the year, tailed frogs would select locations close to a stream and with habitat features, such as coarse woody debris, that provided protection from thermal stress and dehydration; this reflects previous work identifying the importance o f thermal and hydricregulation. Forest retention treatment and resource use —I predicted that as forest overstory declined, tailed frogs would have a greater association with habitat features, such as moist microsites or coarse woody debris that maintained physically suitable microclimates. 61 M ethods Study Area The study area was situated on the leeward side of the Coastal Mountains in northwestern British Columbia. Vegetation communities included the Coastal Western Hemlock (CWH) and the Interior Cedar-Hemlock (ICH) Biogeoclimatic Ecosystem Classification zones (Meidinger and Pojar 1991). The CWH had mild temperatures and heavy rainfall with climax forests o f western hemlock (Tsuga heterophylla) and amabilis fir (Abies amabilis). The ICH, reflective o f a transitional climate between coastal and interior stands, contained climax forests o f western hemlock and subalpine fir (Abies lasiocarpa). The main disturbance across these study sites was timber harvest. See Chapter 1 for a more detailed description o f the study area. I collected radio telemetry data during 2011 and 2012 for 24 adult tailed frogs across 3 watersheds: 1) Gosling (GO); 2) Ascaphus (AS); and 3) Kleanza (KL). Sites in each watershed were classified as 1 of 3 categories representing forest harvest history: 1) mature second growth or no history o f harvest; 2) harvested, but with habitat retention o f 30-50 m from the edge o f the larval stream (i.e., forest retention buffer); and 3) harvested with minimal to no overstory canopy. The GO sites were located within the transition zones between the CWHws and the ICHmc2 (Meidinger and Pojar 1991). Study sites were classified as CWHws 1 or CWHws2 and situated at relatively low elevations (319-463 m above sea level (asl)). Where forested, these sites were dominated by old (> 250 years) western hemlock and amabilis fir. Within the understory, moist pockets o f devil’s club {Oplopanax horridus), skunk cabbage (Lysichiton americanus) and abundant Vaccinium species existed. Most harvesting in this watershed occurred recently (< 5 years) resulting in 62 no tree canopy and the substantial growth o f fireweed (Chamerion angustifolium) and bracken fern (Pteridium aquilinum. Similarly, the AS watershed was located at low elevations (319-582 m asl) within the transitional zone between the CWHws and the ICHmc2 (Meidinger and Pojar 1991). Study sites were classified as CW Hwsl and situated at low elevations (273-582 m asl). Where forested, these sites were dominated by old (>140 years) western hemlock and subalpine fir with some western red cedar (Thuja plicata) and Sitka spruce (Picea sitchemis). Within the understory, moist pockets o f devil’s club and skunk cabbage were present amongst abundant Vaccinium species and false azalea (Menziesia ferruginea). Most harvesting in this watershed occurred 10-15 years in the past resulting in the regeneration o f dense stands of western hemlock and lodegepole pine (Pinus contorta) < 10 m tall. The sites within the KL watershed were found within two variants o f the Coastal Western Hemlock BEC zone (CWHws2, CW Hwsl). Two sites were located at higher elevations (650 m and 801 m asl; CWHws2). The old growth treatment had > 250 year old western hemlock, amabilis and subalpine fir in a multi-storied canopy; while the forest retention treatment contained 50 m o f > 250 year old western hemlock, amabilis and subalpine fir along the stream. Both treatments contained moist pockets with abundant devil’s club and skunk cabbage in the understory. The third site was lower in elevation (528 m asl; CW Hwsl) and had dense regeneration o f western hemlock and lodgepole pine < 20 m tall. This site was harvested in 1987 and had poorly developed understory with moist pockets o f devil’s club and lady fern (Athyrium filix-femina). 63 Data Collection Animal capture and radio telemetry —Adult tailed frogs were captured using 3 methods: 1) pitfall traps; 2) systematic visual encounter surveys (VES); and 3) incidental captures. Frogs were captured during July-October, 2011 in the AS and KL sites using nonsystematic visual searches o f riparian forests adjacent to known larval streams. In 2012, pitfall traps were opened for a total o f 5,184 hours (Chapter 2). The VES consisted o f 3 consecutive survey days in each site during the spring, summer and fall o f 2012. A total o f 15 person-hours were completed per site during each VES. Incidental captures occurred in both years. When captured, tailed frogs were identified as male or female and measured for snout vent length (SVL), shank length (SL) and weighed to the nearest 0.01 g. Only individuals > 5.5 g were fitted with Very High Frequency (VHF) radio transmitters (8-10% o f body weight) and relocated for 5-24 days. I used BD-2N (0.45 g) and BD-2X (0.34 g) transmitters in 2011 and 2012, respectively (Holohil Systems Ltd. Carp, Ontario). To attach the transmitter, I used a belly-belt system (Muths 2003) consisting o f Gossamer Floss™ jewelry cord (B. Toucan Inc., USA) and Japanese glass seed beads (olive green, size 15, Figure 11). In total, the belt and transmitter weighed approximately 0.54 g and 0.46 g, for the respective transmitter models, with a nominal lifespan o f 13-22 days. O f the 24 individuals followed, no mortality was documented; however, the fates o f 7 individuals were unknown due to tag failure (n = 4) or an inability to recapture and remove the tag (n = 3). I used a hand-held receiver (R-1000 Communications Specialist Inc., Orange, CA) to relocate each telemetered tailed frog once per day. 64 Figure 11: VHF radio transmitter (2) with belly-belt attachment (1) and antenna (3) used to relocate coastal tailed frogs. 65 Daily relocations were conducted between 7 am and 9 pm and staggered so that at least 15 hours had passed between consecutive relocations. I concluded the search when I had visual confirmation o f the individual. When tailed frogs were located within or under debris, I used multiple triangulation points and signal strength to identify the relocation with a high level of confidence. Statistical Methods Data analysis I used resource selection functions (RSF) to relate the locations o f monitored tailed frogs to 9 independent variables measured at two spatial scales (Table 7). An RSF can be calculated using a number o f statistical techniques (Johnson et al. 2006). I used conditional fixed-effects logistic regression as this method provided a more sensitive measure o f resource availability, controlling for temporal variability in the availability o f resources, as use locations are paired with available locations sampled from an ecologically relevant area (Johnson and Gillingham 2005). I collected resource data at the observed location o f the tailed frog (used habitat) and a paired random location (available habitat; Manly et al. 2002). Random locations were selected according to a random direction and distance (3-20 m) centered on the most recent location o f the tailed frog. Few studies have documented the mean daily movements o f tailed frogs (Wahbe et al. 2004; Maxcy 2000), thus, I inferred a minimum distance from previous research for a related species (Leiopelma spp.; Cree 1989; Newman 1990). To prevent the macro-habitat sampling plots (100 m2) o f the paired random locations overlapping with the used locations, I added 11.28 m to each random distance (Figure 12). 66 Spatial scales I collected resource data at two spatial scales using nested plots centered over the used or available locations (Figure 12). The micro-habitat scale (Mi; 1-m2 plot), represented the environmental variables assumed to influence the distribution o f monitored tailed frogs relative to fine-scale physiological constraints (e.g., temperature and light). At the macro­ habitat scale (Ma; 25-m2 plot), measured variables accounted for stand-level forest attributes (e.g., site moisture, tree species, CWD volume, and percent canopy opening). Micro-habitat variables - At each tailed frog and paired random location, I measured light intensity (lux; Extech Light Meter) and microclimate (air temperature (°C) and relative humidity (%); Kestrel 4500 weather meter). I used a Gaussian term to represent the nonlinear effect o f temperature on the distribution o f monitored frogs. This equation consisted o f both the linear component, identifying the increase in the relative probability o f use o f a site, and a squared component to identify where the relative probability o f use decreases as a function o f increasing temperature. Macro-habitat variables —Each habitat plot was classified as 1 o f 3 categories according to the distance o f the array from the clearcut (hard) edge: 1) located between the stream edge and > 10 m from the clearcut edge; 2) < 10 m on either side o f the hard edge; and 3) > 10 meters from the hard edge within a clearcut. I used a spherical densiometer to quantify the percent canopy closure above each location. 67 Used Location Available location 5-mCWD transect Random Bearing Micro- habitat -m2ploi Random Distance between 3-20m Macro-habitat 25-m2plot Figure 12: Illustration o f the nested sampling design for recording habitat characteristics at used and available locations for tailed frogs at the micro- (1 -m2plot) and macro-habitat (25-m2 plot) scales. Arrows represent the 5-m coarse woody debris transect lines. I centered a 100-m2 plot (5.64-m radius) on each location. The depth o f the organic layer was identified to the nearest 0.1 cm (BC MELP 1998). Additionally, I measured tree diameter at breast height (dbh), vigor (living or dead), number o f root hollows, and the Wildlife Tree Classification (BC MELP). Root hollows were openings at the base o f a tree where resources such as food, security, or refuge habitat could exist. These openings had to enter the tree cavity to a depth o f 5 cm and could be found within any living or dead standing tree. Openings were classified as 1 o f 3 categories: 1) < 10cm; 2) > 10cm; and 3) both sizes and quantified based on the total number present within a location. Within each plot I assessed the amount and composition of CWD along 3 5-m transects oriented > 120° from an initial random bearing for a total line length o f 15 m at each location (Figure 12). I measured the length, decay class and species for each piece o f CWD > 7.5 cm in diameter that was intercepted by a transect. I condensed the 5 classes o f CWD adopted by the BC MELP (1998) into 3 decay classes: 1) contained pieces that still retained shape and could be identified to species; 2) contained pieces that had bark sloughing and heartwood was beginning to soften; and 3) contained pieces that were decomposed and laying on the forest floor or had vegetation beginning to grow within the log. Using Van Wagner’s (1968) equation, I quantified the CWD volume by decay class at each location: V= (7i2/8L)Xd2 where V is the volume per unit area (m3/m2), L is the length o f transect (m), and d is the diameter o f the log at the intercept (m). I used the Ecosystems o f BC (Meidinger and Pojar 1991) for the Prince Rupert Forest Region and indicator plant species within the 25-m2 plot, to quantify site moisture as: 1) ‘dry’ if no species associated with wet ecosystems were present; 2) ‘mesic’ when devil’s club was 69 < 10% with minimal oak fern (Gymnocarpium dryopteris); 3) ‘subhygric’ when < 20% devil’s club was present and dominant in oak fern; and 4) ‘hygric’ when the site had > 20% devil’s club and was dominated by lady fern or trees with affinity for wet soil (e.g., Salix spp.). Model selection and evaluation I used Akaike’s Information Criterion for small samples (A IC c) to rank and select the most parsimonious model representing the resource selection o f monitored tailed frogs. I used A AICc to identify the ‘best’ model in each o f the sets and Akaike weights (AICc w) to quantify model selection uncertainty. If models were nearly equivalent (i.e., A AICc < 2), then I selected the model with the fewest number o f parameters (Johnson et al. 2006). To control for potential differences in resource selection between gender, I developed model sets using locations from the female (n = 16 individuals) and male (n = 8 individuals) tailed frogs. Also, I pooled all locations and developed a set o f models that represented the general resource selection patterns o f the sample o f tailed frogs that I monitored. I calculated the Area Under the Curve (AUC) for the Receiver Operating Characteristics (ROC) to test the predictive performance o f the most parsimonious models. The curve represents the proportion o f correctly and incorrectly classified predictions over a continuous range of probability thresholds from 0 to 1.0 (Pearce and Ferrier 2000). Scores o f 0.5-0.7 suggest a poor model, 0.7-0.9 a good model, and 0.9-1.0 a highly predictive model (Swets 1988). When performing the ROC test, I withheld paired used and available sites sequentially from the model building process and then used that withheld data to generate a predicted probability. Thus, a bootstrapping-like process resulted in an independent evaluation o f the predictive performance o f the logistic regression model. 