Thermal habitat of giant panda has shrunk by climate warming over the past half century

Biological Conservation 211 (2017) 125–133

Contents lists available at ScienceDirect

Biological Conservation journal homepage: www.elsevier.com/locate/biocon

Thermal habitat of giant panda has shrunk by climate warming over the past half century

MARK

Zhenhua Zanga,b, Guozhen Shenb,⁎, Guofang Renc,d, Cuiling Wangb, Chaoyang Fenge, Wenting Xub, Zongqiang Xieb, Quansheng Chenb, Xuyu Yangf, Junqing Lia,⁎⁎ a

The Key Laboratory for Silviculture and Conservation of Ministry of Education, Beijing Forestry University, Beijing 100083, China State Key Laboratory of Vegetation and Environmental Change, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, China c Century College, Beijing University of Posts and Telecommunications, Beijing 102613, China d School of Information Science and Technology, Beijing Forestry University, Beijing 100083, China e Chinese State Environmental Protection Key Laboratory of Regional Eco-process and Function Assessment, Chinese Research Academy of Environmental Sciences, Beijing 100012, China f Wildlife Conservation Division, Sichuan Forestry Department, Chengdu, Sichuan 610081, China b

A R T I C L E I N F O

A B S T R A C T

Keywords: Climate warming Potential heat stress Thermal habitat Giant panda

Climate warming is increasing the risks of extinction for many species. Giant panda is one of the most vulnerable mammals to climate warming due to its small population size and specialized diet of bamboo. Many studies have quantified projected habitat loss based on climate-change scenarios, but few have employed empirical data to investigate how the thermal habitat of giant panda has changed. In this study, we investigated the frequency, duration, and intensity of potential heat stress (PHS) occurrence that could surpass the biological threshold of giant panda by analyzing daily temperatures throughout the distribution range of giant panda from 1960 to 2010 and giant panda population survey data. We found an increase in the frequency of PHS25 (PHS above threshold of 25 °C) occurrence at a rate of 1.1–5.5 days/decade. The start date of PHS25 occurrence advanced at a rate of − 1.2 to −4.6 days/decade, while the end date of PHS25 occurrence was delayed at a rate of 0.8–3.0 days/ decade. The giant panda habitat is being exposed to an increased PHS occurrence. The area within reserves and densely populated giant panda habitat exposed to PHS occurrence expanded by 32–317% and 38–218%, respectively from the 1960s to the 2000s. Furthermore, PHS occurrence is intensifying; the annual accumulated degree-days of PHS25 and PHS30 occurrence (PHS above threshold of 30 °C) within the reserves increased by 39% and 140%, respectively. These results confirm that the potential threats to giant panda from climate warming are intensifying. It is urgent to expand the extent and range of giant panda habitat to allow giant pandas to move across landscapes in the face of ongoing climate warming.

1. Introduction The fifth assessment report of the Intergovernmental Panel on Climate Change demonstrated an increase in global land and ocean surface temperature of approximately 0.89 °C (0.69–1.08 °C) during the last century, and an increase of 1.0–4.0 °C is projected by the end of the 21st century (Stocker et al., 2013). Climate warming is affecting organisms by shifting their phenology and ranges, and causing habitat fragmentation and loss (Chen et al., 2011; Miller-Rushing and Primack, 2008; Parmesan and Yohe, 2003). Species can cope with ambient temperature increases by moving to more favorable locations, shifting the timing of life-history events, or adjusting their posture and

⁎

physiological functions (Bellard et al., 2012; Porter and Kearney, 2009). However, when ambient temperature increases beyond a species' suitable thermal range, there is a potential for species to become heat stressed, and when populations of a species suffer from prolonged exposure to potential heat stress (PHS), they face increased risk of extinction (Angilletta et al., 2010; Huey et al., 2012; McNab and Morrison, 1963). Multi-taxon reviews suggest that 20–30% of global plant and animal species could be at an increased risk of extinction due to climate warming (Parry et al., 2007), and continued warming will drive 16% species to extinction by 2100 (Urban, 2015). Tolerance to heat is largely conserved across lineages, and hard physiological boundaries exist that constrain evolution of tolerances of terrestrial

Correspondence to: Guozhen Shen, State Key Laboratory of Vegatation and Environmental Change, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, China Correspondence to: Junqing Li, The Key Laboratory for Silviculture and Conservation of Ministry of Education, Beijing Forestry University, Beijing 100083, China E-mail addresses: [email protected] (G. Shen), [email protected] (J. Li).

