i UTILIZING PARNASSIUS CLODIUS BUTTERFLIES AND NECTAR PLANT SPECIES TO EVALUATE ECOLOGICAL RESPONSES TO CLIMATE CHANGE by Janice Simone Durney A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Ecology and Environmental Sciences MONTANA STATE UNIVERSITY Bozeman, Montana December 2023 ©COPYRIGHT by Janice Simone Durney 2023 All Rights Reserved ii DEDICATION I dedicate my dissertation to all the young scientists. Keep being curious. iii ACKNOWLEDGEMENTS I would like to thank all my funding sources and previous funding sources for allowing this research to continue which include the Center for Global and Regional Environmental Research, Decagon Devices, the Disney Conservation Fund, Iowa State University Department of Ecology, Evolution, and Organismal Biology, Iowa State University Graduate Program in Ecology and Evolutionary Biology, Montana State University Ecology Department, Montana State University Graduate School, Montana Wilderness Association – Eastern Wildlands Chapter’s Wildlands Community Grant, Sitka Ecosystem Grants, Teton Conservation District, and the Xerces Society. I would like to thank Montana State University’s Statistical Consulting and Research Services for their statistical advisement concerning some of this research. Thank you to the University of Wyoming NPS Research Station for housing numerous teams of field researchers throughout the years. A very special thanks go to all of those who assisted with the field studies throughout the years: M. Germino, K. Reinhardt, J. Sherwood, K. Szcodronski, A. McCombs, T. Proscholdt, K. McCloskey, C. Gause, A. Taylor, M. Kyer, D. Nelson, A. Binder, J. Pink, H. Pritchard, L. Crees, E. McCall, B. Sisson, A. Engel, B. Servais, E. Knodle, H. Buckbee, L. McMahon, and K. Guajardo. None of this research and culminating dissertation would be possible without the thoughtfulness and patience of my committee members, so, thank you to Dr. Diane Debinski, Dr. Laura Burkle, Dr. Stephen Matter, and Dr. Danielle Ulrich. A huge thank you to the Ecology Department graduate students past and present, my friends, and my family for being my community throughout this process. Jordan thank you for being my person, my sounding board, and the person I can always turn to. iv TABLE OF CONTENTS 1. INTRODUCTION .............................................................................................................. 1 2. MONTANE BUTTERFLY DEVELOPMENTAL LIFE STAGES AS INDICATORS OF INSECT VULNERABILITY TO EXTREME CLIMATIC CONDITIONS ................................................................................................ 6 Contribution of Authors and Co-Authors ........................................................................... 6 Manuscript Information ...................................................................................................... 7 Abstract ............................................................................................................................... 8 Introduction ......................................................................................................................... 8 Materials and Methods ...................................................................................................... 15 Species Description ............................................................................................... 15 Study Area ............................................................................................................. 16 Butterfly Mark-Recapture ..................................................................................... 16 Weather and Climate Variable Selection ............................................................... 17 Analysis................................................................................................................. 18 Results ............................................................................................................................... 21 Egg Life Stage....................................................................................................... 21 Larval-Pupal Life Stage ........................................................................................ 22 Adult Life Stage .................................................................................................... 22 Discussion ......................................................................................................................... 29 References ......................................................................................................................... 34 3. EARLIER SPRING SNOWMELT DRIVES ARROWLEAF BALSAMROOT PHENOLOGY IN MONTANE MEADOWS ...................................... 39 Contribution of Authors and Co-Authors ......................................................................... 39 Manuscript Information .................................................................................................... 40 Abstract ............................................................................................................................. 41 Introduction ....................................................................................................................... 42 Methods and Materials ...................................................................................................... 46 Study Site .............................................................................................................. 46 Experimental Methods and Design ....................................................................... 47 Statistical Analysis ................................................................................................ 49 Results ............................................................................................................................... 53 Flower Onset ......................................................................................................... 53 Flowering Duration ............................................................................................... 53 Maximum Floral Display (i.e., Maximum Number of Concurrent Flowers) ................................................................................................................ 54 Relationships Between Flower Onset, Flowering Duration, and Maximum Floral Display ...................................................................................... 55 Discussion ......................................................................................................................... 59 v TABLE OF CONTENTS CONTINUED Heating with Snow Removal Leads to Earlier Flowering Onset .......................... 60 Heating with Snow Removal Extended Flowering Durations .............................. 61 Maximum Floral Display Increased Under Heating with Snow Removal ................................................................................................................ 62 Earlier Flower Onset Associated with Longer Flowering Durations and Greater Maximum Floral Displays................................................................. 63 Potential Implications to Plant-Pollinator Interactions ......................................... 65 Conclusions ....................................................................................................................... 66 References ......................................................................................................................... 67 4. ROCKY MOUNTAIN FORBS EXHIBIT SPECIES SPECIFIC PHENOLOGICAL AND PHOTOSYNTHETIC RESPONSES TO HEATING ......................................................................................................................... 70 Abstract ............................................................................................................................. 70 Introduction ....................................................................................................................... 71 Materials & Methods ........................................................................................................ 79 Study Site .............................................................................................................. 79 Experimental Methods and Design ....................................................................... 79 Statistical Methods Evaluating Flower Onset, Flowering Duration, and Maximum Floral Display Responses to Heating ........................................... 82 Statistical Methods Evaluating FV/FM Response to Heating ................................ 83 Statistical Methods Evaluating Which Environmental Conditions are the Best Predictors of FV/FM ........................................................................... 84 Statistical Methods Evaluating Flower Onset, Flowering Duration, and Maximum Floral Display Responses to Mean FV/FM .................................... 85 Results ............................................................................................................................... 87 Discussion ....................................................................................................................... 106 References ........................................................................................................................113 5. WINGS THAT MAKE WAVES ......................................................................................117 Contribution of Authors and Co-Authors ........................................................................117 Manuscript Information ...................................................................................................118 6. CONCLUSION ............................................................................................................... 213 REFERENCES CITED ......................................................................................................... 217 APPENDICES ...................................................................................................................... 225 EARLIER SPRING SNOWMELT DRIVES ARROWLEAF BALSAMROOT PHENOLOGY IN MONTANE MEADOWS ........................ 226 vi LIST OF TABLES Table Page 1. Table 1: The nine explanatory weather and climate variables tabulated by each Parnassius clodius life stage included in analyses. .................................................. 20 2. Table 2: AIC comparison table for models testing for the best predictive weather and climate explanatory variables among highly correlated (i.e., r >0.5) variables to be included in full generalized linear models for each life stage (i.e., egg, larval-pupal, adult). The best predictive weather and climate variables for each life stage are in bold font. *Note: due to correlation between maximum and mean snow water equivalent, maximum snow water equivalent was chosen as the predictor variable during the adult life stage since the AIC values were the same between the two variable models. ........................... 26 3. Table 3: AIC comparison table for full generalized linear models including the best predictive weather and climate explanatory variables at each life stage (i.e., egg, larval-pupal, adult), their interactions, and density dependence on population change Rt. The best model for each life stage is listed in bold font. ............................................................................................................. 27 4. Table 4: Hypotheses about how B. sagittata flower onset, flowering duration, maximum floral display may respond to decreased spring snowmelt via removal, warming via passive heating, and warming with decreased spring snowpack. .............................................................................................. 46 5. Table 5: Multiple comparison of means with Tukey contrasts to identify whether treatment affected individual plant flowering onset, flowering duration, or maximum floral display. Each model included heated and snow removal treatments as main effects as well as their interaction with individual nested within plot as a random effect, and year as an additive random effect. Bolded p-values indicate significant effects at α < 0.05. .......................... 