70 Table 7: Independent variables used to derive RSF models for the coastal tailed frog in northwestern BC at the micro- (Mi) and macro­ habitat (Ma) scales. V ariable Description V ariable M acro-habitat distance to edge CWD volume litter root hollows Distance o f trap arrays from hard edge: 1) locations from stream edge to < 10m from hard edge; 2) locations ± 10m of the hard edge; and 3) locations > 10 meters from a hard edge. Straight-line distance from the larval stream edge. Percent canopy opening. Inferred site moisture based on indicator plants in the 25-m2 plot: 1) dry; 2) mesic; 3) subhygric; and 4) hygric. Calculation o f CWD volume for 3 decay classes using Van Wagner’s (1968) equation. Depth (cm) of the organic horizon (litter, fermented and humus; LFH) and duff layer. Number o f openings at base of tree: 1) < 10cm; 2) > 10cm; and 3) both sizes. M icro-habitat temp* humidity light Temperature (°C) at location Relative humidity (%) at location Light intensity (lux) at location distance to stream densiometer site moisture * Variable had Gaussian term applied to represent a non-linear relationship. I used 95% confidence intervals to assess the relative strength of selection or avoidance for each o f the covariates within the most parsimonious model. When intervals do not overlap zero, covariates are considered as a significant factor influencing the distribution o f tailed frogs. I used a modified variance estimator to statistically control for autocorrelation that might be expected from the repeated sampling o f locations from a monitored tailed frog (Williams 2000; Nielson et al. 2001). I used variance inflation factors (VIF) to test for multicollinearity among model covariates for the tested models. If a variable had a VIF value > 10 or a mean VIF value for the model > 1.0 it was removed from the model (Chatterjee et al. 2001). All statistical analysis was conducted using Stata (ver.12.1, StataCorp LP, 2012). Patterns o f movements I calculated the 90% minimum convex polygon (MCP) for 17 tailed frogs with > 10 locations (Arcview 3.2, Animal Movement Analysis extension). A MCP is the smallest polygon that would represent the boundary locations o f an individual’s movement during a monitoring period (Hayne 1949). The estimated area o f use represents the short-term daily activity o f the individual over the period o f monitoring. Given the limited period o f monitoring for each frog, the area o f use could be influenced by short-term seasonal behaviours such as migration to access breading locations. R esults Radio-telemetered animals In 2 years o f radio telemetry, a total o f 24 adult tailed frogs (16 females; 8 males) were followed between 5-24 days (mean 10.75 days, SE = 0.89) for a total o f 573 used and paired locations. Females were on average larger (Chapter 2) and the mean tag weight for both genders was rarely outside o f body weight guidelines o f 5-8% (n = 3 > 8.00% body weight; 72 Richards et al. 1994). O f the 24 tailed frogs, 11 were in the old growth (6 females; 5 males); 10 in the forest retention buffer (7 females; 3 males); and 2 females were monitored in the clearcut treatment (Table 8). No mortality was documented; however, the fates o f 7 individuals were unknown due to tag failure (n = 4) or an inability to recapture and remove the tag (n = 3). Gender differences in movement and space use The distance moved between daily relocations ranged from 0-45.7 m for all frogs ((mean 5.49 m, SE = 0.53). Generally, female frogs had greater daily movements when compared to males (females: 73.6 m, SE = 17.4; males: 41.0 m, SE = 9.5; Figure 13). The MCPs for individual frogs were between 116 to 3386 m2 with a mean estimate o f 628.12 m2 (SE = 190.04). Tailed frogs relocated in the old growth had larger estimates o f space use (977.50 m2, SE = 368.65) than those in the forest retention buffer (320.57 m2, SE = 92.37) or clearcut treatments (307 m2, SE = 71.00); however, the clearcut estimates were based on 2 individuals. Females had larger mean estimates o f space use compared to males (730.50 m2, SE = 317.33; 481.86 m2, S E = 109.83). Forest retention treatment Adult tailed frogs monitored in the old growth treatment traveled 23 m further from the origin o f capture compared to the clearcut and forest retention buffer (77.1 m, SE = 22.8; 54.3 m, SE = 19.2; 40 m, SE = 8.1, respectively); however, only the old growth and the forest retention buffer treatments differed significantly (Figure 14). Additionally, adults had the greatest daily movement rates in the old growth (7.4 m/day, SE = 0.8) compared to the clearcut (5.2 m/day, SE = 2.0) and forest retention buffer treatments (3.7 m/day, SE = 0.8). 73 Table 8: Telemetry data for 24 adult tailed frogs identifying the average number of days followed, weight (grams) and the percent body weight o f the tag for each forest retention treatment in 3 watersheds located east o f Terrace, BC, during 2011 and 2012. Numbers in brackets are standard errors (SE)._________________________________________________ Average days Total Forest retention % body Average captures m onitored treatm en t w eight w eight Old growth Female Male 11 6 5 10.27(1.67) 9.67 (3.07) 11.00(1.00) 9.67 (0.57) 9.92 (0.71) 7.18(0.15) 6.14(0.42) 5.7 (0.31) 7.42 (0.27) Forest retention buffer Female Male 11 8 3 10.91 (1.02) 10.38(1.32) 12.33 (1.20) 9.79 (0.79) 10.91 (2.18) 6.82(1.74) 5.86 (0.50) 5.00 (0.29) 8.15(0.33) C learcut Female Male 2 2 0 12.5 (1.5) 12.5(1.5) 0 10.08 (0.74) 10.08 (0.74) 0 4.90 (0.56) 4.90 (0.56) 0 74 ■3 female (2) C &> O male (0) oS - . - g u female (7) 2 ------------------ u c3 u I/) c ta s> 3 a male (3) £ female (6) S a0 T3 O male (5) 0 2 4 6 8 10 12 14 Mean daily distance (m) moved Figure 13. Comparison o f the mean daily distance (meters) moved (95% confidence intervals) for 24 adult tailed frogs relocated in 3 forest retention treatments (old growth, forest retention buffer, and clearcut) for 3 watersheds east o f Terrace, BC, in 2011 and 2012. Numbers in brackets indicate total number (n) o f frogs relocated within each forest retention treatment. 75 Clear cut (2) ------------- i— £ o E 13 a ^ Forest retention buffer (11) - ' H b Old growth (11) 0 10 20 30 40 50 60 70 80 90 100 110 Mean distance (m) traveled from origin o f capture Figure 14: Comparison o f the mean distance (meters) traveled from the origin o f capture (95% confidence intervals) for 24 adult tailed frogs relocated in 3 forest retention treatments (old growth, forest retention buffer and clearcut) for 3 watersheds east of Terrace, BC, in 2011 and 2012. Numbers in brackets indicate total number (n) o f tailed frogs relocated within each forest retention treatment. 76 The total mean distance traveled from the stream edge decreased in relation to forest harvest and removal o f canopy (Figure 15). In the clearcut, the relocations o f tailed frogs were on average 2.15 m (SE = 0.55) from the stream edge. However, only two female frogs were monitored in this treatment type. In contrast, female tailed frogs in the forest retention buffer (n = 8) and old growth (n = 6) treatments moved further from the stream edge (27.37 m, SE = 1.72; 64.14 m, SE = 5.11, respectively). Conversely, males had similar movement patterns in relation to the stream for the forest retention buffer (n = 3) and old growth (n = 4) treatments (37.58 m, SE = 1.19; 36.88 m, SE = 3.26, respectively). As overstory was reduced, tailed frogs had a greater proportion o f relocations associated with wet site types. In the old growth treatments, 53.9% o f the used locations were in the dry site types. Whereas in the clearcut, 33.33% o f the relocations were associated with the wet site type compared to only 14.81% o f relocations being associated with the dry site types (Table 9). Females appeared to have greater sensitivity to site moisture when canopy overstory was reduced. Females in the old growth were relocated in the drier site types more often (84.4%) than in the wetter site types (15.6 %;). Temporal movement Female tailed frogs generally moved further in the spring (8.32 m, SE = 1.34) while males had greater mean daily movements in the summer (4.49 m, SE = 0.60; Table 10). Although there was no statistical difference in mean daily movements between seasons; in the old growth sites, the greatest mean daily movements occurred during the spring and summer. Conversely, the clearcut treatments had the greatest mean daily movements during the fall although those measures were highly variable and were restricted along the stream edge (Table 11). 77 I 2.15 (n = 2) Clearcut 0c 1O u 27.73 (n = 8) Forest retention buffer IFemale 37.58(n = 3) a Male 64.14(n-6) Old growth 36.88 (n 0 20 40 60 80 Mean distance (m) moved from stream Figure 15: Mean distance (m) moved (95% confidence intervals) from the stream edge for 24 adult tailed frogs relocated in 3 forest retention treatments (old growth, forest retention buffer, and clearcut) for 3 watersheds located east o f Terrace, BC, during 2011 and 2012. 78 Table 9: Radio telemetry data for 24 adult tailed frogs representing the proportions of relocations associated with 4 site moisture types (dry, mesic, subhygric or hygric) for each gender located in 3 forest retention types (old growth, forest retention buffer, and clearcut) for 3 watersheds east o f Terrace, BC, during 2011 and 2012. Numbers in brackets represent sample sizes.________________________ Forest retention treatment Dry Mesic Subhygric Hygric Female Male Fermale Male Female Male Female Male Old growth (128) 33.8 (23) 35.9 (46) 47.1 (32) 5.5 (7) 11.8 (8) 3.9(5) 3.9(5) 1.6(2) Forest retention buffer (134) 24.6 (33) 19.4 (26) 11.9(16) N/A 14.9 (20) 5.2(7) 17.2 (23) 6.7 (9) Clearcut (27) 14.8 (4) N/A 29.6 (8) N/A 22.2 (6) N/A 33.3 (9) N/A Table 10. Mean distance traveled (m) per day for 24 adult tailed frogs relocated during 2011 and 2012 in 3 watersheds east of Terrace, BC, within 3 forest retention treatments (old growth, forest retention and clearcut). Numbers in brackets represent standard errors (SE)._________________________________________________ G ender Summer Spring Fall Female Male 7.09 (1.44; n = 5) 4.49 (0.60; n = 5) 8.32 (1.34; n =5) 3.20 (0.52; n = 7) 3.74 (1.27; n = 6) 0.17 (0.08; n = 1) Table 11: Seasonal daily movement rates for 24 adult tailed frogs located in 3 forest retention treatments (old growth, forest retention buffer and clearcut) for 3 watersheds east of Terrace, BC, during 2011 and 2012. Dates associated with season are located in Table 2. Numbers in brackets represent standard errors and sample sizes (n).