⁎⁎

http://dx.doi.org/10.1016/j.biocon.2017.05.011 Received 21 December 2016; Received in revised form 2 May 2017; Accepted 9 May 2017 0006-3207/ © 2017 Published by Elsevier Ltd.

Biological Conservation 211 (2017) 125–133

Z. Zang et al.

Fig. 1. Maps of A study area which spans six mountain ranges, and the rate of change in a the frequency of PHS25 occurrence; b the start date of PHS25 occurrence; and c the end date of PHS25 occurrence. Light grey circles show stations that had non-significant trends in the frequency of PHS25 occurrence, dark circles show stations that had non-significant trends in the start date of PHS25 occurrence, blue circles show stations that had non-significant trends in the end date of PHS25 occurrence; green circles show stations that had significant negative trends, and red circles show stations that had significant positive trends over time. Size of circles shows the magnitude of change rate. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

126

Biological Conservation 211 (2017) 125–133

Z. Zang et al.

heat stress, and related the temporal and spatial patterns of potential heat stress to the spatial distribution of the giant panda.

organisms to high temperatures (Araújo et al., 2013). Exposure to heat stress accelerates metabolic rate and evaporative water loss of endotherms, thus disrupting their energy and water balance (Huey et al., 2012). Increases in duration and intensity of heat stress will lead to catastrophic mortality of terrestrial endotherms (McKechnie and Wolf, 2010; Welbergen et al., 2008). Species with small populations and specialized trophic relationships are particularly vulnerable to heat stress (Fonseca, 2009; Pearson et al., 2014; Willi et al., 2006). Previous research has shown that genetic variation and potential response to natural selection should be positively correlated with population size (Blows and Hoffmann, 2005). Species with small populations are predicted to have reduced capacity to adapt to increasingly rapid warming (Armbruster and Reed, 2005), and to have lower resistance to heat stress owing to genetic limitations such as inbreeding (Reed, 2005). Specialist herbivores are predicted to be especially susceptible to climate warming because of their dependency on specific plants for food (Mattila et al., 2008). As the responses of herbivores and their prey plants to climate warming may differ, the main impacts of climate warming on herbivore populations will be mediated through their ability to synchronize their habitat and life history with that of their food and habitat resources (Parmesan, 2006). Although specialist herbivores are capable of escaping heat stress caused by climate warming, their population dynamics are closely associated with the distribution of their host plants (Marsico et al., 2009). Alterations to the specialized trophic relationships between prey and predators resulting from climate warming will generally have negative fitness consequences on predator populations (Lavergne et al., 2010; Schweiger et al., 2008; Thackeray et al., 2010). The small population size and specialized diet of the giant panda (Ailuropoda melanoleuca), one of the world's most endangered and recognized mammals, likely make the species extremely vulnerable to potential heat stress resulting from climate warming. Forest-dwelling species like the giant panda tend to favour cooler climates (Frishkoff et al., 2015; Hu, 2001; Pan et al., 2014). The giant panda inhabits a narrow maximum ambient temperature range of 24 to 27 °C throughout the year (Zhang et al., 2014). Ambient temperatures above 25 °C could cause increases in breathing rate to above normal levels and reductions in appetite and physical activity for giant pandas (Hu, 2001). Habitat loss and fragmentation have isolated giant pandas into only 33 populations, with 18 of these populations composed of < 10 individuals (State Forestry Administration, 2015). Fragmentation may limit giant panda dispersal and tracking of climate niches. Furthermore, giant pandas feed almost exclusively on bamboo (Schaller et al., 1985). The giant panda has evolved a suit of adaptations resulting in low energy expenditure to survive on this specialized diet (Nie et al., 2015a). Due to unusual extended sexual reproduction intervals and limited seed dispersal ability (Taylor and Qin, 1993), bamboo species are also vulnerable to climate change (Tuanmu et al., 2013; Li et al., 2015b), which may have negative additional consequences on giant pandas. Several studies have predicted that the climatically suitable habitats of giant panda would shrink and shift towards higher altitudes or latitudes by the end of the 21st century (Fan et al., 2014; Li et al., 2015a). Studies have also predicted that the distribution area of the staple food bamboo of giant panda would substantially shrink (Tuanmu et al., 2013; Li et al., 2015b). Furthermore, the projections have shown that giant panda habitats would greatly fragment from 2011 to 2100 under climate change scenarios (Shen et al., 2015). However, few studies have employed empirical data to investigate how the thermal habitat of giant panda has changed and whether climate warming already poses a potential threat to giant pandas, yet these are crucial for giant pandas conservation in an era of rapid warming. To explore the effects of recent climate warming on the thermal habitat of giant panda, we first investigated the frequency and duration of potential heat stress over the last half century (1960–2010). Then, we explored the intensity, and the temporal and spatial shifts of potential