59 6. Table 6: Mixed effects models to identify whether treatment × year interactions affected individual plant flowering onset, flowering duration, or maximum floral display over time. The treatment × year interactions were tested to evaluate if treatment effects grew stronger over time. Each model included the interaction of year and heated and snow removal treatments with individual nested within plot as a random effect. Bolded p- values indicate significant effects at α < 0.05. .................................................................. 59 vii LIST OF TABLES CONTINUED 7. Table 7: Hypotheses about how Achillea millefolium, Geranium viscosissimum, and Penstemon procerus flower onset, flowering duration, maximum floral display (i.e., maximum number of concurrent flowers), and mean chlorophyll fluorescence may respond to passive heating. Additionally, hypotheses about how Achillea millefolium, Geranium viscosissimum, and Penstemon procerus chlorophyll fluorescence responds to passive heating and various local temperature and moisture environmental conditions. Explanatory variables are listed in the left column and response variables are listed on the top row. ................................................. 78 8. Table 8: Temperature and moisture environmental condition variables collected on site using HOBO loggers and soil dataloggers to identify predictor environmental condition variables affecting FV/FM. Listed metrics derived from each temperature and moisture environmental condition variable. ............................................................................................................................. 87 9. Table 9: Analysis of variance and deviance results testing the effect of treatment (heating versus control) on Julian date of flower onset, flowering duration in days, and maximum floral display (i.e., maximum number of concurrent flowers) for Achillea, Geranium, and Penstemon. Bolded p values indicate significant effects at α <0.05. ................................................................. 101 10. Table 10: Analysis of variance results testing the effect of treatment (heating versus control), year, and the interactive effect of treatment and year on Julian date of flower onset, flowering duration in days, and maximum floral display (i.e., maximum number of concurrent flowers) for Achillea, Geranium, and Penstemon. Bolded p values indicate significant effects at α <0.05. .............................................................................................................................. 101 11. Table 11: Analysis of variance testing the effect of treatment (heating versus control), species (Achillea, Geranium, and Penstemon), and sampling date on FV/FM. Bolded p values indicate significant effects at α <0.05.................................. 102 12. Table 12: Analysis of variance testing the effect of treatment (heating versus control), year (2020 versus 2021), and the interactive effect of treatment and year on FV/FM for each species (Achillea, Geranium, and Penstemon). Bolded p values indicate significant effects at α <0.05. ................................................. 103 13. Table 13: Analysis of variance testing the effect of treatment (heating versus control), species (Achillea, Geranium, and Penstemon), sampling date, and the interactive effect of treatment and sampling date on FV/FM for each species. Bolded p values indicate significant effects at α <0.05. .................................... 103 viii LIST OF TABLES CONTINUED 14. Table 14: Analysis of variance testing the effect of treatment (heating versus control), sampling date, and the interactive effect of treatment and sampling date on FV/FM for each species (Achillea, Geranium, and Penstemon). Bolded p values indicate significant effects at α <0.05. ................................................. 103 15. Table 15: AIC comparison of generalized linear models fit to evaluate which correlated explanatory environmental condition variable metrics were the most informative in explaining FV/FM. Models compared include metrics for each environmental condition variable (i.e., soil temperature, soil moisture, air temperature, and surface temperature). Bold values indicate the lowest AIC and its associated environmental condition variable metric. ................. 104 16. Table 16: AIC comparison of generalized linear models fit with interactions between the most informative temperature and moisture explanatory variables used to evaluate potential interactive effects between environmental condition variables on FV/FM. Bold values indicate the best model with the lowest AIC. ............................................................................................ 104 17. Table 17: Analysis of variance results testing the effect of maximum soil temperature, maximum soil moisture, maximum air temperature, maximum surface temperature, the interactive effect of maximum soil temperature and maximum air temperature, the interactive effect of maximum soil temperature and maximum surface temperature, and treatment on FV/FM. Bolded p values indicate significant effects at α <0.05. ................................................. 105 18. Table 18: Analysis of variance results testing the effect of Mean FV/FM, treatment (heating versus control), and the interactive effect of Mean FV/FM and treatment on Julian date of flower onset, flowering duration in days, and maximum floral display (i.e., maximum number of concurrent flowers) for Achillea, Geranium, and Penstemon. Bolded p values indicate significant effects at α <0.05. .......................................................................................... 105 19. Table 19: The minimum, maximum, mean, and median values for soil temperature, soil moisture, air temperature, and surface temperature for the control and heating treatments. Values were derived from the HOBO temperature loggers and soil dataloggers located in the center of each plot. Air temperature was the same for control and heating plots since the 12 plots shared the same air sensor. ...............................................................................................112 ix LIST OF FIGURES Figure Page 1. Figure 1: Parnassius clodius butterfly. Photo Credit J. Simone Durney. ......................... 20 2. Figure 2: Years of study with associated population change (Rt) values represented in black, no population change represented by the horizontal blue line, and density dependence (logNt) values represented in green. Note: surveys were not carried out in 2012 and 2013. ............................................................... 23 3. Figure 3: The relationship between Parnassius clodius population change (Rt) from 2009 to 2018 by life stage to the precipitation explanatory variables considered in these analyses. The egg life stage is represented by the color purple, the larval-pupal life stages are represented by the color green, and the adult life stage is represented by the color orange. Julian date of snowmelt did not vary across the life stages and is represented by the color black. Lines and points indicate a significant relationship between population change and a descriptive weather or climate variable at a particular life stage. Note: surveys were not carried out in 2012 and 2013. Overall, there were no significant relationships detected between population change (Rt) and date of snowmelt, maximum snow water equivalent (cm), mean snow water equivalent (cm), and accumulated rainfall (cm). ................................. 24 4. Figure 4: The relationship between Parnassius clodius population change (Rt) from 2009 to 2018 by life stage to the temperature explanatory variables considered in these analyses. The egg life stage is represented by the color purple, the larval-pupal life stages are represented by the color green, and the adult life stage is represented by the color orange. Darker, thicker lines and points indicate a significant relationship between population change and a descriptive weather variable at a particular life stage. Note: surveys were not carried out in 2012 and 2013. Overall, a significant negative relationship between extreme maximum temperatures during the larval- pupal life stage and population change (Rt) was detected. There were no significant relationships detected between population change (Rt) and mean temperature (C), mean maximum temperature (C), mean minimum temperature (C), and extreme minimum temperature (C). ............................................... 25 5. Figure 5: Parnassius clodius population change (Rt) from 2009 to 2018 relative to the extreme maximum air temperature of its environment during the larva-pupa life stages. Note: surveys were not carried out in 2012 and 2013................................................................................................................................... 34 x LIST OF FIGURES CONTINUED 6. Figure 6: The effect of treatment on the least squares means flower onset date, flowering duration, and maximum floral display. Treatments included: decreased spring snowpack via snow removal (SR: not heated, snow removal), warming via passive heating (H: heated, no removal), decreased spring snowpack and warming (H+SR: heated, snow removal), and control (C: not heated, no removal). Estimated marginal means pairwise comparison are indicated by letters where the same letter indicates overlap and different letters indicate variation. The error bars represent the upper and lower 95% confidence levels. .................................................................................... 56 7. Figure 7: Treatment effects over time on flower onset, flowering duration, and maximum floral display. Treatments included: decreased spring snowpack via snow removal (SR: not heated, snow removal), warming via passive heating (H: heated, no removal), decreased spring snowpack and warming (H+SR: heated, snow removal), and control (C: not heated, no removal). Figure displays results of treatment interactions with year. Asterisks indicate a significant treatment × year interaction (indicated by treatment color) compared to the control on the response variable (i.e., flowering onset, flowering duration, and maximum floral display). The normal confidence intervals (indicated by the gray shaded areas along each line) were constructed using generalized linear smoothing to observe linear change in each treatment over time. Flowering onset did not occur in snow removal plots in 2019, therefore, there is no data available for that treatment and year. ............................................................................................................................ 57 8. Figure 8: The relationships between the raw values for plant individuals of Balsamorhiza sagittata flower onset and flowering duration, flowering duration and maximum floral display, and flower onset and maximum floral display within each treatment across the seven-year study. Treatments included: decreased spring snowpack via snow removal (Snow Removal: not heated, snow removal), warming via passive heating (Heated: heated, no removal), decreased spring snowpack and warming (Heated and Snow Removal: heated, snow removal), and control (Control: not heated, no removal). A best fit line was added to display the relationship between the variables, with bolded lines indicating a significant correlation....................................... 58 9. Figure 9: Plot layout where H represents the heating treatment using passive heating structures (N = 6) and C represents the control treatment (N = 6). Plots were 2.4 m2 (8 ft X 8 ft), 59 cm tall, and spaced 4.5 m apart on the north and south sides and 6 m apart on the east and west sides. Green represents control plots and brown represents heating plots. ........................................... 