__________________________________________________ Fall Forest retention Spring Sum m er treatm ent Male Female Male Female Male Female Old growth N/A 9.56 3.1 11.65 5 N/A (1.78, n = 3) (0.73; n = 1) (0.74; n = 1) (2.5; n = 3) Forest retention buffer Clearcut 6.2 (1.92; n = 2) 3.32 (0.79; n = 1) 6.76 (2.08; n = 1) 2.93 (0.74; n = 1) 2.82 (1.08; n = 5) 0.17 (0.08; n = 1) N/A N/A 3.44 (0.88; n = 1) N/A 7.5 (4.45; n = 1) N/A Resource selection When considering the location data for male and female tailed frogs combined, two logistic regression models were responsible for the majority of the AICc weight (Table 12). Although, there were nearly twice as many relocations for female frogs, thus, that gender had a disproportionate effect on model selection. The highest ranked model (light + temp2 + moisture + distance to stream + dc2 + dc3 + densiometer) for the pooled location data contained treatment and environmental variables. The second highest ranked model (light + temp2 + moisture + distance to stream + dc2 + dc3) was slightly more parsimonious as it did not include a covariate for canopy closure. This model was selected as the ‘best’ o f the set, but the ROC score suggested ‘poor’ to ‘good’ (considering statistical uncertainty) predictive performance (AUC = 0.696, SE = 0.022). The relocation data for females resulted in 5 logistic regression models representing > 95% o f the AICc weight. The top ranked model contained a combination o f environmental and treatment variables (light + temp2 + moisture + distance to stream + dc3). However, the second ranked model (light + temp2 + moisture + distance to stream) had fewer parameters (k = 8) and was therefore selected as the ‘best’ o f the model sets. The ROC score for this model suggested ‘good’ predictive performance (0.734, SE = 0.027). For male tailed frogs, > 95% o f the AICc weight was represented by 5 models. The top ranked model consisted of only treatment variables (distance to stream + del + dc2 + dc3; Table 13). Resource selection models generated for male tailed frogs had consistently ‘poor’ predictive performance (A U C < 0.676). 81 Table 12: Most parsimonious logistic regression models (AICc) for relocation data o f 24 tailed frogs in 3 watersheds east of Terrace, BC, in 2011 and 2012. Top ranked models represented > 95% of the AICc weight (AICc w); the area under the curve (AUC, Standard Error) represents the measure of predictability for each model. Model R ank k AICc AAICc AICc w AU C (SE) light+temp2+distance to stream+dc2+dc3+densiometer+moisture* 1 11 274.381 0.0 0.573 0.706 (0.022) light+temp2+distance to stream+dc2+dc3+ moisture*f 2 10 275.001 0.6 0.420 0.696 (0.022) light+temp2+distance to stream +dc3+moisture* 1 9 134.860 0.0 0.389 0.764 (0.025) light+temp2+distance to stream +moisture*f 2 8 135.409 0.5 0.294 0.734 (0.027) light+temp2+distance to stream+dc2+dc3+ moisture* 3 10 137.353 2.5 0.111 0.772 (0.025) light+temp2+distance to stream+dc3+densiometer+ moisture* 4 10 137.590 2.7 0.099 0.765 (0.025) light+temp2+distance to stream+densiometer+ moisture* 5 9 137.940 3.1 0.083 0.736 (0.026) dcl+dc2+dc3+distance to streamf 1 4 103.702 0.0 0.401 0.655 (0.039) dcl+dc2+dc3+distance to stream+light 2 5 104.150 0.4 0.321 0.663 (0.039) dcl+dc2+dc3 3 3 106.448 2.7 0.102 0.676 (0.038) dcl+dc2+dc3+light+temp2+ distance to stream* 4 7 107.001 3.3 0.077 0.630 (0.040) dcl+dc2+dc3+distance to stream+temp2 +densiometer* 5 6 107.786 4.1 0.052 0.635 (0.040) Female and M ale Female Male *Both the linear and quadratic terms were applied to model. tM ost parsimonious model identifying forest attributes post-metamorphic tailed frogs use. Table 13: Coefficients and measure of statistical significance, including 95% confidence intervals (Cl), for covariates from the most parsimonious logistic regression models (Table 13) generated using relocation data for tailed frogs collected in 2011 and 2012 from 3 watersheds east of Terrace, BC.____________________________________________________________________________________ Coefficient SE z P Lower 95% Cl Upper 95% Cl Female and Male light temp temp2 distance to stream decay class 2 decay class 3 dry mesic subhygric hygric -0.001 0.878 -0.029 -0.019 0.001 <0.001 -0.750 -0.146 -0.119 1.015 0.001 0.377 0.016 0.010 <0.001 <0.001 0.310 0.231 0.242 0.498 -1.71 2.33 -1.87 -1.83 1.7 2.36 -2.42 -0.63 -0.49 2.04 0.087 0.02 0.062 0.068 0.089 0.018 0.016 0.527 0.623 0.042 -0.002 0.140 -0.059 -0.040 <-0.001 <0.001 -1.358 -0.599 -0.594 0.039 <0.001 1.617 0.001 0.001 0.001 <0.001 -0.142 0.307 0.356 1.991 Female light temp temp2 distance to stream dry mesic subhygric hygric -0.003 0.208 0.008 -0.013 -1.195 -0.333 0.011 1.516 0.001 0.319 0.020 0.017 0.440 0.291 0.275 0.691 -2.91 0.65 0.42 -0.75 -2.71 -1.14 0.04 2.19 0.004 0.515 0.671 0.454 0.007 0.253 0.967 0.028 -0.005 -0.417 -0.030 -0.046 -2.057 -0.904 -0.527 0.161 -0.001 0.833 0.047 0.020 -0.332 0.238 0.550 2.871 Male distance to stream decay class 1 decay class 2 decay class 3 -0.034 0.001 0.003 <0.001 0.013 0.001 0.001 <0.001 -2.57 1.49 2.80 0.16 0.010 0.135 0.005 0.812 -0.061 <-0.001 0.001 <-0.001 -0.008 0.002 0.004 <0.001 Variable Adult tailed frogs were significantly associated with non-vascular and shrubby vegetation found in the wetter site types and with coarse woody debris in decay classes > 2 (p < 0.05; Table 14). The distribution o f tailed frogs was negatively associated with decreasing canopy closure, increasing light, sites containing drier vegetation, and an increased distance from the stream; however, only drier site types were statistically significant (p < 0.05). Furthermore, the nonlinear temperature term (linear and quadratic) was positively associated with tailed frog presences in the model using the pooled data up to 11°C after which the probability o f a tailed frog using a location decreased (Appendix VI). Females used locations with temperatures o f approximately 11°C and demonstrated a strong negative association with increasing light and a drier moisture regime, while having a greater association with wetter site types (p < 0.05). Males used locations with temperatures o f approximately 12°C, were negatively associated with an increasing distance from the stream edge, and positively associated with coarse woody debris, principally decay class 2 (p <0.05; Table 14). D is c u s s io n Currently, there is little information describing the movement patterns and habitat requirements o f tailed frogs in the terrestrial environment. Previous studies have used pitfall trapping and visual encounter surveys (Maxcy 2000; Wahbe et al. 2004; Matsuda and Richardson 2005; Hayes et al. 2006; Burkholder and Diller 2007). These techniques have limited utility for describing the fine-scale movements, patterns o f resource use or distribution o f the species as the recapture rate o f individual tailed frogs can be low (Wahbe et al. 2004; Matsuda and Richardson 2005; McEwan Unpub. Data). A better understanding 84 o f the terrestrial habitat requirements o f the tailed frog is essential if conservation and management activities are to be effective. Researchers have quantified the average movement rates for populations o f tailed frogs in southern BC within seasons (e.g., 3.0 m/day in the fall; Wahbe et al. 2004) and relative to habitat alterations (e.g., average daily movement rate of 12.27 m, SE = 3.48 in forested stands compared to 8.53 m, SE = 5.01 in buffered stands; Maxcy 2000). The average minimum distance traveled within fixed pitfall grids in southern BC populations has been estimated as 42.86 m (SE = 6.96; Maxcy 2000). The patterns o f movement within my study are representative o f daytime activity o f a nocturnal species. However, this study increases our knowledge on the movement between potential daytime refugia; particularly when climatic conditions are not favorable (increased temperature and decreased moisture). This population in northwestern BC, had a mean daily distance traveled o f 5.49 m/day (SE = 0.53), with some individuals traveling up to 45 m between daily relocations. Gender differences in movement Understanding movement rates reflective o f the biological demands o f each gender can assist with the conservation o f the species. For example, Hayes et al. (2006) suggested that female tailed frogs undergo movements to productive headwaters after ovipositioning, likely to acquire food resources for future yolk production. In my study, 9 females had a total distance traveled > 40 m and had greater estimated space use than males. However, the distances that females traveled were highly variable suggesting a considerable range o f seasonal ecological (reproduction) or physiological (thermal- and hydric-regulatory) demands. Conversely, 5 males had a total distance traveled < 40 m implying a greater 85 philopatry to particular areas, although the mechanisms for this behavior are unclear. Similar patterns in movement by male and female A. truei were observed by Burkholder and Diller (2007) in northwestern California, where females underwent greater parallel movements within the stream compared to both adult and immature males. Forest retention treatment and movement o f tailed frogs Previous research has suggested that timber harvest is a barrier to amphibian movement restricting dispersal and reducing migration (Johnston and Frid 2002; Matsuda and Richardson 2005; Baldwin et al. 2006; Hawkes and Gregory 2012). Forest harvesting reduces or eliminates canopy overstory, altering the microclimate at the forest floor, including the availability o f moist microsites that can act as thermal- and hydric-refugia (Com and Bury 1991; deMaynadier and Hunter 1999; Brosofske et al. 1997; Johnston 1998). Tailed frogs are known to be sensitive to local environmental changes because o f their ectothermic nature and fidelity to breeding locations (Daugherty and Sheldon 1982; Nussbaum et al. 1983; Wahbe et al. 2004; Matsuda and Richardson 2005; McEwan Pers. Obs.). The movement and habitat selection o f female frogs reflected the interaction between the ecology and physiology o f the species with the history o f forest harvesting at each site. For example, females in the old growth traveled further from the stream suggesting that the overstory and on-ground structural complexity created appropriate habitat conditions relative to the thermo- and hydro-regulatory requirements o f the species. Likewise, as overstory was reduced, the distance traveled from the stream decreased and movement became more parallel with the stream (Figure 15). These differences in movement suggest that physical constraints (increased temperature and decreased moisture) limited perpendicular movement. 