2. Data and methods 2.1. Study area and data sources The distribution range of giant panda spans six mountains (Qinling, Minshan, Qionglai, Daxiangling, Xiaoxiangling, and Liangshan) along the eastern edge of the Qinghai-Tibetan Plateau in China (Hu, 2001). It encompasses 51 administrative counties in Sichuan, Shanxi, and Gansu provinces, with an area of 164,411 km2 (27°50′–34°19′N, 101°4′–108°56′E) and elevations between 260 and 7500 m (Fig. 1A). We obtained giant panda occurrence data from the third (1999–2003) national survey on giant pandas (State Forestry Administration, 2006). The third national survey used a global positioning system, remote sensing, and geographic information systems to identify the distribution of giant panda. A stratified random sample of 2 km2 grids was laid out across the distribution range of giant panda. At least one continuous ≥0.75 km long U-shaped or Z-shaped transect was established on each grid, and > 9000 transect lines were surveyed. Giant panda occurrences along the transects were identified using the bamboo stem fragment (BSF) method (State Forestry Administration, 2006). According to the BSF method, individual giant pandas were distinguishable through the geographical separation distance and bite size of bamboo stem fragments between their scats. If the geographical separation distance between each of two scats exceeded the distance threshold (i.e., 1.0 km for scats that were defecated within 1 day, 1.5 km for 1–3 days, 2.5 km for 4–15 days, and 3.5 km for > 15 days, respectively), they were considered to be from different giant pandas (State Forestry Administration, 2006). When the geographical separation distance between each of two scats was not beyond the distance threshold, if the difference in the average length of bamboo stem fragments between two scats exceeded 2 mm, the scats were assumed to be excreted by different giant pandas; otherwise, the scats were considered to be from the same individual (State Forestry Administration, 2006). The correct distinguishing ratio between wild giant panda individuals according the technique was 90.5% (Yin et al., 2005). Since the fourth national survey (2011 − 2013) data on the giant panda was not officially released, we developed a density distribution map of giant panda by performing a point density analysis on giant panda occurrence data based on the third national survey using the spatial analyst tools in ArcGIS 10.1. To facilitate visual representation of our results, we presented the density distribution for the grids with giant panda density > 0.015 individuals/km2 (Fig. S1). We then defined the densely populated giant panda habitat as the densest 20% of the above grids. By 2015, sixty-seven nature reserves had been established for giant panda conservation, and we obtained the boundaries of these reserves from China's State Forestry Administration (2015). We extracted the elevation data of the distribution range of giant panda from a digital elevation model (DEM) at a spatial resolution of 90 × 90 m from the National Geomatics Center of China (2015). We obtained daily maximum, minimum, and mean surface air temperature (Tmax, Tmin, and Tmean, < 1% of the data was missing) at meteorological stations throughout the distribution range of giant panda during 1960–2010 from the China Meteorological Administration (2016). Temperatures were recorded daily at 02:00, 08:00, 14:00, 20:00. Daily Tmean was the mean of these records, and daily Tmax and Tmin was the record at 14:00 and 02:00, respectively. We filled data gaps spanning up to seven consecutive days using a simple linear interpolation algorithm. When the data gaps spanned more than seven consecutive days, we filled the gaps using stepwise regressions based on data from stations with no missing data during the closet 5 years. We used the short-cut Bartlett test to examine the homogeneity of the daily Tmax, Tmin, and Tmean series for each station (Hasanean and 127