86 xi LIST OF FIGURES CONTINUED 10. Figure 10: Plant layout within each plot which included 30 individuals total with 10 individuals for each of the three species. The three species included Achillea millefolium (orange circle), Geranium viscosissimum (blue square), and Penstemon procerus (yellow triangle). Each number corresponds to that plants individual identification number within each plot. ........................................................................................................................................... 87 11. Figure 11: Effect of treatment on the least squares mean flower onset date, flowering duration, and maximum floral display for Achillea, Geranium, and Penstemon accounting for both years. Treatments included heating via passive heating and control. Estimated marginal mean pairwise comparison is indicated by letters where the same letter indicates overlap and different letters indicate variation. The error bars represent the upper and lower 95% confidence levels. .............................................................................................................. 91 12. Figure 12: Treatment effects between 2020 and 2021 on flower onset, flowering duration, and maximum floral display for Achillea, Geranium, and Penstemon. Treatments included heating via passive heating and control. The figure displays results of treatment interactions with year. Lines indicate a significant treatment × year interaction (indicated by treatment color) compared with the control on the response variable (i.e., flowering onset, flowering duration, and maximum floral display). The normal confidence intervals (indicated by the gray-shaded areas along each line) were constructed using generalized linear smoothing to observe a linear change in each treatment over time. There was a significant difference in flower onset, flowering duration, and maximum floral display for each species between 2020 and 2021. ....................................................................................... 92 13. Figure 13: Effect of treatment on the least squares mean FV/FM for Achillea, Geranium, and Penstemon. Treatments included heating via passive heating and control. Estimated marginal mean pairwise comparison is indicated by letters where the same letter indicates overlap and different letters indicate variation. The boxplot error bars represent the upper and lower 95% confidence levels. .............................................................................................................. 93 14. Figure 14: Effect of treatment between 2020 and 2021 on the least squares mean FV/FM for Achillea, Geranium, and Penstemon. Treatments included heating via passive heating and control. Estimated marginal mean pairwise comparison is indicated by letters where the same letter indicates overlap and different letters indicate variation. The boxplot error bars represent the upper and lower 95% confidence levels. .......................................................................... 94 xii LIST OF FIGURES CONTINUED 15. Figure 15: Treatment effects during the growing seasons for the two-year study on FV/FM of Achillea, Geranium, and Penstemon. Treatments included heating via passive heating and control. The figure displays results of treatment interactions with sampling date. Lines indicate a significant treatment × date interaction (indicated by treatment color) compared with the control on FV/FM. The normal confidence intervals (indicated by the gray-shaded areas along each line) were constructed using generalized linear smoothing to observe a linear change in each treatment over time. There was a significant difference in FV/FM for each species over time. ......................... 95 16. Figure 16: Treatment effects during the growing seasons for the two-year study on FV/FM for each species (Achillea, Geranium, and Penstemon). Treatments included heating via passive heating and control. The figure displays results of treatment interactions with sampling date. Lines indicate a significant treatment × date interaction (indicated by treatment color) compared with the control on FV/FM. The normal confidence intervals (indicated by the gray-shaded areas along each line) were constructed using generalized linear smoothing to observe a linear change in each treatment over time. There was no significant difference in FV/FM between treatments. ........................................................................................................................................... 96 17. Figure 17: Relationships between best predictive temperature and moisture variables (maximum soil moisture, maximum soil temperature, maximum air temperature, maximum surface temperature) from best model and FV/FM across all species (Achillea, Geranium, and Penstemon). Treatments included heating via passive heating and control. Lines indicate a significant effect of that variable on FV/FM. The normal confidence intervals (indicated by the gray-shaded areas along each line) were constructed using generalized linear smoothing to observe a linear relationship. Lines color coded by treatment to evaluate differences between the two treatments. There was no significant difference in FV/FM between treatments. .................................. 97 xiii LIST OF FIGURES CONTINUED 18. Figure 18: Relationships between mean maximum soil moisture, mean maximum soil temperature, mean maximum air temperature, mean maximum surface temperature and mean FV/FM across all species (Achillea, Geranium, and Penstemon). Treatments included heating via passive heating and control. Lines were added to display the relationship between the variables and mean FV/FM. The normal confidence intervals (indicated by the gray-shaded areas along each line) were constructed using generalized linear smoothing to observe a linear relationship. Lines color coded by treatment to evaluate differences between the two treatments. There was no significant difference in mean FV/FM between treatments when each of the following abiotic variables were included in analysis of variance tests: mean maximum soil moisture test – Treatment: F1,13 = 0.31, p value = 0.5870; mean maximum soil temperature test – Treatment: F1,13 = 0.20, p value = 0.6590; mean maximum air temperature test – Treatment: F1,9 = 0.07, p value = 0.7991; mean maximum surface temperature test – Treatment: F1,16 = 2.24, p value = 0.1539. ........................................................................ 98 19. Figure 19: Relationships between mean maximum soil moisture, mean maximum soil temperature, mean maximum air temperature, mean maximum surface temperature and mean FV/FM across species (Achillea, Geranium, and Penstemon). Lines were added to display the relationship between the variables and mean FV/FM. The normal confidence intervals (indicated by the gray-shaded areas along each line) were constructed using generalized linear smoothing to observe a linear relationship. Lines color coded by plant genus to evaluate differences between the three plant species. There were significant differences in mean FV/FM across plant species when each of the following abiotic variables were included in analysis of variance tests: mean maximum soil moisture test – Genus: F2,274 = 101.49, p value = <0.0001; mean maximum soil temperature test – Genus: F2,275 = 99.12, p value = <0.0001; mean maximum air temperature test – Genus: F2,275 = 102.25, p value = <0.0001; mean maximum surface temperature test – Genus: F2,275 = 98.44, p value = <0.0001. ........................................... 99 20. Figure 20: Significant relationships between mean FV/FM, Achillea flowering duration, and Geranium maximum floral display. Lines indicate a significant mean FV/FM × treatment interaction (indicated by treatment color) compared with the control on flowering duration and maximum floral display. The normal confidence intervals (indicated by the gray-shaded areas along each line) were constructed using generalized linear smoothing to observe a linear relationship. Lines color coded by treatment to evaluate differences between the two treatments. There was no significant difference in mean FV/FM between treatments. ................................................................................ 100 xiv LIST OF FIGURES CONTINUED 21. Figure A1: Schematic of the study area with plots of corresponding treatments of decreased spring snowpack via snow removal (SR: not heated, snow removal), warming via passive heating (H: heated, no removal), decreased spring snowpack and warming (H+SR: heated, snow removal), and control (C: not heated, no removal). ........................................................................ 134 22. Figure A2: Individual Balsamorhiza sagittata with identification tag for continued monitoring. ..................................................................................................... 135 xv ABSTRACT Climate change is a pressing global issue leading to species declines, altering species interactions. Changing environmental conditions result in species adaptation or movement to suitable areas. High elevations and latitudes are susceptible to rapid change, potentially leaving no refuge for species. Plants and pollinators are particularly vulnerable to changing environmental conditions with their life cycles dependent on the timing of environmental cues that initiate species morphology, physiology, and behavior patterns. A disruption in a species’ typical life-cycle pattern may lead to phenological shifts affecting biological events potentially inducing phenological mismatches between interacting species. This research addresses (1) how current and predicted environmental conditions affect the population growth rate of Parnassius clodius butterflies, (2) investigates the effect of warming and advanced spring snowmelt on arrowleaf balsamroot phenology and morphology, and (3) understands the effect of warming on the phenology, morphology, and photosynthetic performance of three prominent native flowering species in the Rocky Mountain Region. A mark-recapture-release program paired with life-stage hypothesis modeling was used to evaluate population growth rates of Parnassius clodius. Plots with plant species of interest were experimentally manipulated (warming, snow removal) and phenology, morphology, and photosynthetic performance of individuals monitored and evaluated using mixed-effects models. Findings from this dissertation determined that (1) extreme maximum temperatures during the spring led to decreased population growth rates of Parnassius clodius butterflies, (2) heating via passive warming and advanced spring snowmelt via snow removal led to earlier flower onset, extended flowering durations, and more flowers produced by arrowleaf balsamroot, and (3) there was no real effect of our heating treatment, but rather species-specific phenological, morphological, and photosynthetic performance differences with extended flowering durations among yarrow and geranium, more flowers produced by geranium, the greatest photosynthetic performance by yarrow followed by geranium, and maximum soil moisture, soil temperature, air temperature, and surface temperature were the best predictors of photosynthetic performance. Additionally, an iBook science communication tool was produced. These findings suggest that floral communities are faring better under a changing climate with more sensitive butterfly counterparts facing declines. Overall, this could lead to cascading effects on floral populations due to declining butterfly populations. 