86 Thus, the riparian area close to the stream and moist microsites within a site may act as habitat refugia facilitating upstream/downstream movement and ameliorate the dry conditions found across the broader clearcut matrix. Temporal distribution Previous research has proposed that A truei undergo migration to facilitate reproduction or to accommodate changing environmental conditions (Landreth and Ferguson 1967; Brown 1975; Wahbe et al. 2004; Hayes et al. 2006). Wahbe et al. (2004) suggested that A. truei males in southwestern BC traveled upstream to locate mature females during the breeding season and gravid females moved towards the stream for ovipositioning; although, there is little direct evidence relating movement to reproduction. A. montanus is known to move downstream during the fall in response to decreasing water temperatures (Adams and Frissell 2001). Pre-oviposited females were located moving towards the stream (Chapter 2). After eggs were deposited under rocks or debris within the larval streams (Karracker et al. 2006), females moved away from the larval stream (Chapter 2). During this time, females had the greatest daily movements (8.32 m, SE = 1.34). With the onset o f cooler temperatures and increased precipitation in the fall, females were seen in old stream channels containing moving water where breeding was occurring (McEwan Pers. Obs.). During this time, daily movements decreased (3.73 m SE = 1.27). These movement patterns suggest that female tailed frogs may express migratory behaviour in response to annual or biennial reproductive events. Conversely, I did not observe large movements by male tailed frogs. Individuals remained close to the origin o f capture with the lowest rate of movement occurring during 87 the fall (0.16 m SE = 0.08). Reduced movement within the breeding season has perhaps evolved with the inability o f the species to vocally communicate (Schmidt 1970) or is a function o f spatially discrete breeding locations (Todd et al. Unpub Data). Additionally, tailed frogs fitted with transmitters became subterranean in the late fall and remained there until the transmitter signal stopped. This may suggest that over-wintering occurs in the same vicinity as the breeding locations. Statistical models identifying habitat use by tailedfrogs Understanding habitat use at multiple scales is essential if we are to consider the full range o f habitat requirements for species o f concern like the tailed frog. Distributional choices an individual makes at a broad scale can influence or interact with behaviours, including habitat use, at a finer level (Rettie and Messier 2000). For example, the thermoand hydro-regulatory demands o f amphibians suggests that the use o f habitat at the coarse scale, such as forest retention treatment, is mechanistically linked to finer-scale use o f individual habitat features including wet micro-sites or CWD (Blomquist and Hunter 2010). With the miniaturization o f radio-telemetry devices, the application o f species distribution models to data for small (< 10 g) herpetofauna has become more common. For example, Faccio (2003) used radio-telemetry to quantify the habitat and area used by Ambystoma salamanders (A. maculatum and A. jejfersoniaum) near breeding ponds. Blomquist and Hunter (2010) used radio-telemetry data for wood frogs to study fine-scale habitat selection across 4 timber forest retention treatments in Maine. My study, however, is the first to use radio-telemetry to quantify resource selection and movement o f Ascaphus spp. Also, I considered resource selection relative to habitat features and site conditions that were measured at two spatial scales across a 4-month growing season. 88 Previous work suggests that CWD is an important structural component for maintaining amphibian diversity in forested stands o f the Pacific Northwest (reviewed by deMaynadier and Hunter 1995). This forest attribute provides shelter, protection from predation and ameliorates the thermo- and hydro-regulatory limits o f these taxa (Com and Bury 1991; Petranka et al. 1994; deMaynadier and Hunter 1995; Butts and McComb 2000; Bull 2002; Faccio 2003; Kluber et al. 2008; Blomquist and Hunter 2010). Older decay classes are typical in mature stands (Spies and Cline 1988) and are known to have greater water holding capabilities (Jager 1980). In addition, downed wood contains high proportions o f invertebrates (Harmon et al. 1986; Lockaby et al. 2002) and amphibians forage in and around CWD (Loebl999; Whiles and Grubaugh 1996). The RSF models identified a significant association between the location o f tailed frogs and CWD in older decay classes (i.e., decay classes 2 and 3; Table 14). This relationship was strongest for male frogs. Male frogs also demonstrated greater site fidelity and limited dispersal, as indicated by decreased movement (< 40 m) from the origin o f capture. Coarse woody debris would provide shelter and a microclimate required by the tailed frog (low temperature and increased moisture) thereby reducing movement. Although CWD can increase at a site following forest harvesting (Harmon et al. 1996), Spies and Cline (1988) suggest that this newly added CWD is often not decayed and smaller in diameter with less volume, which reduces its ability to hold water. A relatively small sample o f monitored tailed frogs within the clearcut treatments limits our understanding o f the resources used within those stands. However, the positive association with CWD in upper decay classes suggests that this habitat feature is a valuable resource for tailed frogs. 89 The RSF models identified a positive association between the locations o f tailed frogs and wetter site types, indicated by lady fern, devil’s club and leafy mosses. These wet site types would meet the physical and ecological demands o f tailed frogs and facilitate movement from stream edge, especially in sites where timber harvest has resulted in increased temperatures and decreased moisture. For example, females relocated in the clearcut, had a greater proportion (55.5%) o f their relocations in either subhygric or hygric site types (Table 8). When compared to random locations, male tailed frogs had greater association with locations near the stream (< 40 m). This may have been reflective o f their origin o f capture and limited movement (36.85 m, SE = 7.87) during the monitoring period. However, populations o f A. truei in southern (Wahbe et al. 2004; Matsuda and Richardson 2005) and northwestern BC (Chapter 2) also used habitats adjacent to the maternal stream. Tailed frogs are more prone to desiccation than other anurans (Claussen 1973a), thus, habitats adjacent to or within riparian areas may provide wet and cool microclimates (Brosofske et al. 1997; Chen etal. 1999). Temperature is critically important for the timing o f emergence, breeding, embryological development and growth o f amphibians (Brattstrom 1963). Previous works have demonstrated that Ascaphus have ambient air temperature thresholds between 2427.6°C. For example, Brown (1975) reported that A. montanus has a maximum temperature tolerance o f 24°C; Claussen (1973a) reported that populations o f A. truei along their southern range have critical thermal maxima o f 27.6°C; and A. truei in northwestern BC were not found in traps when ambient air temperatures were > 26.5°C (McEwan Unpub. Data). 90 However, many studies use air temperatures that reflect the macro-environment not temperatures associated with the micro-habitat selected by frogs. Removal o f forest cover at > 17m from the stream can result in temperature increases o f 2-4°C and a decrease in relative humidity o f 2.5-13.8% (Chen et al. 1999). Within this study, the most parsimonious model suggested tailed frogs used locations with temperatures between 11-12°C (Appendix VI); however, tailed frogs were found at locations with temperatures up to 24°C. Furthermore, the 2 females relocated in the clearcut treatments were on average monitored close to the stream (2.15 m, SE = 0.55) and used locations with understory vegetation representative o f wetter site types (Table 9) that would provide thermal and hydric refugia. Removal o f overstory increases the penetration o f light to the forest floor thereby altering the microclimate and vegetative structure (Greenberg 2001). Even with partial canopy removal, an increase in light can adversely affect amphibian abundance (deMaynadier and Hunter 1995, 1999). For example, Knapp et al. (2003) reported that > 41% o f canopy removal resulted in declines o f populations o f plethodontid salamanders in the southern Appalachians. Likewise, Baldwin et al. (2006) reported that during post­ breeding migration wood frogs selected locations that were dark, moist and had closed canopies. The RSF model for the female data identified a negative association with increasing light levels (p < 0.01). Similarly, I observed female tailed frogs using microhabitats with low levels of light: only 9.2% o f used locations had light levels > 103 lux, comparable to an overcast day (Appendix VII). Use o f habitats with lower light may be an artifact o f the species’ nocturnal behaviour (Metter 1967; Bury 1970), especially as monitored frogs were relocated during the day. However, the use o f locations with lower 91 ambient light could also indicate a relationship between canopy closure and its influence on the microclimate and the use o f certain forest attributes like CWD. Both variables were included in the top RSF models (Table 14). C o n c l u s io n I tested 3 main predictions based on existing knowledge o f forest-dependent amphibians like the tailed frog. Previous studies have reported the positive association o f amphibians with forest features such as CWD that ameliorate site conditions by providing shelter and increasing the availability o f food resources. Within the scope o f my study, adult tailed frogs demonstrated a positive association with CWD o f the oldest decay classes. Second, I predicted that during the warmer and drier portions of the year, tailed frogs would be located close to the stream and with habitat features that reduced thermal stress. Although tailed frogs were located closer to the stream in the summer compared to the spring, the greatest adjacency occurred during the fall. This suggests that the breeding phenology o f the species is a strong driver o f seasonal distribution. Additionally, tailed frogs were associated with locations with temperatures o f 11°C as seen in the RSF model. Lastly, I predicted that as overstory declined A. truei would have a greater association with habitat features that ameliorate dry or warm stand conditions. Adult tailed frogs had a significant association with wet microsites and avoided drier types; this was especially true in the clearcut treatments. R e c o m m e n d a t io n s Forest management and the tailed fro g Forest management practices can directly affect the integrity o f habitat required by the tailed frog at a number o f ecological and evolutionary scales. Our understanding o f the 92 ecology o f semi-aquatic species, like the tailed frog, is often focused on the larval stage or near-stream locations o f the post-metamorphs. This is in part reflective o f the species’ cryptic and fossorial nature. Thus, we have a limited understanding o f the use and importance o f upslope habitats where tailed frogs spend the majority o f their life. Documenting the species’ movement and distribution at multiple scales, including habitat use across fragmented landscapes, is necessary for maintaining important forest features, stand conditions, and habitat connectivity that facilitates long-term gene flow among populations. Tailed frogs in northwestern BC demonstrated a positive association with moist or wet microhabitats, as indicated by vegetation type and CWD in the older decay classes. These habitat features likely ameliorate extremes in temperature allowing tailed frogs to meet thermal- and hydro-regulatory limits. Additionally, tailed frogs used locations with cooler temperatures (1 1-12°C) that were well below the thermal tolerance o f the species (24°C; Brown 1975). However, the structure o f the best RSF model did vary when considering locations separately for male and female tailed frogs. Females used locations with low light levels (< 103 lux) and avoided drier site types, while males remained in locations close to a stream (< 40 m) with moderately decomposed CWD (decay class 2). Furthermore, assessments of the predictive capacity o f each set of models suggested that females were more selective in their choice o f habitat (i.e., greater AUC). These results confirm past work that has related the ecological requirements o f tailed frogs to thermal and hydric limits (Metter 1967; Brown 1975; Daughtery and Sheldon 1982) and recommended that land management practices retain intact overstory and complex forest structure (Wahbe et al. 2004; Matsuda and Richardson 2005). For example, Spears and Strofer (2008) found that certain landscape features, such