Biological Conservation 211 (2017) 125–133

Z. Zang et al.

Basset, 2006). Finally, we removed data from five stations due to nonhomogeneity, resulting in a final temperature dataset of daily Tmax, Tmin, and Tmean at 42 stations with elevations ranging from 291 to 3728 m.

(for stations where the frequency of PHS25 occurrence was < 30 days per year, the start date and the end date of PHS25 occurrence was inconsistent and not consecutive, and these stations were not considered in further analysis). Finally, we tested the significance of the trends in the time series of the frequency, start, and end date of PHS25 occurrence using the Mann-Kendall method, and calculated the change rates using the cumulative method. We performed all statistical analyses in R.

2.2. Definition of potential heat stress to giant panda The ambient temperature at the onset of behavioural thermoregulation (panting, soaking, and postural adjustments), along with preferred ambient temperatures species experienced in the wild, could be useful proxies to estimate thermal performance curves of endotherms (Childress and Letcher, 2017; Huey et al., 2012). Based on the behavioural and physiological response of giant panda to heat environment, Wang (2007) proposed that the stressful ambient temperature for captive giant pandas was between 23.2 and 29.5 °C (increased behavioural thermoregulation such as postural adjustments, panting, and soaking have been observed when giant panda individuals were exposed to these temperatures). When ambient temperature rises above 25 °C, the breathing rate of giant pandas has been shown to significantly increase above normal levels, to > 70 breaths per minute (Hu, 2001), and the Na+ content of giant panda blood serum has been shown to precipitously decline (Wang, 2007). These physiological effects likely explain observed reductions in the appetite and physical activity of giant pandas at high ambient temperatures (Hu, 2001). Therefore, we defined 25 °C as the threshold temperature of potential heat stress (PHS25) for the giant panda. We defined the frequency of PHS25 occurrence as the total number of the days where daily Tmax exceeds 25 °C in a year. The start date of PHS25 occurrence was defined as the median date of the initial 5% of the days (which ranged from 2 to 10 days, depending on the year) with daily Tmax exceeding 25 °C in a year, and the end date of PHS25 occurrence as the median date of the final 5% of the days with daily Tmax exceeding 25 °C in a year (Shen et al., 2016).

2.4. Temperature interpolation analyses To interpolate the Tmax of each station into the distribution range of giant panda, we calculated the lapse rate of Tmax (the rate at which the air temperature drops with elevation increase) within the distribution range of giant panda during 1960–2010 using the linear lapse rate adjustment method (Dodson and Marks, 1997). The linear lapse rate adjustment method is appropriate for temperature interpolation because it is computationally feasible and accounts for elevation effects (Dodson and Marks, 1997). We used three sets of lapse rates to interpolate Tmax for each month (April to October; Appendix S1). We assessed the accuracy of lapse rates through cross-validated interpolation of Tmax and by comparing the mean absolute interpolation error of Tmax among the three sets of lapse rates (Appendix S1). For each month, the set of lapse rates with the lowest mean absolute interpolation errors were interpreted to be the most accurate. Using the most accurate lapse rates, we interpolated the decadal monthly Tmax (the arithmetic mean of the daily Tmax in a specified month during the 1960s or the 2000s) and the decadal daily Tmax (the arithmetic mean of the Tmax on a specified day during the 1960s or the 2000s) of each station into the distribution range of giant panda. Then we prepared 7 maps of the decadal monthly Tmax, and 214 maps of the decadal daily Tmax for the 1960s and 2000s, to quantify the area and intensity of potential heat stress experienced by giant pandas in the area.