1 INTRODUCTION Climate change is a growing issue leading to species declines, altering species interactions globally (Valiente-Banuet et al. 2015). In montane ecosystems, these changes often pressure plants and animals to adapt or move up in elevation and/or latitude (Crozier 2004, Aguirre-Gutiérrez 2016). For example, Parmesan & Yohe (2003) found that for 99 species examined globally, ranges had shifted poleward an average of 6.1 km per decade. These shifts are further exacerbated in subalpine and alpine environments, potentially leaving species no refuge from a changing climate. Species distribution shifts like this can have dramatic effects on community and ecosystem dynamics and functions (Fontaine et al. 2006). Changes in the timing of environmental cues that initiate species morphology, physiology, and behavior patterns may lead to phenological shifts in a species’ typical life cycle pattern affecting biological events and potentially inducing phenological mismatch between interacting species (Walther et al. 2002, Parmesan and Yohe 2003, Cleland et al. 2007, Scaven and Rafferty 2013). A meta-analysis found that for 172 species of plants, birds, butterflies, and amphibians, there was a mean shift of 2.3 days per decade of earlier spring timing (Parmesan and Yohe 2003). Even seemingly minute shifts like this can have cascading effects on community dynamics by shifting species abundance and distribution (Parmesan and Yohe 2003) and have the potential to lead to phenological mismatches between species (Visser and Both 2005). Phenological mismatches between plants and pollinators are occurring rapidly because both taxonomic groups are especially vulnerable to environmental shifts (Cleland et al. 2007, Memmott et al. 2007), and such changes can have lasting effects on the overall ecosystem (Burkle and Alarcon 2011). 2 The Greater Yellowstone Ecosystem (GYE) is a relatively intact temperate ecosystem, home to many species of mammals, amphibians, reptiles, fishes, plants, and insects that are facing the effects of a changing climate. In particular, the GYE has over 120 species of butterflies. Butterfly lifecycles are directly linked to seasonality changes, and they show immediate responses to their environment with many species completing their lifecycle within one year. These butterfly species are important bioindicators that can be used to observe and monitor the effects of climate change (Kevan 1999). One species of particular interest in the GYE is Parnassius clodius, a non-migratory butterfly specialized to living in subalpine and alpine meadows and dependent on only a few flowering plant species found in high elevation and high latitude environments. Parnassius clodius populations in the GYE have been studied since 1992 and involve ongoing collaborations with Canadian and European researchers examining other Parnassius species (Auckland et al. 2004, Nakonieczny et al. 2007, Roland and Matter 2013, Filazzola et al. 2020). With a changing climate, these butterflies may be forced to seek refuge at higher elevations, potentially risking displacement to areas where the resources they depend on are not available. Parnassius clodius are an important climatic indicator because a) their lifecycles are dependent on environmental cues and changes in seasonality, b) they are well studied, c) they are widespread across the GYE, and d) they live in small populations with low fecundity, making their populations susceptible to change. For these reasons, Parnassius butterflies may serve as a “canary in a coalmine” of how climate change may be affecting ecosystems where they reside. They also expand our understanding of plant-pollinator interactions that provide ecosystem services with economic and social implications to ecosystem health in the GYE. 3 Another species of interest is arrowleaf balsamroot (Balsamorhiza sagittata), a vital early-spring floral resource for pollinators in montane regions just after snowmelt. In addition to its role as a nectar source for many generalist pollinators, B. sagittata is a valuable source of nectar for the butterfly Parnassius clodius (Sherwood et al. 2017). Additionally, it serves as a critical nectar source for some specialized bee species in the genus Osmia, which forage exclusively on Balsamorhiza and Wyethia species (Cane 2011). Therefore, B. sagittata’s demography, phenology, and flower production are important for the stability of pollinator community dynamics in montane regions in parts of the western United States in early spring. Additional species of interest in the GYE include yarrow (Achillea millefolium), sticky geranium (Geranium viscosissimum), and littleflower penstemon (Penstemon procerus). These species are native and abundant flowering plants in the western United States, that vary in morphology from one another, and are important for pollinator communities in the western United States (Bauer 1983, Scheinost and Stannard 2010, Krampien and Low 2020, Glenny et al. 2022, Mann et al. 2022). Therefore, evaluating the floral phenology, morphology, and photosynthetic performance responses to a warming climate will aid in understanding how flowering plants in the western United States will cope and respond to climate change. This has cascading implications on species that interact with and rely on these prominent floral resources providing valuable information on community dynamics of project climatic scenarios. The research presented in this dissertation focuses on investigating the following questions: 4 (1) How does the interannual variation in population size of P. clodius butterflies in Grand Teton National Park, Wyoming, USA correlate with weather variables assumed to affect Parnassius butterflies within specific life stages (Chapter 2)? (2) How do treatment effects of heating, advanced spring snowmelt, and heating with advanced spring snowmelt affect arrowleaf balsamroot flower onset, flowering duration, maximum floral display, the cumulative treatment effects over time, and the correlations among plant responses to treatments to understand how flower onset, flowering duration, and maximum floral display are associated with one another in Grand Teton National Park, Wyoming, USA (Chapter 3)? (3) How does heating affect the photosynthetic performance, phenology, morphology, the relationships between them, and the effect of local abiotic environmental conditions on photosynthetic performance of yarrow, sticky geranium, and littleflower penstemon in the Rocky Mountain Region of the USA (Chapter 4)? The final chapter (i.e., Chapter 5) of this dissertation is not an investigative research report, but rather a science communication tool highlighting Parnassius clodius butterflies and their connection to their environment and other organisms in Grand Teton National Park, Wyoming, USA. The tool encompasses a free downloadable interactive iBook promoting observation and public scientist efforts (i.e., anyone participating in the scientific process) along with teaching lessons covering phenology and morphological adaptations written for the general public, ages nine and up. Activities are associated with each lesson and include how to build your own field guide, building a butterfly character card, and phenological tracking over time. 5 The research included in this dissertation helps address how environmental conditions over time affect (1) the life stages of the montane butterfly, Parnassius clodius, reflected in population growth rates, (2) the phenological and morphological responses of an early-season native flowering plant, arrowleaf balsamroot, to predicted warming and advanced spring snowmelt, (3) the phenological, morphological, and photosynthetic performance responses of three prominent native flowering plants (i.e., yarrow, sticky geranium, and littleflower penstemon) to heating, and (4) science communication of environment-plant-pollinator interactions and public engagement in the scientific process. Answers to these knowledge gaps will improve our understanding of how native flowering plants and a butterfly species will respond to predicted climate change scenarios with implications to community and ecosystem dynamics. 6 CHAPTER TWO MONTANE BUTTERFLY DEVELOPMENTAL LIFE STAGES AS INDICATORS OF INSECT VULNERABILITY TO EXTREME CLIMATIC CONDITIONS Contribution of Authors and Co-Authors Manuscript(s) in Chapter(s) 1 Author: J. Simone Durney Contributions: contributed to data collection, conducted analyses, and was the primary author drafting the manuscript for publication Co-Author: Diane M. Debinski Contributions: received funds for research, conducted data collection and contributed to drafting manuscript for publication Co-Author: Stephen F. Matter Contributions: consulted on data analysis methodology and contributed to drafting manuscript for publication 7 Manuscript Information J. Simone Durney, Diane M. Debinski, and Stephen F. Matter Oikos Status of Manuscript: [Put an x in one of the options below, then delete instruction in brackets] ☐ Prepared for submission to a peer-reviewed journal  Officially submitted to a peer-reviewed journal ☐ Accepted by a peer-reviewed journal ☐ Published in a peer-reviewed journal The Nordic Society Oikos Submitted September 21, 2023 8 Abstract Butterflies are important bioindicators that can be used to monitor the effects of climate change, particularly in montane environments. Changes in butterfly population size over time can signal changes that have occurred or are occurring in their environment indicating ecosystem health. From the perspective of understanding butterflies as bioindicators in these systems, it is essential to identify influential environmental variables at each life stage that have the greatest effect on population dynamics. Life stage hypothesis modelling was used to assess the effects of multiple temperature and precipitation metrics on the population growth rate of a Parnassius clodius butterfly population from 2009 to 2018. Extreme maximum temperatures during the larval-pupal life stages were identified to have a significant effect on population growth rate, leading to a decline in the size of the population’s next generation. We hypothesize that higher temperatures during the spring ephemeral host plant’s flowering, and P. clodius’s larval stage together, may lead to earlier plant senescence and lower P. clodius growth. Because Parnassius butterflies are well studied from a global perspective, our results may aid in understanding the potential vulnerabilities of other insect species in montane environments to climatic changes. Findings from this study demonstrate the value in assessing a butterfly species’ response to short- term weather variation or long-term climatic changes at each life stage in order to protect and conserve insects and their interactions with other organisms. Introduction Insects are responding to climate change globally (Parmesan et al. 1999, Parmesan and Yohe 2003, Diamond et al. 2011, Forister et al. 2021) – sparking a need to investigate how 9 climate affects insect population dynamics. It is becoming increasingly crucial to identify influential weather variables and climatic patterns that play a role in insect population change and distributional shifts (Roland and Matter 2016). The term weather refers to atmospheric conditions for a specific time and place changing hourly, daily, and seasonally, while climate is a statistical characterization of weather represented by means, variability, and extremes over many years (Moore et al. 2010). Once identified, these influential weather variables and climatic patterns may help in predicting population changes that could result from either short-term weather variation or long-term climatic changes (Roland and Matter 2016) and reflect greater changes occurring in the environment (Samways et al., 2010). Insects serve as bioindicators because (1) they are highly sensitive to local environmental conditions due to their small size, (2) they move towards optimal conditions due to mobility and reactivity to changing conditions, (3) their abundance is highly responsive to short- and long- term change due to short breeding rates and generation times, (4) they can be linked to specific environmental variables due to different growth rates, life history traits, body sizes, food preferences, and ecological preferences, and (6) they are relatively easy to survey (Samways et al., 