as forest cover, had a strong 93 influence on the population structure o f tailed frogs. A reduction in the overstory canopy creates fragmented habitats ultimately limiting dispersal and gene flow among and potentially within populations. Hawkes and Gregory (2012) documented a time lag in the response by tailed frogs to overstory removal such that the number o f individuals decreased 10-years post-harvest compared to pre-harvest numbers. Previous studies reported a positive association between the distribution o f tailed frogs and specific forest attributes including CWD, decreased light intensity, and moist microsites that ameliorate site conditions after a disturbance. For example, Butts and McComb (2000) suggested that the retention o f CWD in managed stands closely mimic those volumes found in natural stands (248 m3/ha) in the western Cascades o f Washington. Baldwin et al. (2006), Rittenhouse and Semlitsch (2008), and Blomquist and Hunter (2010) all speak to the importance o f preserving forest attributes, including CWD and moist microsites that create thermal- and hydric-refugia. These refugia can facilitate movement across landscapes fragmented by anthropogenic disturbances such as timber harvest. Consistent with past studies, my findings suggest that conservation planners should consider fine-scale habitat features when designing reserves (e.g., Wildlife Habitat Areas) and managing forest harvest to conserve populations o f the tailed frog. First, the most effective, but costly strategy is to maintain old forest types that provide suitable climatic and structural conditions required by tailed frogs (i.e., moist microsites, CWD in decay classes > 2 and temperatures between 11-12°C). These habitat conditions are essential for maintaining individual populations o f tailed frog and will facilitate movement by the reproductive core o f the population (i.e., females). Second, across managed forests, strategies should be designed to preserve important microsites and habitat features, such as wet drainages and CWD that 94 provide thermo- and hydro-regulatory refugia. Third, where harvesting occurs, a wind-firm forest retention buffer should be maintained around maternal streams. This buffer should be o f sufficient size to ameliorate changes in the understory microclimate resulting from adjacent forest harvesting. These buffers will maintain essential reproductive and rearing habitats and facilitate seasonal migration and population connectivity (Spears and Strofer 2008). By maintaining functional habitat for the tailed frog across forest patches and landscapes, we can conserve this species at their northern range limit, including the maintenance o f evolutionary processes. 95 Chapter 4 G eneral Sum m ary There have been relatively few studies on the biology or ecology o f the postmetamorphic coastal tailed frog, especially at the northern distribution o f the species. Hence, the majority of our knowledge is attributed to interior and southern populations o f Ascaphus with a particular focus on the larval life stage (Daugherty and Sheldon 1982; Bury and Adams 1999; Dupuis and Steventon 1999; Maxcy 2000; Wahbe and Bunnell 2001; Wahbe et al. 2004; Matsuda and Richardson 2005; Hayes et al. 2006; Karracker et al. 2006; Burkholder and Diller 2007). I used an Information Theoretic Model Comparison approach to quantify the distribution, movement patterns, and resource selection o f post-metamorphic coastal tailed frog (A. truei). For A. truei in northwestern BC, this study was the first to: 1) describe the movement and fine-scale distribution o f the post-metamorphic tailed frog in response to ecological, biological and disturbance factors; and 2) create resource selection functions identifying the fine-scale habitat requirements o f the species. Although this study is relevant to populations o f A. truei in northwestern BC, the use o f these techniques can be applied to populations o f Ascaphus throughout their range. A better understanding o f the distribution, movements, and resource selection o f this species can assist with conservation actions and the mitigation o f impacts from anthropogenic disturbances including timber harvest, independent power projects and linear features such as oil and gas pipelines. In Chapter 2 , 1 used 2 consecutive years o f capture data to quantify the movement patterns and relative abundance o f post-metamorphic tailed frogs across terrestrial habitats within two watersheds. The fine-scale distribution o f tailed frogs was systematically monitored using pitfall traps installed at 4 forest retention treatments. Emergence from over­ wintering sites and movement by females for ovipositioning appeared to correspond with warmer temperatures and melting o f the snow pack. Furthermore, direction o f movement 97 was specific to time o f year and the reproductive status or gender o f the tailed frog. Juveniles are known dispersers (Daughtery and Sheldon 1982) and were documented moving away from the stream in the old growth and clearcut treatments, females moved away from the stream after ovipositioning, and newly metamorphosed frogs were located moving upstream after emergence. The majority o f captures were in the 5 m arrays, associated with rain events (< 1 day) and during the months associated with the breeding phenology o f the species. Furthermore, canopy cover influenced movement and capture rate as the number o f frogs caught per unit effort decreased as the overstory canopy was reduced following forest treatment. This was particularly evident with the large sample size from the 2012 trap year. These results are consistent with past research in other regions that noted a relationship between the distribution and activity o f tailed frogs relative to micro-site and broader climatic conditions as influenced by forest overstory and structural complexity (Daughtery and Sheldon 1982; Aubry 2000; Matsuda and Richardson 2005). Temporal variation in capture success suggests that the pitfall trap method and resulting data are influenced by season, likely mediated by climate trends, and the reproductive biology and resulting behaviour o f tailed frogs. Thus, the timing and extent of trap sessions can greatly influence capture results and the observed movements o f frogs, at least at the northern extent o f their range. These findings are useful when designing inventory protocols for A. truei, specifically: 1) the time o f year to conduct surveys designed to monitor fine-scale distribution and detect the presence o f post-metamorphic tailed frogs within a stream reach or watershed; and, 2) the proximity o f searches to a stream (< 50 m) either for the placement o f pitfall traps or areas to focus visual encounter surveys. 98 In Chapter 3 , 1 used radio-telemetry data to quantify the resource selection o f adult A. truei across forest retention treatments. These results allowed me to identify the features and the micro-scale attributes o f habitat likely important to the thermo- and hydro-regulatory processes o f A. truei. In general, adult tailed frogs used locations that afforded the greatest protection from desiccation (increased moisture) and were characterized by decayed coarse woody debris, low light, and cool temperatures (11-12°C). However, movement patterns and habitat use differed between female and male tailed frogs. Females moved greater distances and were more sensitive to temperature and moisture constraints. In addition, females made large-scale movements associated with the reproductive phenology o f the species. Conversely, male frogs demonstrated greater philopatry to a location (< 40 m from origin o f capture), remained close to the stream (< 40 m) and increased their daily movement during the summer compared to the spring (1.4 times greater). Such movements were likely associated with foraging or to access habitat that provided protection from physical constraints (increased temperature and decreased moisture). Through this study I documented or inferred seasonal movement patterns at a range o f ecological scales and I identified the habitat requirements, including micro- and macro­ scale features that were related to the ecological and physical requirements o f A. truei found at the northern extent o f its range. For instance, female movements coincided with reproductive needs such as ovipositioning (spring) and breeding (fall). Additionally, tailed frogs were seen using breeding locations annually and these sites have the potential to be important over-wintering habitat. Furthermore, the observed movement patterns and finescale distribution revealed the dependency o f this species to the near-stream environment, especially when canopy cover was reduced creating warmer and drier stand conditions. 99 Finally, the biological and physical requirements o f this species, in addition to observed habitat use and fine-scale distribution, suggested that the most appropriate forest types had mature standing timber with complex structure in the understory, specifically cool, moist microsites and CWD in decay classes > 3. Conservation Recommendations Findings from this study, in combination with conservation theory and current understanding o f A. truei, can be used by land managers to identify appropriate habitat reserves (e.g., Wildlife Habitat Areas) and may assist with the design o f movement corridors across landscapes. Such a conservation network would be composed o f riparian reserve zones, riparian management zones and linkages to adjacent watersheds. O f greatest priority, habitat should be conserved near the larval stream. These areas support all life stages o f A. truei and are essential for reproduction and ultimately population persistence and dispersal. To protect the intact overstory and structural complexity o f the understory near these areas, a Riparian Reserve Zone should be placed on both sides of the larval stream (Chen et al. 1999), specifically along stream reaches where ovipositioning occurs. In addition to being the source o f juvenile frogs, these areas have high densities o f adult frogs during periods o f the year associated with reproduction. Data from this study revealed that post-metamorphic tailed frogs are found at distances beyond the 30-50 m buffers currently proposed for habitat reserves (BCMWLAP 2004). However, to fully quantify the buffer width in riparian areas, more work is required to assess long-term movement. Such a study would require multiple successive transmitters placed on individuals allowing for the measurement o f movement dynamics that exceed the shorter period o f monitoring (mean = 10.75 days, SE = 0.89) applied to this research. Such 300 data would better quantify the distance traveled by females seasonally thereby increasing the certainty o f dispersal capabilities o f the species. Consistent with theory, past work, and the current strategy o f habitat reserves for tailed frogs in BC, I recommend a Riparian Management Zone adjacent to the Riparian Reserve Zone. Forest harvesting would be permitted in this zone, but there would be an emphasis on retaining ground-level forest structure and moist microsites. These habitat features would ensure physical and foraging refugia with temperatures reflective o f locations used by tailed frogs (1 1-12°C) in this study. Similar to past research on amphibians, the tailed frog demonstrated a positive association with CWD specifically in decay classes 3-5 (BC MELP 1998); however, the minimum quantity that should be retained within the Riparian Management Zone requires further investigation. Riparian Management Zones should follow the current stand retention and best management practices for S5 (non-fish bearing) streams located in valley bottoms: 1) the retention o f 50% o f the dominant and co-dominant overstory that would reduce the risk of windthrow; 2) retention of non-merchantable timber and herbaceous vegetation within 10 m of the stream channel, 3) ensure that falling and yarding occur away from the stream and remove any slash or debris that enters the stream; and, 4) retention o f wildlife trees (BC Ministry o f Forests and Ministry o f Environment 1995). Finally, movement corridors or linkage zones should be created between riparian reserves allowing for genetic dispersal (Beier et al. 2008). In southerly populations o f A. truei, Spear and Strofer (2008) demonstrated genetic similarities among watersheds up to 25-30 km apart. Thus, a study o f genetic connectivity for northern populations o f the tailed frog will increase our 101 understanding o f the importance and proper placement o f movement corridors (Murphy et al. 2010 ). If the aforementioned recommendations are followed, then A. truei may persist across a landscape consisting o f a patchwork of timber harvest and intact forests connected by movement corridors. Moreover, by creating useable habitat for the tailed frog that retains mature stand attributes and intact corridors between watersheds, other old-forest flora and fauna will benefit. Likewise, retention o f intact headwaters that are non-fish bearing and contain A. truei also will protect the downstream waters o f fish-bearing streams by decreasing sedimentation and temperature fluctuations (BC Ministry o f Environment 2004). 102 L i t e r a t u r e C it e d Adams, S.B. and C.A. Frissell. 2001. Thermal habitat use and evidence o f seasonal migration by Rocky Mountain tailed frogs, Ascaphus montanus, in Montana. Canadian Field-Naturalist 115: 251-256. Araujo M.B., P.J. Densham and P.H. Williams. 2004. Representing species in reserves from patterns o f assemblage diversity. Journal o f Biogeography 31: 1037-1050. Ascaphus Consulting. 2003. Watershed-level protection and management measures for the maintenance o f Ascaphus truei populations in the Skeena region. Report to: Leonard Vanderstar, Habitat Protection Biologist MWLAP, Skeena Region, Smithers, BC. Aubry, K.B. and P.A. Hall. 1991. Terrestrial amphibian communities in the Southern Washington Cascade Range. In Wildlife and vegetation o f unmanaged Douglas-fir forests. L. F. Ruggiero, K. B. Aubry, A. B. Carey, and M. H. H uff (editors). U.S. Department of Agriculture and Forest Service, PNW-GTR-285, pp. 326-338. Aubry, K.B. 2000. Amphibians in managed, second-growth Douglas fir forests. The Journal o f Wildlife Management 64: 1041-1052. Baldwin, R.F., A.J.K. Calhoun and P.G. deMaynadier. 2006. Conservation Planning for Amphibian Species with Complex Habitat Requirements: A Case Study Using Movements and Habitat Selection o f the Wood Frog Rana sylvatica. Journal o f Herpetology 40: 443-454. Bartelt, P.E. and C.R. Peterson. 2000. A description and evaluation o f a plastic belt for attaching radio transmitters to western toads (Bufo boreas). Northwestern Naturalist 81: 122-128. BC Conservation Data Centre (CDC). 2012. BC species and ecosystems explorer. British Columbia Ministry o f Environment Victoria, BC. (accessed January 24, 2012). BC Environment. 2002. Environmental trends in British Columbia 2002. British Columbia Ministry o f Environment, Lands, Parks, Victoria, BC. BC Ministry o f Environment, Lands and Parks (BC MELP). 1998. Field manual for describing terrestrial ecosystems. Land Manager Handbook 25 ISSN-0229-1622. BC Ministry o f Forests and Ministry o f Environment. 1995. Riparian management area guidebook. Forest Practices Code o f British Columbia. Ministry o f Forests and Ministry o f Environment, Lands and Parks, Victoria, British Columbia p.68. BC Ministry o f Forests and Range and BC Ministry o f Environment (BCMFR and BCMOE). 2010. Field manual for describing terrestrial ecosystems. Second edition. Forest Science Program, Victoria, BC. Land Management Handbook No. 25. p.266. 103 BC River Forecast Centre. 2012. Snow survey and water supply bulletin —June 1st 2012. http://bcrfc.env.gov.bc.ca/bulletins/watersupDlv/archive/2012/2012 Junl SnowBulletin. pdf. BC River Forecast Centre. 2013. Sow survey and water supply bulletin- June 1st 2013. http://bcrfc.env.gov.bc.ca/bulletins/watersupply/archive/2013/2013_Jun 1.pdf. BC Ministry o f Water, Land and Air Protection (BCMWLAP). 2002. Indicators o f climate change for British Columbia, http://www.env.gov.bc.ca/cas/pdfs/indcc.pdf. B.C. Ministry o f Water, Land and Air Protection (BCMWLAP). 2004. Coastal Tailed Frog (Ascaphus truei). In Accounts and Measures for Managing Identified Wildlife Accounts V. 2004. BC Ministry o f Water, Land and Air Protection, Victoria, BC. http://www.env.gov.bc.ca/wld/ffpa/iwms/documents/Amphibians/a_coastaltailedfrog.pdf Blomquist, S.M. and M.L. Hunter Jr. 2010. A multi-scale assessment of amphibian habitat selection: wood frog response to timber harvesting. Ecoscience 17: 251-264. Boyce, M.S., P.R. Vernier, S.E. Neilsen and F.K.A. Schmiegelow. 2002. Evaluating resource selection functions. Ecological Modeling 157: 281-300. Brosofske, K.D., J. Chen, R.J. Naiman and J.F. Franklin. 1997. Harvesting effects on microclimatic gradients from small streams to uplands in western Washington. Ecological Applications 7: 1188-1200. Brown, H.A. 1975. Temperature and development o f the tailed frog, Ascaphus truei. Comparative Biochemical Physiology 50: 397-405. Burkholder, L.L. and L.V. Diller. 2007. Life history o f post metamorphic coastal tailed frogs (Ascaphus truei) in northwestern California. Journal o f Herpetology 41: 251-262. Bull, E.L. and B.E. Carter. 1996. Tailed frogs: distribution, ecology, and association with timber harvest in northeastern Oregon. Research Paper PNW-RP-497. Portland, OR: United States Department o f Agriculture, Forest Service, Pacific Northwest Research Station. Bull, E.L. 2002. The value o f coarse woody debris to vertebrates in the Pacific Northwest. USDA Forest Service General Technical Report, PSW-GTR-181. 171-178. Bury, R.B. and M.J. Adams. 1999. Variation in age at metamorphosis across a latitudinal gradient for the tailed frog, Ascaphus truei. Herpetologica 55: 283-291. Butts, S.R. and W.C. McComb. 2000. Associations o f forest-floor vertebrates with coarse woody debris in managed forests o f western Oregon. The Journal o f Wildlife Management 64: 95-104. 104 Cabeza, M, M.B. Araujo, R.J. Wilson, C.D. Thomas, M.J.R. Cowley and A. Moilanen. 2004. Combining probabilities o f occurrence with spatial reserve design. Journal o f Applied Ecology 41: 252-262. Chen, J., Saunders, S.C., Crow, T.R., Naiman, R.J., Brosofske, K.D., Mroz, G.D., Brookshire, B.L., and Franklin, J.F. 1999. Microclimate in forest ecosystem and landscape ecology. BioScience 49: 288-297. Claussen, D.L. 1973a. The thermal relations o f the tailed frog, Ascaphus truei and the Pacific treefrog, Hyla regilla. Comparative Biochemistry and Physiology 44: 137-153. Claussen, D.L. 1973b. The water relations o f the tailed frog, Ascaphus truei and the Pacific treefrog, Hyla regilla. Comparative Biochemistry and Physiology 44: 155-171. Com, P.S. and R.B. Bury. 1991. Terrestrial amphibian communities in the Oregon Coast Range. In Ruggiero, L. F., K. B. Aubry, A. B. Carey, M. H. Huff, technical coordinators. Wildlife and vegetation o f unmanaged Douglas fir forests. Portland, Oregon: USDA Forest Service Pacific Northwest Research Station. General Technical Report PNW-GTR 285: 305-317. Cree, A. 1989. Relationship between environmental conditions and nocturnal activity o f the terrestrial frog, Leiopelma archeyi. Journal o f Herpetology 23: 61-68. Cushman, S.A. 2006. Effects o f habitat loss and fragmentation on amphibians: a review and prospectus. Biological Conservation 128:231-240. Daugherty, C.H. and A.L. Sheldon. 1982. Age-specific movement patterns o f the frog Ascaphus truei. Herpetologica 38: 468-474. Deguise, I. and J.S. Richardson. 2009. Movement behaviour o f adult western toads in a fragmented, forest landscape. Canadian Journal o f Zoology 87: 1184-1194. deMaynadier, P.G. and M.L. Hunter. 1995. The relationship between forest management and amphibian ecology: a review o f the North American literature. Environmental Review 3: 230-261. deMaynadier, P.G. and M.L. Hunter. 1998. Effects o f silvicultural edges on the distribution and abundance o f amphibians in Maine. Conservation Biology 12: 340-352. deMaynadier, P.G. and M.L. Hunter. 1999. Forest canopy closure and juvenile emigration by pool-breeding amphibians in Maine. Journal o f Wildlife Management 63: 441-450. Dodd, C.K. Jr., 2009. Amphibian Ecology and Conservation. A Handbook o f Techniques. Oxford University Press Inc, New York. 105 Dupuis, L.A. and F.L. Bunnell. 1997. Status and distribution of the tailed frog in British Columbia. Final report for Forest Renewal BC. University o f BC, Centre o f Applied Conservation Biology, Vancouver, BC. Dupuis, L. and D. Steventon. 1999. Riparian management and the tailed frog in northern coastal forests. Forest Ecology and Management 124: 35-43. Dupuis, L., F.L. Bunnell and P.A. Friele. 2000. Determinants o f the tailed frog’s range in British Columbia, Canada. Northwest Science 74: 109-115. Dupuis, L.A. and F.L. Waterhouse. 2001. Response o f amphibians to partial cutting in a coastal mixed-conifer forest: management practices for retaining amphibian habitat in the Vancouver forest region. Forest Research Extension Note. Vancouver Forest Region March EN-005. Farhig, L., J.H. Pedlar, S.E. Pope, P.D. Taylor and J.E. Wegner. 1995. Effect o f road traffic on amphibian density. Biological Conservation 73: 177-182. Faccio, S.D. 2003. Postbreeding emigration and habitat use by Jefferson and spotted salamanders in Vermont. Journal o f Herpetology 37: 479-489. Fenolio, D.B., G.O. Graening and J.F. Stout. 2005. Seasonal movement patterns o f the pickerel frog (Ranapalustris) in an Ozark cave and trophic implications supported by stable isotope evidence. The Southwestern Naturalist 50: 385-389. Forsythe, P., B. Flitz and S.J. Mullin. 2004. Radio telemetry and post-emergent habitat selection of neonate box turtles (Emydidae: Terrapene Carolina ) in central Illinois. Herpetological Review 35: 333-335. Gibbs, J.P. 1998. Amphibian movement in response to forest edges, roads, and streambeds in southern New England. Journal o f Wildlife Management 62: 584-589. Green, D.M. and R.W. Campbell. 1984. The amphibians o f British Columbia. Royal British Columbia Museum, Victoria, BC Handbook 45. Greenberg, C.H. 2001. Response o f reptile and amphibian communities to canopy gaps created by wind disturbance in the southern Appalachians. Forest Ecology and Management 148: 135-144. Griffiths, R.A. and M. Dos Santos. 2012. Trends in conservation biology: Progress or procrastination in a new millennium? Biological Conservation 153: 153-158. Hamann, A. and T. Wang. 2006. Potential effects o f climate change on ecosystem and tree species distribution in British Columbia. Ecology 87: 2773-2786. Harmon, M.E., J.F. Franklin, F J . Swanson, P. Sollins, S.V. Gregory, J.D. Lattin, N.H. 106 Anderson, S.P. Cline, N.G. Aumen, J.R. Sedell, G.W. Lienkaemper, K. Cromack Jr., and K.W. Cummins. 1986. Ecology o f coarse woody debris in temperate ecosystems. Advance Ecological Research 15: 133-302. Harmon, M.E., S.L. Garman and W.K. Ferrel. 1996. Modeling historical patterns of tree utilization in the Pacific Northwest: carbon sequestration implications. Ecological Applications 6: 641-652. Hawkes, V.C. and P.T. Gregory. 2012. Temporal changes in the relative abundance o f amphibians relative to riparian buffer width in western Washington, USA. Forest Ecology and Management 274: 67-80. Hawkins C.P., L.J. Gottschalk and S.S. Brown. 1988. Densities and habitat o f tailed frog tadpoles in small streams near Mt. St. Helens following the 1980 eruption. Journal o f North American Benthological Society 7: 246-252. Hayes, M.P., T. Quinn, D.J. Dugger, T.L. Hicks, M.A. Melchiors and D.E. Runde. 