2010). Montane insect species, i.e., those associated with mountain environments, can serve as bioindicators for the effects of climate change in higher elevations and latitudes, because their species distribution and abundance patterns reflect the subtle changing conditions of these high- elevation environments (Samways et al., 2010). A broad set of insect taxa have been documented to respond to climatic changes in recent decades. For instance, in response to changing environmental conditions (i.e., climate change & habitat modification), northern, cold-adapted moth species declined, while southern, warm- 10 adapted moth species grew in abundance in Great Britain over a 40-year period (Fox et al. 2014, Halsch et al. 2021). Warmer winters in the northeastern U.S. over 45 years drove an 83% decline in beetle abundance and a 39% decrease in number of beetle species captured, with reduced snow cover depths and duration associated with fewer beetles captured the following growing season (Harris et al. 2019, Halsch et al. 2021). Alpine grasshopper communities over 30 years in the northern Alps displayed expansion and increased densities of generalists, while specialists declined in response to warmer and dryer conditions (Illich and Zuna-Kratky 2022). A German headwater stream within a protected area experienced an 82% decline in aquatic insect abundance over 45 years associated with warmer water temperatures and altered discharge patterns (Baranov et al. 2020, Halsch et al. 2021). When occupancy and richness of 66 bumble bee species across North America and Europe were compared between 1901-1974 and 2000- 2014, occupancy declined on average by 46% in North America and 17% in Europe, while species richness declined on average by 20% in North America and 35% in Europe. Both bumble bee metrics declined relative to increasing frequency of high temperatures (Soroye et al. 2020). These long-term insect datasets demonstrate that insect distribution and abundance patterns, across a broad range of orders and families, can serve as bioindicators for climate change in montane ecosystems across various environmental stressors. Butterflies in particular are important bioindicators that can be used to monitor the effects of climate change over time (Kevan 1999, Diamond et al. 2011). As an easily surveyed, small ectothermic organism with a rapid generation time, butterflies are an ideal model organism to evaluate biodiversity and shifts in species distributions geographically in response to weather and climate variation (Kevan, 1999; Samways et al., 2010). Changes in butterfly population size over 11 time can be a strong signal of changes that have occurred or are occurring in their environment (Kevan 1999, Horn 2003). Suitable butterfly habitat is decreasing globally due to human land use intensification, landscape-scale degradation due to pollution, and climate change, leading to declining butterfly population sizes and shifting range distributions in North America, the U.K. and other locations (White and Kerr 2006, 2007, Thomas 2016, Forister et al. 2021). The abundances of butterfly species in North America and Europe have been declining during the last half century in response to habitat changes due to warming and drying (Forister et al. 2021), habitat changes from direct anthropogenic influences (Nakonieczny et al. 2007, Boggs and Inouye 2012), and distributions have shifted poleward and to higher elevations in response to declining suitable habitat (Parmesan et al. 1999, Hill et al. 2002, Parmesan and Yohe 2003, Condamine and Sperling 2018). Butterflies that persist in small, isolated populations with specific habitat requirements are at greater risk of decline or extinction compared to other butterfly species (Szcodronski et al. 2018). For example, the critically imperiled Fender’s blue butterfly (Icaricia icarioides fenderi) exists in small, isolated patches of Oregon prairie where <0.5% of its original habitat contains its larval host plant, Kincaid’s lupine (Lupinus sulphureus kincaidii) (Schultz 2001, Schultz and Crone 2005). Further, butterfly phenological responses are advancing three times stronger than their plant counterparts (Parmesan 2007) in response to climate change. Identification of influential weather variables and climatic patterns at each butterfly life stage is needed so that scientists and managers can prioritize butterfly conservation efforts during the most vulnerable life stage in the most threatened geographic locations. A study conducted over 17 years in two mountainous regions of the Czech Republic tested the effect of climatic 12 conditions during two larval phases (i.e., fall/spring activity and overwintering) on subsequent adult abundances of two Erebia butterflies (E. epiphron and E. sudetica) to evaluate which weather and climate variables drive population dynamics of these montane insects (Konvicka et al. 2021). Decreased precipitation and warmer temperatures during overwintering were associated with decreased adult abundances, while increased adult abundances were observed with warmer temperatures in the fall and warmer and drier springs during active larval phases (Konvicka et al. 2021). Additionally, higher minimum temperatures during the adult flight season were associated with decreased butterfly abundances (Konvicka et al. 2021). This research demonstrates that evaluating life stages may identify vulnerabilities of bioindicator insect species life cycles to weather patterns. A meta-analysis found a 1.6% annual reduction in the number of individual butterflies in the western US over the past 40 years in response to warmer fall temperatures, with single brood (univoltine) species declining more severely (Forister et al. 2021). Conversely, more butterflies were observed in response to warmer temperatures during summer months in some western US areas and in response to increased precipitation associated with positive effects on interacting plant species (Forister et al. 2021). These research projects detail the positive and negative effects that short-term weather variation or long-term climatic changes can have on butterfly species. Observational research conducted over 15 years in Alberta, Canada investigated the direct and indirect effects of various weather variables on sub- populations of Parnassius smintheus to predict their response to a changing climate (Matter et al. 2011, Filazzola et al. 2020). Direct effects included increased overwintering mortality in response to extreme minimum temperatures reaching -27.9°C and frequent loss of snow cover, and premature larval development under elevated temperatures, associated with earlier 13 emergence which negatively affected survival and population growth (Matter et al. 2011, Filazzola et al. 2020). Indirect effects included temperature and precipitation changes affecting growth and survival of host and nectar plant species limiting larvae and adult resources, thus affecting butterfly distribution (Matter et al. 2011, Filazzola et al. 2020). Due to the increasing frequency of short-term extreme weather events, such as extreme temperature, precipitation, and natural disasters (van Aalst 2006), and longer-term changes in variables such as minimum and maximum daytime and nighttime temperatures, annual snowpack, etc. (Whitlock et al. 2017, IPCC 2021), understanding these relationships will become even more important in the future. Parnassius butterflies serve as model species for insect conservation in montane ecosystems across the globe (Nakonieczny et al. 2007, Todisco et al. 2010). Parnassius are non- migratory butterflies, making them susceptible to short-term weather variation and the long-term effects of climate change. These butterflies live in local populations with limited fecundity, making their populations susceptible to change. Parnassius are restricted to high elevation montane meadows, environments predicted to be at greater risk and expected to undergo rapid change under a changing climate (Kim et al. 2002, Thuiller et al. 2005, Nakonieczny et al. 2007, Roland and Matter 2013). Further, these butterflies’ lifecycles are intricately dependent on the climatic patterns associated with seasonality in montane meadows. Changing climate patterns will likely lead to earlier larval emergence due to prolonged warm weather in the fall and/or spring (Crees and Debinski 2021), altering butterfly life stage timing inducing mismatch between larval emergence and host plant emergence and availability (Calabrese et al. 2008, Crees and Debinski 2021). 14 Here, we analyzed the population dynamics of a Parnassius clodius population in Grand Teton National Park, WY, USA from 2009 to 2021 in relation to weather and climate data to assess how this species is responding to changing climate in a montane region. Populations of P. clodius are well distributed in the Teton region, with 78% of suitable surveyed meadows occupied in 2013 (Szcodronski et al. 2018). However, many of the monitored populations had low numbers of individuals. Like many insect species, P. clodius population dynamics have yet to be evaluated in determining their conservation status. Parnassius clodius is currently not threatened or endangered in its native range of North America. However, it faces similar threats as its European sister species, Parnassius apollo, which is listed on the IUCN red list of threatened species due to habitat loss from forest encroachment, climatic changes, heavy metal contamination, over collection, and genetic erosion (Nakonieczny et al. 2007). By studying how P. clodius population dynamics are correlated with interannual variation in weather variables, we may be able to identify similar factors affecting other Parnassius species and predict how these factors may impact P. clodius in the future. The goal of this research was to evaluate how the interannual variation in population size of Parnassius clodius was correlated with weather and climate variables assumed to affect Parnassius butterflies within specific life stages. Weather conditions during the overwintering period (i.e., egg life stage) were hypothesized to have the greatest effect on population size compared to the larval/pupal and adult life stages based on previous findings evaluating population growth of Parnassius smintheus (Matter et al. 2011). 