2006. Dispersion o f coastal tailed frog (Ascaphus truei): An hypothesis relating occurrence of frogs in non-fish-bearing headwater basins to their seasonal movements. Journal of Herpetology 40: 531-543. Houlahan, J.E., C.S. Findlay, B.R. Schmidt, A.H. Meyers and S.L. Kuzmin. 2000. Quantitative evidence for global amphibian declines. Nature 404: 752-755. Jaeger, R.G. 1980. Microhabitats o f a terrestrial salamander. Copeial 980: 265-268. Johnson, C.J., D.R. Seip and M.S. Boyce. 2004. A quantitative approach to conservation planning: using resource selection functions to map the distribution o f mountain caribou at multiple spatial scales. Journal o f Applied Ecology 41: 238-251. Johnson, C.J. and M.P. Gillingham. 2005. An evaluation o f mapped species distribution models used for conservation planning. Environmental Conservation 32: 1-12. Johnson, C.J., S.E. Nielsen, E.H. Merrill, T.L. McDonald and M.S. Boyce. 2006. Resource selection functions based on use-availability data: Theoretical motivation and evaluation methods. Journal o f Wildlife Management 70: 347-357. Johnston, B. 1998. Terrestrial Pacific giant salamanders (Dicamptodon tenebrosus Good): Natural history and their response to forestry practices. MSc. thesis. University o f British Columbia, Vancouver BC. Johnston, B. and L. Frid. 2002. Clear-cut logging restricts the movements o f terrestrial Pacific giant salamanders (Dicamptodon tenebrosus Good). Canadian Journal of Zoology 80:2170-2177. 107 Karracker, N. E. 2001. String theory: reducing mortality o f mammals in pitfall traps. Wildlife Society Bulletin 29: 1158-1162. Karracker, N.E., D.S. Pilliod, M.J. Adams, E.L. Bull, P.S. Com, L.V. Diller, L.A. Dupuis, M.P. Hayes, B.R. Hossack, G.R. Hodgson, E.J. Hyde, K. Lohman, B.R. Norman, L.M. Ollivier, C.A. Pearl and C.R. Peterson. 2006. Taxonomic variation in oviposition by tailed frogs (Ascaphus spp.). Northwestern Naturalist 87: 87-97. Karracker, N.E., and H.H. Welsh Jr. 2006. Long-term impacts o f even-aged timber management on abundance and body condition o f terrestrial amphibians in Northwestern California. Biological Conservation 131: 132-140. Ketcheson, M.V., T.F. Braumandl, D.V. Meidinger, G. Utzig, D.A. Demarchi and B.M. Wikeem. 1991. Chapter 11: Interior Cedar Hemlock Zone In Ecosystems o f British Columbia. D.V. Meidinger and J. Pojar (eds.) pp. 95-111. Kluber, M.R., D.H. Olson and K.J. Puettmann. 2008. Terrestrial salamander distributions in riparian and upslope areas and their habitat associations on managed forest landscapes of the Oregon Coast Range. Forest Ecology and Management 256: 529-535. Knapp, S.M., C.A. Hass, D.N. Harpole and R.L. Kirkpatrick. 2003. Initial effects o f clear cutting and alternative silvicultural practices on terrestrial salamander abundance. Conservation Biology 17: 752-762. Koper, N. and M. Manseau. 2010. A guide to developing resource selection functions from telemetry data using generalized estimating equations and generalized linear mixed models. Rangifer20: 195-204. Landreth H.F. and D.E. Ferguson. 1967. Movements and orientation o f the tailed frog, Ascaphus truei. Herpetologica 23: 81-93. Lockaby, B.G., B.D. Keeland, J.A. Stanturf, M.D. Rice, G. Hodges and R.M. Govemo. 2002. Arthropods in decomposing wood o f the Atchafalaya river basin. Southeastern Naturalist 1: 339-352. Loeb, S.C. 1999. Responses o f small mammals to coarse woody debris in a southeastern pine forest. Journal o f Mammalogy 80: 460-471. Manly, B.F.J., L.L. McDonald, D.L. Thomas, T.L. McDonald, and W.P. Erickson. 2002. Resource selection by animals: statistical analysis and design for field studies. Second Edition. Kluwer, Boston, Massachusetts, USA. Marsh, D.M. and P.C. Trenham. 2000. Metapopulation dynamics and amphibian conservation. Conservation Biology 51: 40-49. Matsuda, B.M. and J.S. Richardson. 2000. Clear-cut timber harvest and movement patterns 108 in tailed frogs. In Proceedings o f a Conference on the Biology and Management of Species and Habitats at Risk 2: 485-488. Matsuda, B.M. 2001. The effects o f clear-cut timber harvest on the movement patterns o f tailed frogs {Ascaphus truei) in southwestern British Columbia. MSc. Thesis. University o f British Columbia Vancouver BC. Matsuda, B.M. and J.S. Richardson. 2005. Movement patterns and relative abundance o f coastal tailed frogs in clearcuts and mature forests stands. Canadian Journal o f Forest Research 35: 1131-1138. Maxcy, K.M. 2000. The response o f terrestrial salamanders to forest harvesting in southwestern British Columbia. MSc. Thesis. University o f British Columbia Vancouver BC. Mazerolle, M.J. 2006. Improving data analysis in herpetology: Using Akaike’s Information Criterion (AIC) to assess the strength o f biological hypotheses. Amphibia-Reptilia 27: 169-180. McCaffery, R.M. and B.A. Maxell. 2010. Decreased winter severity increases viability of montane frog population. Proceedings o f the National Academy o f Sciences 107: 86448649. Meidinger, D. and J. Pojar (Eds.). 1991. Ecosystems o f British Columbia. Ministry of Forests, Victoria, British Columbia, Special Report Series 6: Chapter 6: 95-112. Metter, D.E. 1967. Variation in the ribbed frog Ascaphus truei Stejneger. Copeia 1967: 634-649. Muths, E. 2003. A radio transmitter belt for small ranid frogs. Herpetological Review 34: 345-348. Newman, D.G. 1990. Activity, dispersion and population densities o f Hamilton’s frog {Leiopelma hamiltoni) on Maud and Stephens Islands, New Zealand. Herpetologica 46: 319-330. Nielson, M., K. Lohman and J. Sullivan. 2001. Phylogeography o f the tailed frog {Ascaphus truei): Implications for the biogeography o f the Pacific Northwest. Evolution 55: 147160. Nielsen, S.E., M.S. Boyce, G.B. Stenhouse and R.H.M. Munro. 2001. Modeling grizzly bear habitats in the Yellowhead Ecosystem o f Alberta: Taking autocorrelation seriously. Ursus 13: 45-56. Nielsen, S.E., C.J. Johnson, D.C. Heard and M.S. Boyce. 2005. Can models o f presenceabsence be used to scale abundance? Two case studies considering extremes in life history. Ecography28: 197-208. 109 Nussbaum, R.A., E.D. Brodie Jr. and R.M. Storm. 1983. Amphibians and reptiles of the Pacific Northwest. University Press o f Idaho, Moscow. Nuzzo, V.A. and K.S. Mierzwa. 2000. The effect o f forest structure on amphibian abundance and diversity in the Chicago Region. Citizens for Conservation, U.S. Environmental Protection Agency, Great Lakes National Program Office, Lake County Forest Preserve District, Forest Preserve District o f Will County. Olson, D.H. and C. Rugger. 2007. Preliminary study o f the effects o f headwater riparian reserves with upland thinning on stream habitats and amphibians in western Oregon. Forest Science 53: 331-342. Olson D.H., P.D. Anderson, C.A. Frissell, H.H. Welsh Jr. and D.F. Bradford. 2007. Biodiversity management approaches for stream-riparian areas: Perspectives for Pacific Northwest headwater forests, microclimates, and amphibians. Forest Ecology and Management 246: 81-107. Palis, J.G. 1997. Breeding migration o f Ambystoma cingulatum in Florida. Journal of Herpetology 31: 71-78. Palis, J.G. 1998. Breeding biology of the gopher frog, Rana capito, in western Florida. Journal o f Herpetology 32: 217-223. Pearce, J. and S. Ferrier. 2000. Evaluating the predictive performance o f habitat models developed using logistic regression. Ecological Modeling 133: 225-245. Perkins, D.W. and M.L. Hunter. 2006. Use o f amphibians to define riparian zones of headwater streams. Canadian Journal o f Forest Research 36: 2124-2130. Petranka, J.W., M.P. Brannon, M.E. Hopey and C.K. Smith. 1994. Effects o f timber harvesting on low elevation populations o f southern Appalachian salamanders. Forest Ecology and Management 67: 135-147. Pojar, J., K. Klinka and D.A. Demarchi. 1991. Chapter 6: Coastal Western Hemlock Zone. In Ecosystems of British Columbia, y D.V. Meidinger and J. Pojar (eds.) pp. 95-111. Rappole, J.H., M.A. Ramos and K. Winker. 1989. Movements and mortality in Wood Thrushes. Auk 106: 402-410. Rettie, W.J. and F. Messier. 2000. Hierarchical habitat selection by woodland caribou: its relationship to limiting factors. Ecography 23: 466-478. Richardson, J.S. and W.E. Neill. 1998. Headwater amphibians and forestry in British Columbia: Pacific giant salamanders and tailed frogs. Northwest Science 72: 122-123. 110 Ritland, K., L.A. Dupuis, F.L. Bunnell, W.L.Y. Hung and J.E. Carlson. 2000. Phylogeography of the tailed frog {Ascaphus truei) in British Columbia. Canadian Journal of Zoology 78: 1749-1758. Rittenhouse, T.A.G. and R.D. Semlitsch. 2007. Post-breeding habitat use by wood frogs in Missouri oak-hickory forest. Journal o f Herpetology 41: 645-653. Rowley, J.J.L. and R.A. Alford. 2007. Techniques for tracking amphibians: the effects o f tag attachment, and harmonic direction finding versus radio telemetry. AmphibiaReptilia 28: 367-376. Schmidt, R.S. 1970. Auditory receptors in two mating call-less anurans. Copeia 1:169-170. Semlitsch, R.D. 2002. Critical elements for biologically based recovery plans o f aquaticbreeding amphibians. Conservation Biology 16: 619-629. Semlitsch, R.D. 2008. Differentiating migration and dispersal processes for pond-breeding amphibians. Journal o f Wildlife and Management 72: 260-267. Spear S.F. and A. Storfer. 2008. Landscape genetic structure o f coastal tailed frog {Ascaphus truei) in protected versus managed forest. Molecular Ecology 17: 46424652. Spies, T.A. and S.P. Cline. 1988. Coarse woody debris in forests and plantations o f coastal Oregon. In From the forest to the sea: a story o f fallen trees Maser, C., S.P. Cline, K. Cromack Jr, J.M. Trappe, and E. Hansen (eds.) United States Department o f Agriculture, Forest Service, General Technical Report PNW-229: 1-53. Stebbins, R.C. 2003. Western Reptiles and Amphibians. Boston New York: Houghton Mifflin Company, p 471. Stoddard, M.A. and J.P. Hayes. 2005. The influence o f forest management on headwater stream amphibians at multiple spatial scales. Ecological Applications 15: 811-823. StrahlerA.N. 1957. Quantitative analysis o f watershed geomorphology. Transactions o f the American Geophysical Union 38: 913:920. Swets, J.A. 1988. Measuring the accuracy o f diagnostic systems. Science 240: 1285-1293. Van Wagner, C.E. 1968. The line-intersect method in forest fuel sampling. Forest Science 14: 20-26. Wahbe, T.R. 1996. Tailed frogs {Ascaphus truei Stejneger) in natural and managed coastal temperate rainforests o f southwestern British Columbia, Canada. M.Sc. thesis. University o f British Columbia, Vancouver, BC. p.49. Ill Wahbe, T.R., F.L. Bunnell and R.B. Bury. 2000. Defining wildlife habitat areas for tailed frogs. In Proceedings o f the Conference on the biology and management o f species and habitats at risk. L. M. Darling (ed.) Kamloops, BC, Feb. 15-19, 1999. British Columbia Ministry of Environment, Lands and Parks, Victoria, BC and University College o f the Cariboo, Kamloops, BC 489-496. Wahbe, T.R. and F.L. Bunnell. 2001. Preliminary observations on movements o f tailed frog tadpoles (Ascaphus truei) in streams through harvested and natural forests. Northwest Science 75: 77-83. Wahbe, T.R., F.L. Bunnell and R.B. Bury. 2004. Terrestrial movements o f juvenile and adult tailed frogs in relation to timber harvest in coastal British Columbia. Canadian Journal o f Forest Research 34: 2455-2466. Wallace, R.L. and L.V. Diller. 1998. Length o f the larval cycle of Ascaphus truei in coastal streams in the Redwood Region, Northern California. Journal o f Herpetology 32: 404409. Welsh, H.H. Jr. and L.M. Olliver. 1998. Stream amphibians as indicators o f ecosystem stress: A case study from California’s redwoods. Ecological Applications 8: 1118-1132. Whiles, M.R., and J.W. Grubaugh. 