15 Materials and Methods Species Description Parnassius clodius (Lepidoptera: Papilionidae) is a medium-sized butterfly that is mostly white with a hairy yellow body and inhabits western Canada and U.S.A. (Figure 1). They are often found in open montane woodlands and meadows but can extend to higher elevations and reside on mountaintops. Their host-plant is steer’s head (Dicentra uniflora) (Fumariaceae) (Scott, 1986), a small herbaceous perennial and a spring ephemeral (Crees & Debinski, 2021; Scott, 1986). The primary nectar source of P. clodius in this region is sulphur flower buckwheat (Eriogonum umbellatum) (Auckland et al. 2004) along with several other flowering plant species (Szcodronski et al. 2018, Crees and Debinski 2021). Previous research revealed suspected positive associations between P. clodius occupancy and low sagebrush (Artemisia arbuscula) and Dicentra uniflora (i.e., P. clodius’s host plant), with the presence of big sagebrush (Artemisia tridentata) and lupine (Lupinus spp.) having a clear negative association with P. clodius occupancy (Szcodronski et al. 2018). Parnassius clodius is a univoltine species that overwinters as a pharate first instar larvae (i.e., egg) (Scott, 1986). Following spring snowmelt, the larvae emerge to feed and take two to three weeks to mature (Crees and Debinski 2021). Larvae form a chrysalis on the ground typically among rocks and debris to pupate over a one-to-two-week period. Typically, adults emerge from pupae in mid to late June with males emerging prior to females. Peak flight is usually late June to early July in Grand Teton National Park, although this can vary by several weeks, depending upon spring conditions (pers. obs.). Parnassius clodius females oviposit on vegetation, such as sagebrush and the undersides of balsamroot leaves, near host plants for the 16 remainder of the flight season, which often ends in mid-July (Crees and Debinski 2021). Microclimatic conditions influence the timing of when each life stage takes place (i.e., the phenological progression) (Crees and Debinski 2021). Study Area The study area (Pilgrim Creek) was a relatively homogenously vegetated sagebrush meadow with dry, gravelly soils surrounded by coniferous woodlands along the meadow edge in Grand Teton National Park, Wyoming, USA at 43.915915°N, -110.578850°W and 1500m X 300m in size. The meadow is relatively flat, and at an elevation of 2120 m. The vegetation community of the meadow included mostly low sagebrush (Artemisia arbuscula), some big sagebrush (Artemisia tridentata), several known P. clodius nectar plant species listed in the species description section, and P. clodius’ host plant – Dicentra uniflora. Butterfly Mark-Recapture Mark-recapture methods were used to study a population of Parnassius clodius at Pilgrim Creek across annual flight seasons between 2009 and 2018 during June and July. Surveys were not carried out in 2012 and 2013. Six 50m x 50m plots a minimum of 100m apart were located using GPS units, flagged prior to the flight season of P. clodius, and surveyed each year. Mark- recapture surveys began a few days after the beginning of the flight season and continued until only one or two butterflies per plot were caught during a survey period. Plots were monitored daily if weather permitted throughout each flight season. Surveys were conducted when temperatures were above 21°C, wind was <16kmh-1, and clouds were not obscuring the sun. If weather prevented all six plots from being sampled in one day, the rotation was completed the 17 following day. The order in which plots were surveyed rotated throughout the flight season to vary the time of day surveys were conducted for each plot. Two people surveyed each plot for 20 minutes between 10:00 and 17:00 hours. All Parnassius butterflies in a plot were netted by hand and held in glassine envelopes until the survey ended. The time of capture, sex, and activity of the butterfly when captured (i.e., nectaring, chasing, stationary) were noted. Captured butterflies were marked with a permanent marker on each hindwing with an identification number indicating the plot it was caught in and the butterfly’s individual number (Figure 1). The number corresponded to the consecutive butterfly count throughout a season. All butterflies were released in the center of each plot by placing butterflies on surrounding plants, rather than being placed on the ground where they could be stung by ants. Weather and Climate Variable Selection A list of meaningful weather variables and climatic patterns expected to affect populations of Parnassius clodius butterflies was formulated based upon previous research that identified influential weather and climate variables affecting the population dynamics of other Parnassius species (Matter et al. 2011, Roland and Matter 2013, 2016, Matter and Roland 2017) as well as our observations at Pilgrim Creek over time. The list of explanatory variables included: date of snowmelt, maximum snow water equivalent (cm), mean snow water equivalent (cm), mean temperature (C), mean maximum temperature (C), extreme maximum temperature (C), mean minimum temperature (C), extreme minimum temperature (C), accumulated rainfall (cm) from 2008 to 2018 (Table 1). All temperature data were sourced from the USDA Base Camp SNOTEL site located at 43.93333°N, -110.45000°W at 2151 m in elevation and 10.7 km 18 from Pilgrim Creek meadow. All snow water equivalent and precipitation data were sourced from the USDA Snake River Station SNOTEL site located at 44.13333°N, -110.66667°W at 2109 m in elevation and 25.1 km from Pilgrim Creek meadow. Analysis The goal of these analyses was to identify which weather variables, known to affect other butterfly species, explain variation in Parnassius clodius population growth through each life stage. Estimates of population change were calculated for the Pilgrim Creek population of P. clodius for the duration of the study (i.e., 2009-2011, 2014-2018). Calculations of population change (Rt) were estimated as the ratio of the number of caught butterflies (Nt) for one flight season after accounting for effort, relative to the number of caught butterflies of the previous flight season after accounting for effort. Effort was measured as the number of two-person 20- minute plot surveys conducted during each flight season. Rt was calculated as log10((Nt+1/effortt+1)/(Nt/effortt)) (Roland and Matter 2016). Our estimate of population size was included in each model to evaluate density dependence as an explanatory variable on population growth. Population size was estimated by taking the log of the number of caught butterflies divided by the number of surveys conducted for each flight season (i.e., log10(Nt/effortt)). A life stage hypothesis modeling approach was used to evaluate the effect of each chosen weather variable during each life stage on the population growth (i.e., Rt) of Parnassius clodius butterflies. The observed values for each weather variable were partitioned by life stage to evaluate each variable’s effect during each stage of the P. clodius lifecycle. Dates of each life stage were estimated from researcher observations (L. Crees, personal communication, October 2022) of P. clodius over time to capture the estimated beginning and end dates of each life stage. 19 We categorized the egg stage as July 21 - April 30, the larva-pupa stage as May 1 - May 31, and the adult stage June 1 - July 20. Each weather and climate variable was tabulated for the egg stage, larva-pupa stage, and the adult butterfly stage (Table 1). The larva and pupa life stages were combined due to the uncertainty of the date of when one stage ended and the other began. Correlation tests were conducted between all explanatory weather variables to prevent the inclusion of highly correlated variables in models, which could lead to multicollinearity and unreliable estimates. Uncorrelated precipitation and temperature explanatory variables and logNt (i.e., density dependence) were included in each generalized linear model. Then, generalized linear models using the glm function in R were fit to evaluate which explanatory weather variables were the most descriptive in explaining population change (Rt). AIC comparison between models was conducted to determine the most informative explanatory variables. Generalized linear models that included interactions between the most informative temperature and precipitation explanatory variables and logNt were used to evaluate potential interactive effects between weather and climate variables along with density dependence on population change. Final AIC comparison was conducted to determine the best model to describe population change for each P. clodius life stage. 20 Figure 1: Parnassius clodius butterfly. Photo Credit J. Simone Durney. Table 1: The nine explanatory weather and climate variables tabulated by each Parnassius clodius life stage included in analyses. Weather and Climate Variables Evaluated Julian date of snowmelt Maximum snow water equivalent (centimeters) Mean snow water equivalent (centimeters) Mean temperature (Celsius) Mean maximum temperature (Celsius) Extreme maximum temperature (Celsius) Mean minimum temperature (Celsius) Extreme minimum temperature (Celsius) Accumulated rainfall (centimeters) 21 Results A Parnassius clodius population was monitored from 2009-2018, excluding 2012 and 2013, producing eight interannual estimates of population change. During the study, 532 surveys were conducted and 3,868 P. clodius individual butterflies were caught. Population size and change (Rt) varied over the study with population increases occurring in 2011 and population declines occurring in 2009 (Figure 2). Egg Life Stage Correlation tests revealed that extreme maximum temperature was not highly correlated (i.e., -0.5< r <0.5) with the explanatory weather variables and thus was included in the full generalized linear model. Due to collinearity among the remaining weather variables, generalized linear models were constructed with population change (i.e., response variable) and each remaining explanatory weather variable (i.e., Julian date of snowmelt, maximum and mean snow water equivalent, mean temperature, mean maximum temperature, mean minimum temperature, extreme minimum temperature, accumulated rainfall) (Table 2). AIC comparison between those models was conducted to determine which weather variables were the most descriptive in explaining population change (Table 2). Thus, mean snow water equivalent, mean minimum temperature (Table 2) and extreme maximum temperature were the best descriptors of population change during the P. clodius egg life stage. However, including those weather variables with and without interactions along with previous populations size in generalized linear models resulted in no explanation of significant independent variation of population change during the egg life stage and did not significantly differ from the null model (F(3,7) = 4.0271; p = 0.1059) (Table 3). Further, including each descriptive weather variable in its own generalized 22 linear model with population change as the response revealed that none of the descriptive weather variables during the egg stage significantly explained population change (Figures 3 & 4). Larval-Pupal Life Stage All explanatory weather variables were highly correlated during the larval-pupal life stage. Therefore, generalized linear models were constructed with population change as the response variable and each explanatory weather variable. AIC comparison determined that accumulated rainfall and extreme maximum temperature were the best descriptors (i.e., lowest AIC value) (Table 2) of population change during the P. clodius larva-pupa life stages. The full model including all descriptive weather variables, their interactions, and density dependence resulted in no significant explanation of independent variation on population change. The same was true for the full model excluding interactions. Each descriptive weather variable was evaluated in its own generalized linear model with population change as the response. For the larval-pupal life stage, we found that extreme maximum temperature significantly explained population change (β = -0.095± 0.031(SE), t = -3.032, p = 0.023) (Figures 3 & 4; Table 3). Adult Life Stage Correlation tests determined that extreme minimum temperature during the adult stage was not highly correlated with the other explanatory weather variables and so was included in the full generalized linear model for the adult stage. Of the remaining variables, maximum snow water equivalent and extreme maximum temperature had the lowest AIC values during AIC comparison (Table 2). Thus, maximum snow water equivalent, extreme maximum temperature, and extreme minimum temperature were the best descriptors of population change during the P. 23 clodius adult life stage. However, including those weather variables with and without interactions along with population size in generalized linear models resulted in no explanation of significant independent variation of population change during the adult life stage and did not significantly differ from the null model (F(3,7) = 2.3194; p = 0.2170) (Table 3). Including each descriptive weather variable in its own generalized linear model with population change as the response revealed that none of the descriptive weather variables during the adult life stage significantly explained population change (Figures 3 & 4). Figure 2: Years of study with associated population change (Rt) values represented in black, no population change represented by the horizontal blue line, and density dependence (logNt) values represented in green. Note: surveys were not carried out in 2012 and 2013. 24 Figure 3: The relationship between Parnassius clodius population change (Rt) from 2009 to 2018 by life stage to the precipitation explanatory variables considered in these analyses. The egg life stage is represented by the color purple, the larval-pupal life stages are represented by the color green, and the adult life stage is represented by the color orange. Julian date of snowmelt did not vary across the life stages and is represented by the color black. Lines and points indicate a significant relationship between population change and a descriptive weather or climate variable at a particular life stage. Note: surveys were not carried out in 2012 and 2013. Overall, there were no significant relationships detected between population change (Rt) and date of snowmelt, maximum snow water equivalent (cm), mean snow water equivalent (cm), and accumulated rainfall (cm). 25 Figure 4: The relationship between Parnassius clodius population change (Rt) from 2009 to 2018 by life stage to the temperature explanatory variables considered in these analyses. The egg life stage is represented by the color purple, the larval-pupal life stages are represented by the color green, and the adult life stage is represented by the color orange. Darker, thicker lines and points indicate a significant relationship between population change and a descriptive weather variable at a particular life stage. Note: surveys were not carried out in 2012 and 2013. Overall, a significant negative relationship between extreme maximum temperatures during the larval- pupal life stage and population change (Rt) was detected. There were no significant relationships detected between population change (Rt) and mean temperature (C), mean maximum temperature (C), mean minimum temperature (C), and extreme minimum temperature (C). 26 Table 2: AIC comparison table for models testing for the best predictive weather and climate explanatory variables among highly correlated (i.e., r >0.5) variables to be included in full generalized linear models for each life stage (i.e., egg, larval-pupal, adult). The best predictive weather and climate variables for each life stage are in bold font. *Note: due to correlation between maximum and mean snow water equivalent, maximum snow water equivalent was chosen as the predictor variable during the adult life stage since the AIC values were the same between the two variable models. Life Stage Weather or Climate Variable Type Model df AIC Egg Precipitation Rt ~ Mean Snow Water Equivalent 3 -1.3919 Rt ~ Max Snow Water Equivalent 3 -0.7237 Rt ~ Julian Date of Snowmelt 3 -0.7088 Temperature Rt ~ Mean Minimum Temperature 3 -1.3306 Rt ~ Mean Maximum Temperature 3 -0.9940 Rt ~ Mean Temperature 3 -0.6472 Larval-Pupal Precipitation Rt ~ Accumulated Rainfall 3 -1.3621 Rt ~ Mean Snow Water Equivalent 3 -0.8933 Rt ~ Max Snow Water Equivalent 3 -0.7500 Rt ~ Julian Date of Snowmelt 3 -0.7088 Temperature Rt ~ Extreme Maximum Temperature 3 -8.0743 Rt ~ Mean Maximum Temperature 3 -1.5279 Rt ~ Mean Temperature 3 -1.1596 Rt ~ Mean Minimum Temperature 3 -0.7731 Rt ~ Extreme Minimum Temperature 3 -0.6695 Adult Precipitation Rt ~ Max Snow Water Equivalent* 3 -2.7788 Rt ~ Mean Snow Water Equivalent* 3 -2.7788 Rt ~ Accumulated Rainfall 3 -0.8856 Rt ~ Julian Date of Snowmelt 3 -0.7088 Temperature Rt ~ Extreme Maximum Temperature 3 -3.3465 Rt ~ Mean Maximum Temperature 3 -1.6039 Rt ~ Mean Temperature 3 -1.4181 Rt ~ Mean Minimum Temperature 3 -0.6702 27 Table 3: AIC comparison table for full generalized linear models including the best predictive weather and climate explanatory variables at each life stage (i.e., egg, larval-pupal, adult), their interactions, and density dependence on population change Rt. The best model for each life stage is listed in bold font. Life Stage Model df AIC Egg Rt ~ logNt + Mean Snow Water Equivalent + Mean Minimum Temperature 5 -7.7721 Rt ~ logNt + Extreme Maximum Temperature + Mean Minimum Temperature 5 -7.2273 Rt ~ logNt + Mean Snow Water Equivalent + Extreme Maximum Temperature + Mean Minimum Temperature 6 -6.2467 Rt ~ logNt + Mean Snow Water Equivalent*Extreme Maximum Temperature + Mean Minimum Temperature 7 -4.8884 Rt ~ logNt + Extreme Maximum Temperature + Mean Snow Water Equivalent*Mean Minimum Temperature 7 -4.2761 Rt ~ logNt + Mean Snow Water Equivalent*Extreme Maximum Temperature + Mean Snow Water Equivalent*Mean Minimum Temperature 8 -3.3665 Rt ~ logNt 3 -3.0856 Rt ~ 1 2 -2.6412 Rt ~ logNt + Mean Snow Water Equivalent + Extreme Maximum Temperature 5 -1.6514 Rt ~ Mean Snow Water Equivalent 3 -1.3919 Rt ~ Mean Minimum Temperature 3 -1.3306 Rt ~ Extreme Maximum Temperature 3 -0.8014 Rt ~ Extreme Maximum Temperature + Mean Minimum Temperature 4 0.1655 Rt ~ Mean Snow Water Equivalent + Extreme Maximum Temperature 4 0.1886 Rt ~ Extreme Maximum Temperature + Mean Snow Water Equivalent*Mean Minimum Temperature 6 0.2423 Rt ~ Mean Snow Water Equivalent + Mean Minimum Temperature 4 0.3441 Rt ~ Mean Snow Water Equivalent*Mean Minimum Temperature 5 1.3311 Rt ~ Mean Snow Water Equivalent + Extreme Maximum Temperature + Mean Minimum Temperature 5 1.7033 Rt ~ Mean Snow Water Equivalent*Extreme Maximum Temperature 5 2.0951 Rt ~ Mean Snow Water Equivalent*Extreme Maximum Temperature + Mean Minimum Temperature 6 3.5334 Likelihood Ratio Test of Null Model to Best Model F = 4.0271 3 0.1059 Larval-Pupal Rt ~ Extreme Maximum Temperature 3 -8.0743 Rt ~ Accumulated Rainfall + Extreme Maximum Temperature 4 -7.6590 Rt ~ Accumulated Rainfall*Extreme Maximum Temperature 5 -6.4838 Rt ~ logNt + Extreme Maximum Temperature 4 -6.0747 Rt ~ logNt + Accumulated Rainfall + Extreme Maximum Temperature 5 -5.6658 28 Rt ~ logNt + Accumulated Rainfall*Extreme Maximum Temperature 6 -4.7182 Rt ~ logNt 3 -3.0856 Rt ~ 1 2 -2.6412 Rt ~ logNt + Accumulated Rainfall 4 -1.8830 Rt ~ Accumulated Rainfall 3 -1.3621 Likelihood Ratio Test of Null Model to Best Model F = 9.1939 1 0.02303 Adult Rt ~ logNt + Extreme Minimum Temperature + Extreme Maximum Temperature 5 -4.7035 Rt ~ Maximum Snow Water Equivalent + Extreme Maximum Temperature 4 -4.3000 Rt ~ Maximum Snow Water Equivalent*Extreme Maximum Temperature 4 -4.3000 Rt ~ logNt + Maximum Snow Water Equivalent + Extreme Maximum Temperature 5 -3.8363 Rt ~ Extreme Maximum Temperature 3 -3.3465 Rt ~ logNt 3 -3.0856 Rt ~ Maximum Snow Water Equivalent 3 -2.7788 Rt ~ logNt + Maximum Snow Water Equivalent + Extreme Minimum Temperature + Extreme Maximum Temperature 6 -2.7225 Rt ~ logNt + Extreme Minimum Temperature + Maximum Snow Water Equivalent*Extreme Maximum Temperature 6 -2.7225 Rt ~ logNt + Maximum Snow Water Equivalent*Extreme Minimum Temperature + Extreme Maximum Temperature 6 -2.7225 Rt ~ logNt + Maximum Snow Water Equivalent*Extreme Minimum Temperature + Maximum Snow Water Equivalent*Extreme Maximum Temperature 6 -2.7225 Rt ~ 1 2 -2.6412 Rt ~ Extreme Minimum Temperature + Extreme Maximum Temperature 4 -2.4239 Rt ~ Maximum Snow Water Equivalent + Extreme Minimum Temperature + Extreme Maximum Temperature 5 -2.3274 Rt ~ Extreme Minimum Temperature + Maximum Snow Water Equivalent*Extreme Maximum Temperature 5 -2.3274 Rt ~ Maximum Snow Water Equivalent*Extreme Minimum Temperature + Extreme Maximum Temperature 5 -2.3274 Rt ~ Extreme Minimum Temperature 3 -1.0617 Rt ~ Maximum Snow Water Equivalent + Extreme Minimum Temperature 4 -0.8080 Rt ~ Maximum Snow Water Equivalent*Extreme Minimum Temperature 4 -0.8080 Rt ~ logNt + Maximum Snow Water Equivalent + Extreme Minimum Temperature 5 0.4432 Likelihood Ratio Test of Null Model to Best Model F = 2.3194 3 0.2170 29 Discussion Life stage hypothesis modelling of a Parnassius clodius population from 2009 to 2018 identified extreme maximum temperatures during the larval-pupal life stage as a mechanism driving population dynamics in the next generation of the population (Figures 4 & 5). These findings do not support our original hypothesis, that weather conditions during the overwintering period (i.e., egg life stage) would have the greatest effect on population size. However, they clearly identify that the larval-pupal life stages are the most sensitive to changing weather and climate conditions for the montane species Parnassius clodius. When extreme maximum temperatures reach 22.5°C or above in May (i.e., larval-pupal life stage) there is a noted decline in P. clodius population growth rates, while additional unaccounted factors must be contributing to the deviation from this pattern observed in 2018 (Figure 5). This finding suggests that the predicted increasing frequency of short-term extreme weather events (van Aalst 2006), especially in May, and long-term climatic changes to increasing temperatures (Hostetler et al. 2021, IPCC 2021) could lead to continual declining P. clodius populations. Related research investigating the response of population dynamics of the related non- migratory montane univoltine butterfly, Parnassius smintheus, identified direct and indirect effects of weather and climate on population growth rates. A field study showed that a quadratic interaction between extreme minimum temperature and absence of snow cover had the greatest direct negative effect on P. smintheus overwintering egg survival, with temperatures being too hot or cold and little to no snow negatively affecting population growth (Roland and Matter, 2016). This finding supported our hypothesis that weather and climate during the overwintering period would primarily affect population growth, however, our findings for P. clodius indicate 30 that weather conditions during the larval-pupal life stages have the greatest effect. In a rearing experiment using P. smintheus, elevated temperatures during larval development had a direct negative effect on survival and population growth due to early larval emergence (Matter et al. 2011). The rearing experiment detected similar findings to ours, with elevated temperatures during larval development having a negative effect on population growth. However, our research did not allow us to distinguish direct from indirect effects such as early host plant senescence on population dynamics. Increasing the years of observation and data on host plant phenology would aid in detecting a definitive relationship for P. clodius. From the perspective of developing a more generalized understanding of montane insect responses to climatic variation, we can summarize the responses of these two related butterfly species by saying that the indicator life stage of P. smintheus is the overwintering egg, whereas the indicator life stage of P. clodius is the larval-pupal life stages in spring. Extreme summer temperatures (i.e., hot and cold) are speculated to have direct negative effects on P. smintheus adult survival (Matter et al., 2011). These results are inconsistent with our findings, suggesting that population dynamics in response to weather conditions and climate change vulnerability may not be consistent across a genus. In fact, P. clodius and P. smintheus population dynamics responded to different weather and climate variables at different life stages. From the perspective of plant-insect interactions under varying climate, altered temperature and precipitation affecting growth and survival of host and nectar plant species limiting larvae and adult resources was identified as a negative indirect effect on Parnassius smintheus presence, distribution, and genetic structure (Matter et al. 2011, Filazzola et al. 2020). Similarly, research evaluating the reproductive success of the threatened univoltine butterfly, 31 Euphydryas editha bayensis, found that later adult emergence and early host plant senescence had direct and indirect negative effects on female reproductive success (Cushman et al., 1994; USFWS, 2023). Prediapause (larval life stage) starvation resulting from early host plant senescence was identified as the leading cause of E. editha bayensis mortality (Cushman et al., 1994; USFWS, 2023). The indirect effects of early host plant senescence on Euphydryas editha bayensis population dynamics are applicable to our findings, given that