1996. Importance o f coarse woody debris to southern forest herpetofauna. In Proceedings o f the Workshop on Coarse Woody Debris in Southern Forests: Effects on Biodiversity. Biodiversity o f Coarse Woody Debris in Southern Forests. J.W. McMinn, D.A. Crossley Jr. (eds.). U.S. Forest Service General Technical Report SE-94. Williams, R.L. 2000. A note on robust variance estimation for cluster-correlated data. Biometrics 56: 645-646. Wind, E. 2009. Coastal tailed frog inventory and habitat assessment. E. Wind ConsultingFinal Report BCRP project number: 09.W.BRG.01. Young, K.A. 2000. Riparian zone management in the Pacific Northwest: W ho’s cutting what? Environmental Management 26: 131 -144. 112 Appendix I- Diagram demonstrating the pitfall trap with plastic insert and escape rope. During the 2012 spring session, the escape rope did not contain knotted sections. After discussion with M. Todd, knotted sections were implemented for the remainder o f the 2012 trap session and all o f 2013 sessions. Appendix II- Logistic regression models used for the 2012 trap data conducted in 2 watersheds located east o f Terrace, BC, TEM PORAL trap month TREATM ENT densiometer site moisture canopy code de transition distance stream distance stream de transition site moisture distance stream de transition distance stream site moisture de transition canopy code site elevation site elevation2 site elevation site elevation2 distance stream densiometer orientation distance stream de transition ENVIRONM ENTAL temp difference hum difference temp difference tem p m in temp_min2 temp max temp max2 days rain hum difference ENVIRONMENTAL & TREATM ENT COMBINED distance stream temp difference distance stream hum difference de transition temp max2 temp max k Log Likelihood 4 -2LL AIC AICc -514.932 1029.865 1037.865 6 9 5 9 2 5 4 5 3 4 2 9 -506.428 -488.819 -489.436 -486.976 -491.792 -511.260 -508.890 -510.637 -506.884 -481.048 -510.837 -485.402 1012.856 977.639 978.872 973.953 983.584 1022.519 1017.781 1021.273 1013.768 962.096 1021.674 970.803 3 2 3 3 2 2 -504.073 -507.130 -514.469 -512.726 -505.395 -504.106 3 3 6 -484.228 -482.626 -506.506 114 A AIC AIC weight 1038.891 85.3 <0.0001 1024.856 995.639 988.872 991.953 987.584 1032.519 1025.781 1031.273 1019.768 970.096 1025.674 988.803 1027.127 1000.933 990.450 997.247 987.877 1034.098 1026.807 1032.852 1020.368 971.121 1025.967 994.098 73.6 47.4 36.9 43.7 34.3 80.6 73.3 79.3 66.8 17.6 72.4 40.6 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 1008.145 1014.259 1028.938 1025.453 1010.789 1008.212 1014.145 1018.259 1034.938 1031.453 1014.789 1012.212 1014.745 1018.552 1035.538 1032.053 1015.082 1012.505 61.2 65.0 82.0 78.5 61.5 59.0 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 968.457 965.252 1013.012 974.457 971.252 1025.012 975.057 971.852 1027.283 21.5 18.3 73.7 <0.0001 <0.0001 <0.0001 de transition temp difference tem pm ax temp max2 site moisture tem pdifference site moisture temp difference canopy code de transition canopy code days rain days rain site elevation site elevation2 distance stream densiometer days rain orientation canopy code days rain temp difference site elevation site elevation2 orientation days rain TREATM ENT & TEM PORAL orientation trap month distance stream orientation trap month orientation trap month de transition trap month distance stream trapjnonth de transition trap month distance stream de transition site elevation site elevation2 trap month trap month densiometer TREATM ENT, ENVIRONMENTAL & TEM PORAL trap month orientation days rain distance stream trap month days rain distance stream trap_month site elevation site elevation2 days rain de_transition trap_month site_elevation site elevation2 temp difference de transition trap month days rain distance_stream densiometer trap_month site elevation site elevation2 days rain 5 7 6 6 9 4 4 10 4 6 -502.708 -505.679 -499.182 -503.723 -491.180 -497.050 -478.801 -494.004 -494.004 -501.388 1005.415 1011.359 998.364 1007.447 982.360 994.100 957.603 988.009 988.008 1002.775 1015.415 1025.359 1010.364 1019.447 1000.360 1002.100 965.603 1008.009 996.008 1014.775 1016.994 1028.470 1012.634 1021.717 1005.654 1003.126 966.628 1014.675 997.033 1017.046 63.4 74.9 59.1 68.2 52.1 49.6 13.1 61.1 43.5 63.5 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.0007 <0.0001 <0.0001 <0.0001 9 8 11 5 7 8 6 5 -485.514 -510.939 -502.656 -489.550 -506.661 -487.191 -504.654 -508.614 971.029 1021.878 1005.312 979.100 1013.322 974.383 1009.307 1017.229 989.029 1037.878 1027.312 989.100 1027.322 990.383 1021.307 1027.229 994.323 1041.992 1035.562 990.679 1030.433 994.497 1023.577 1028.808 40.8 88.4 82.0 37.1 76.9 41.0 70.0 75.3 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 9 6 -496.935 -475.193 993.870 950.386 1011.870 962.386 1017.164 964.656 63.6 11.1 <0.0001 0.0020 8 -466.747 933.493 949.493 953.608 0.1 0.4908 10 8 -480.949 -492.185 961.898 984.370 981.898 1000.370 988.565 1004.484 35.0 50.9 <0.0001 <0.0001 9 -465.125 930.251 948.251 953.545 0.0 0.5064 115 Appendix II- Logistic regression models used for the 2013 trap data conducted in 2 watersheds located east o f Terrace, BC, identifying the distribution o f tailed frogs within forest retention treatments. __________ i_____ Log k AIC weight Likelihood -2LL AIC AIC c A AIC TEM PORAL code trap month (trap month) <0.0001 6 -196.277 392.553 404.553 406.8233 40.7 TREATM ENT densiometer site moisture -197.382 394.765 5 406.344 40.2 <0.0001 404.765 canopy code de transition distance stream 9 -193.506 387.012 405.012 <0.0001 410.306 44.2 distance stream de transition <0.0001 5 -195.630 391.260 401.260 402.839 36.7 <0.0001 site moisture distance stream de transition 9 -194.296 388.591 406.591 45.7 411.885 distance stream <0.0001 2 -198.672 397.343 401.343 401.636 35.5 <0.0001 site moisture 5 -198.023 396.046 406.046 407.625 41.5 <0.0001 de transition 4 401.410 402.436 36.3 -196.705 393.410 <0.0001 canopycode 40.4 5 -197.488 394.977 404.977 406.556 <0.0001 site elevation site elevation2 28.2 2 -195.016 390.032 394.032 394.325 <0.0001 393.871 28.3 site elevation site elevation2 distance stream 3 -193.935 387.871 394.471 <0.0001 densiometer 36.2 2 -199.021 398.042 402.042 402.335 <0.0001 36.7 orientation distance stream de transition 9 -189.791 379.581 397.581 402.875 ENVIRONMENTAL <0.0001 temp difference hum difference 3 -195.498 390.995 397.595 31.4 396.995 <0.0001 31.2 temp difference 397.378 2 -196.543 393.086 397.086 <0.0001 temp min temp min2 3 -198.023 396.045 402.645 36.5 402.045 <0.0001 37.9 temp max temp max2 403.491 404.091 3 -198.746 397.491 days rain 8.5 0.0133 2 -185.171 370.343 374.343 374.635 <0.0001 hum difference 2 -198.477 396.955 400.955 401.248 35.1 ENVIRONMENTAL & TREATM ENT COMBINED 398.276 32.1 <0.0001 3 -195.838 391.676 397.676 distance stream temp difference <0.0001 35.7 3 -197.619 395.238 401.238 401.838 distance stream hum difference 6 402.994 <0.0001 de transition temp max2temp max -195.497 390.994 405.264 39.1 de transition temp difference temp max temp max2 site moisture temp difference site moisture temp difference canopy code de transition canopy code days rain days rain site elevation site elevation2 distance stream densiometer days rain orientation canopy code days rain temp difference site elevation site elevation2 orientation days rain TREATM ENT & TEM PORAL orientation trap month distance stream orientation trap month orientation trap month de transition trap month distance stream trap month de transition trap month distance stream de transition site elevation site elevation2 trap month trap month densiometer TREATM ENT, ENVIRONMENTAL & TEM PORAL trap month orientation days rain distance stream trap month days rain distance_stream trap_month site_elevation site elevation2 days rain de_transition trap_month site_elevation site elevation2 temp difference de transition trap month days rain distance_stream densiometer code_trap_month site elevation site elevation2 days rain 5 7 6 6 9 4 4 10 3 6 -192.911 -197.008 -194.730 -192.943 -178.084 -178.563 -182.718 -177.083 -190.63614 -180.009 385.822 394.015 389.459 385.886 356.168 357.126 365.437 354.167 381.272 360.019 395.822 408.015 401.459 397.886 374.168 365.126 373.437 374.167 387.272 372.019 397.401 411.126 403.729 400.156 379.462 366.152 374.462 380.834 387.872 374.289 31.2 45.0 37.6 34.0 13.3 0.0 8.3 14.7 21.7 8.1 <0.0001 <0.0001 <0.0001 <0.0001 0.0012 0.9262 0.0145 0.0006 <0.0001 0.0158 11 10 13 7 9 10 8 7 -189.212 -190.436 -187.612 -195.057 -193.456 -192.370 -191.637 -195.260 378.424 380.872 375.224 390.113 386.912 384.741 383.273 390.520 400.424 400.872 401.224 404.113 404.912 404.741 399.273 404.520 408.674 407.539 413.357 407.224 410.206 411.408 403.388 407.631 42.5 41.4 47.2 41.1 44.1 45.3 37.2 41.5 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 11 8 -176.629 -180.406 353.259 360.812 375.259 376.812 383.509 380.926 17.4 14.8 0.0002 0.0006 10 -173.438 346.875 366.875 373.542 7.4 0.0230 12 10 -184.880 -178.652 369.760 357.305 393.760 377.305 403.825 383.972 37.7 17.8 <0.0001 0.0001 11 -173.288 346.577 368.577 376.827 10.7 0.0045 Appendix IV: Temperature and days since last rain at the 5- and 80-m arrays for 4 forest retention treatments in 2 watersheds located east o f Terrace, BC during 2012. Gosling watershed. Kleanza watershed ■ i 17 10 tt 45 4U 35 30 25 6 4 20 15 10 2 J O 1 7 Jul 1 8 Jig 1 9 Jill 2 0 Jill 2 1 Jul S p rin g tra p sp w io n 5 o L 22-Jul 2 2 Jul 73-Jul 24-Jul 2S 4u l 76-lul 27-Jul cm days nncc rain Spring tra p session 12 in H t> 4 45 -♦-O ld growth 5 m 40 35 ••■••Old growth 80 m 30 25 20 -♦ -5 0 - forest retention 5 m 15 ID 2 0 2 S A l« 26-A t« 27-Aur 28-A t« 29-Aur S u m m e r tra p s e s sio n 12 40 lO AS JO A 25 20 6 15 4 10 2 O ItVOrt 11-Ort 12-Orr Fall tr a p s e s s io n 13-Ort 14-flrt ■••50- forest retention 80 m I9-AUH 20-AUR 21-Aun 22-AUR 23-AgR S u m m e r tra p s e s s io n 3M m 45 9-Orf 5 O 2 4-A w ° Clearcnt/Rflgeneratioii 5 m 17 lO 45 40 35 TO A 25 6 70 4 15 ID 2 5 O 0 ^ 2 -O tl 3-Out 4-O cl t a il tra p s e s s io n 5-Oul ■ Cleaivat/Regeneratkxi 80m Appendix V: Temperature and days since last rain for 5-m and 80-m arrays in 4 forest retention treatments in 2 watersheds located east o f Terrace, BC during 2013. Gosling watershed Kleanza watershed o Temperature(°C) a lul 4 till .Villi ft-tal 7 lul A-lul fl-liin 4-lw n __________________ Sprtnfl trap sessio n _____________ lul lO a o A O 1/JU n IMAlft 1*#A M R iOAir. f> lu n /lu n S tu n czndnys since rtin V-lnn Spring trap sessio n 14 IbAUR S ta n 45 40 35 30 14 1? lO 20 15 lO 5 O 7-Atrn »A u g ^Au« IO A i« llA « g I2A u r «•" deucut/Regeneratioa 5 m Summertrap sgwlon • Clement/Regeneration 80 m 40 lO lO s o KtSnp K ir* ___________________ f ail (1 ) trap sessio n _____________ 45 < » S rrn ms'll n v (» i7Sr.|i mS'n us-p la li trap 14 40 35 1J .HK) 1(1 25 B 20 O 15 4 *>* J. I i a-ocr I all (2 ) trap sessio n w •••••Old growth 80 m .3 “ -•-5 0 - forest retention 5 m s>* » ••*••50-fbietf retention 80 m ilAUR ___________________ S u m m er tra p sessio n ___________ 121 ” —•—Old growth 5 m o Appendix VI: Graphical representation o f optimal temperature o f used locations by the tailed frog in 3 watersheds located east o f Terrace, BC, during 2011 and 2012. Largest bar represents the optimal temperature for the model based on the Gaussian term (linear and quadratic). The pooled tailed frog model and the female tailed frog model are associated with 11°C, while the male tailed frog data is associated with 12°C. _________________ < D J—T' o o wEu cn o£ IFemale data IM ale data •5 lFemale and male data 0 2 4 6 8 10 11 12 13 14 16 18 20 22 Temperature (°C) 121 Appendix VII: Proportion of used locations associated with light levels >103 and <103 lux for each gender across 3 forest retention treatments (old growth, forest retention buffer and clearcut) for 24 radio telemetered frogs in 3 watersheds east of Terrace, BC during 2011 and 2012. Male Female Forest Clearcut Total (102) Clearcut Total Old Forest Old retention (27) (184) growth retentio growth (0) (60) buffer (42) n buffer (132) (91) < 103 lux 93.41 93.59 90.76 88.33 83.33 0 86.27 93.18 > 103 lux 6.81 6.59 7.41 9.24 16.67 0 13.73 11.67