the timing of host plant emergence, senescence, and survival dictated larval and pupal survival and were reflected in the population dynamics of this univoltine butterfly species. Therefore, conservation of univoltine butterfly species should be directed towards evaluating the distribution, phenology, fecundity, and survival of a butterfly species at each life stage in addition to its nectar and host plant species (Demarse et al. 2023), especially when the butterfly is specific to one host plant species. The next steps to foster additional insights on P. clodius growth rates would include evaluating the distribution and phenology of its host plant, Dicentra uniflora, an early-spring ephemeral forb relative to climatic data that we used in assessing P. clodius population dynamics. The life history of D. uniflora has been examined in the southern extent of the Cascade Range in California, U.S.A., but little is known about its phenology in the Rocky Mountain Region in the Western U.S.A. (Schlising and Mackey 2020, Crees and Debinski 2021). Dicentra uniflora emergence and senescence has been observed to vary by up to two weeks at the same elevation depending on sun exposure within the region where P. clodius was evaluated in this research (Crees and Debinski 2021). No research has been conducted evaluating the temperature thresholds of D. uniflora (i.e., timing of emergence and senescence in response to temperature). This has direct implications for the research presented here as we speculate that increasing 32 temperatures during D. uniflora’s flowering duration, and P. clodius’s larvae life stage, could lead to early senescence thereby initiating larvae starvation and declining P. clodius larval-pupal survival. Ultimately, increasing our understanding of D. uniflora’s phenology and sensitivity to climatic variation will aid in conserving both D. uniflora plants and P. clodius butterflies, and both could serve as bioindicators of weather patterns or climatic change in montane ecosystems. Host plant life history and phenology affect the population change of numerous butterfly species, including P. clodius and P. smintheus. Parnassius smintheus’ host plant, Sedum lanceolatum, is a succulent species (Roland and Matter 2016) persisting throughout the summer months withstanding moderate drought and heat, while Dicentra uniflora, P. clodius’ host plant, is a spring ephemeral and much more vulnerable to drought and heat. The differences between these host plant’s life histories influence their responses to weather conditions and phenology ultimately affecting those species that interact and depend on them. For instance, the research presented here identified vulnerability using P. clodius population growth rates during the larval- pupal life stage in response to extreme heat. This could be the direct result of larvae and/or pupae being vulnerable to heat or the indirect result of host plant vulnerability leading to early senescence and larval starvation. We were unable to clearly identify at this time which metric was responding to extreme heat driving the declining population growth trends, however, we suspect an indirect effect of host plant vulnerability due to D. uniflora’s life history characteristics. Understanding host plant life histories and potential vulnerabilities may aid in identifying indicator life stages for butterfly species. Findings from this study and previous research demonstrate the value in assessing butterfly responses to short-term weather variation or long-term climatic changes at each life 33 stage in order to protect and conserve a species and its interactions with other organisms (Radchuk et al. 2013, Halsch et al. 2021, Buckley 2022). This research serves as a model for understanding more threatened Parnassius species and other butterfly species. Life stage hypothesis modelling would aid in identifying vulnerable life stages of threatened butterfly species, helping guide conservation efforts to preserve populations. Population predictions and conservation efforts based on a single life stage would likely incorrectly indicate that a species is not directly at risk of global climate change (Radchuk et al. 2013). Therefore, evaluation of all life stages is warranted to correctly identify vulnerability and an indicator life stage. Further, an understanding of host plant life histories along with data on host plant phenology in response to weather variations would provide guidance to conservation efforts. Additionally, it is pertinent to evaluate species responses during fall and spring when extreme fluctuations in weather occur most frequently and have been observed to have the greatest impact on species, especially montane butterflies (Matter et al. 2011). The research presented here, using life stage hypothesis modelling, determined extreme maximum temperatures during the larval-pupal life stages of Parnassius clodius led to a decline in the next generation (Figure 5). As exemplified in this research, evaluating an insect species’ response to external stressors by life stage reveals at which life stage conservation efforts should be focused in order to efficiently protect that species. 34 Figure 5: Parnassius clodius population change (Rt) from 2009 to 2018 relative to the extreme maximum air temperature of its environment during the larva-pupa life stages. Note: surveys were not carried out in 2012 and 2013. References Auckland, J.N., Debinski, D.M., Clark, W.R., 2004. Survival, movement, and resource use of the butterfly Parnassius clodius. Ecol. Entomol. 29, 139–149. https://doi.org/10.1111/j.0307-6946.2004.00581.x Baranov, V., Jourdan, J., Pilotto, F., Wagner, R., Haase, P., 2020. Complex and nonlinear climate driven changes in freshwater insect communities over 42 years. Conserv. Biol. 34, 1241– 1251. https://doi.org/10.1111/cobi.13477 35 Boggs, C.L., Inouye, D.W., 2012. A single climate driver has direct and indirect effects on insect population dynamics. Ecol. Lett. 15, 502–508. https://doi.org/10.1111/j.1461-0248.2012.01766.x Buckley, L.B., 2022. Temperature-sensitive development shapes insect phenological responses to climate change. Curr. Opin. Insect Sci. 52, 100897. https://doi.org/10.1016/j.cois.2022.100897 Calabrese, J.M., Ries, L., Matter, S.F., Debinski, D.M., Auckland, J.N., Roland, J., Fagan, W.F., 2008. Reproductive asynchrony in natural butterfly populations and its consequences for female matelessness. J. Anim. Ecol. 77, 746–756. https://doi.org/10.1111/j.1365-2656.2008.01385.x Condamine, F.L., Sperling, F.A.H., 2018. Anthropogenic threats to high-altitude parnassian diversity. News Lepid. Soc. 60, 94–99. Crees, L., Debinski, D.M., 2021. Natural History Observations on Parnassius clodius altaurus and Parnassius smintheus magnus (Papilionidae) in Grand Teton National Park , Wy. J. Lepid. Soc. 75, 70–74. Cushman, J.H., Boggs, C.L., Weiss, S.B., Murphy, D.D., Harvey, A.W., Ehrlich, P.R., 1994. Estimating female reproductive success of a threatened butterfly: influence of emergence time and hostplant phenology. Oecologia 99, 194–200. https://doi.org/10.1007/BF00317101 Demarse, A., Trendos, E., Linton, J., Flockhart, T., Brewster, A., Keyghobadi, N., Custode, L., Norris, R., 2023. Phenology, population size, and factors influencing variation in density of an endangered butterfly, the mottled duskywing Erynnis martialis. Endanger. Species Res. 50, 195–208. https://doi.org/10.3354/esr01230 Diamond, S.E., Frame, A.M., Martin, R.A., Buckley, L.B., 2011. Species’ traits predict phenological responses to climate change in butterflies. Ecology 92, 1005–1012. https://doi.org/10.2307/41739278 Filazzola, A., Matter, S.F., Roland, J., 2020. Inclusion of trophic interactions increases the vulnerability of an alpine butterfly species to climate change. Glob. Chang. Biol. 26, 2867–2877. https://doi.org/10.1111/gcb.15068 Forister, M.L., Halsch, C.A., Nice, C.C., Fordyce, J.A., Dilts, T.E., Oliver, J.C., Prudic, K.L., Shapiro, A.M., Wilson, J.K., Glassberg, J., 2021. Fewer butterflies seen by community scientists across the warming and drying landscapes of the American West. Science (80). 371, 1042–1045. https://doi.org/10.1126/science.abe5585 Fox, R., Oliver, T.H., Harrower, C., Parsons, M.S., Thomas, C.D., Roy, D.B., 2014. Long-term 36 changes to the frequency of occurrence of British moths are consistent with opposing and synergistic effects of climate and land-use changes. J. Appl. Ecol. 51, 949–957. https://doi.org/10.1111/1365-2664.12256 Halsch, C.A., Shapiro, A.M., Fordyce, J.A., Nice, C.C., Thorne, J.H., Waetjen, D.P., Forister, M.L., 2021. Insects and recent climate change. Proc. Natl. Acad. Sci. U. S. A. 118, 1–9. https://doi.org/10.1073/PNAS.2002543117 Harris, J.E., Rodenhouse, N.L., Holmes, R.T., 2019. Decline in beetle abundance and diversity in an intact temperate forest linked to climate warming. Biol. Conserv. 240, 108219. https://doi.org/10.1016/j.biocon.2019.108219 Hill, J.K., Thomas, C.D., Fox, R., Telfer, M.G., Willis, S.G., Asher, J., Huntley, B., 2002. Responses of butterflies to twentieth century climate warming: Implications for future ranges. Proc. R. Soc. B Biol. Sci. 269, 2163–2171. https://doi.org/10.1098/rspb.2002.2134 Horn, H.S., 2003. Butterflies as Bioindicators. Am. Sci. 91, 565–567. IPCC, 2021. Sixth Assessment Report of the Intergovernmental Panel on Climate Change. https://doi.org/10.1260/095830507781076194 Kevan, P.G., 1999. Pollinators as bioindicators of the state of the environment: Species, activity and diversity. Agric. Ecosyst. Environ. 74, 373–393. https://doi.org/10.1016/S0167-8809(99)00044-4 Kim, J., Kim, T.K., Arritt, R.W., Miller, N.L., 2002. Impacts of increased atmospheric CO2 on the hydroclimate of the western United States. J. Clim. 15, 1926–1942. https://doi.org/10.1175/1520-0442(2002)015<1926:IOIACO>2.0.CO;2 Konvicka, M., Kuras, T., Liparova, J., Slezak, V., Horázná, D., Klečka, J., Kleckova, I., 2021. Low winter precipitation, but not warm autumns and springs, threatens mountain butterflies in middle-high mountains. PeerJ 9, e12021. https://doi.org/10.7717/peerj.12021 Matter, S.F., Doyle, A., Illerbrun, K., Wheeler, J., Roland, J., 2011. An assessment of direct and indirect effects of climate change for populations of the Rocky Mountain Apollo butterfly (Parnassius smintheus Doubleday). Insect Sci. 18, 385–392. https://doi.org/10.1111/j.1744-7917.2011.01407.x Matter, S.F., Roland, J., 2017. Climate and extreme weather independently affect population growth, but neither is a consistently good predictor. Ecosphere 8. https://doi.org/10.1002/ecs2.1816 Nakonieczny, M., Kedziorski, A., Michalczyk, K., 2007. Apollo butterfly (Parnassius apollo L.) 37 in Europe - its history, decline and perspectives of conservation. Funct. Ecosyst. Communities 1, 56–79. Parmesan, C., 2007. Influences of species, latitudes and methodologies on estimates of phenological response to global warming. Glob. Chang. Biol. 13, 1860–1872. https://doi.org/10.1111/j.1365-2486.2007.01404.x Parmesan, C., Ryrholm, N., Stefanescu, C., Hill, J.K., Thomas, C.D., Descimon, H., Huntley, B., Kaila, L., Kullberg, J., Tammaru, T., Tennent, W.J., Thomas, J.A., Warren, M., 1999. Poleward shifts in geographical ranges of butterfly species associated with regional warming. Nature 399, 579–583. https://doi.org/10.1038/21181 Parmesan, C., Yohe, G., 2003. A globally coherent fingerprint of climate change impacts across natural systems. Nature 421, 37–42. https://doi.org/10.1038/nature01286 Radchuk, V., Turlure, C., Schtickzelle, N., 2013. Each life stage matters: The importance of assessing the response to climate change over the complete life cycle in butterflies. J. Anim. Ecol. 82, 275–285. https://doi.org/10.1111/j.1365-2656.2012.02029.x Roland, J., Matter, S.F., 2016. Pivotal effect of early-winter temperatures and snowfall on population growth of alpine Parnassius smintheus butterflies. Ecol. Monogr. 86, 412–428. https://doi.org/10.1002/ecm.1225 Roland, J., Matter, S.F., 2013. Variability in winter climate and winter extremes reduces popul