2.3. Time series analyses of temperatures and potential heat stress 2.5. Spatial distribution of the accumulated degree-days of potential heat stress

We applied the Mann-Kendall method to determine whether statistically significant temporal trends existed for daily temperatures (Tmax, Tmin, and Tmean) between 1960 and 2010. The nonparametric Mann-Kendall method has been widely used for identifying time series trends in climatological data (Xu et al., 2008). We derived the daily temperatures of the entire distribution range of giant panda from the arithmetic means of the daily temperatures of the 42 stations. We calculated the change rates of the daily temperatures between 1960 and 2010 using the cumulative method (Haynes et al., 2014). The cumulative method can minimize the effects of seasonal amplitudes and daily random fluctuations on estimates of climate variables (Haynes et al., 2014). The change rates calculated by the cumulative method show a higher degree of accuracy than those calculated by the simple linear regression method (Haynes et al., 2014). The cumulative daily temperature of a specific date between 1960 and 2000 was calculated by summing the daily temperature from the first date of 1960 to the specific date. We calculated the cumulative daily temperature for every date across the period of 1960–2010. Then, we performed a second-power polynomial regression analysis on the time series of the cumulative daily temperature. Finally, we estimated the change rate of daily temperature by calculating the second derivative of the polynomial. For a specific station, if the number of days where daily Tmax exceeded 25 °C was > 1% of the total number of days from 1960 to 2010, PHS25 was considered to have occurred at that station. From 1960 to 2010, PHS25 occurred in 38 of the 42 stations. We calculated the frequency of PHS25 occurrence for the 38 stations. Then, we identified 31 stations where the frequency of PHS25 occurrence exceeded 30 days in every year between 1960 and 2010, and calculated the start date and the end date of PHS25 occurrence for the 31 stations

We quantified the total area of potential heat stress based on the 7 maps of the decadal monthly Tmax in the 1960s and the 2000s. To account for the potential cooling effects of forests, we investigated threshold temperatures ranging from 25 to 30 °C (i.e., PHS25, PHS26, PHS27, PHS28, PHS29, PHS30). We defined potential heat stress (PHS25–PHS30) occurrence areas within the distribution range of giant panda as the grids with the decadal monthly Tmax above 25–30 °C, respectively. We extracted the distribution maps of potential heat stress for each month, and overlapped the distribution maps of potential heat stress with the density distribution map of giant panda and the boundaries of nature reserves. We extracted the potential heat stress occurrence areas within the entire distribution range of giant panda, the nature reserves, and the densely populated giant panda habitat between the 1960s and the 2000s, respectively. Since the area that experienced Tmax above 25 °C was negligible (< 0.2% of the area of the distribution range of giant panda) in April and October, we only included potential heat stress occurrence area results for May, June, July, August, and September. We defined the upper elevation of potential heat stress occurrence as the median of the top 5% of the elevations where potential heat stress occurred, and the median elevation of potential heat stress occurrence as the median of the elevations where potential heat stress occurred. Then, we extracted the upper elevation and the median elevation of potential heat stress occurrence separately for each month in the 1960s and the 2000s from the elevation surface of the DEM. Because the elevations were not normally distributed, we used a Wilcoxon test to compare the upper elevation and median elevation of potential heat stress occurrence between the 1960s and the 2000s for 128

Biological Conservation 211 (2017) 125–133

Z. Zang et al.

82–162% during the 40 years from the 1960s to the 2000s (Fig. 4, Table S2). The annual accumulated degree-days of PHS25 and PHS30 occurrence in the 2000s was 23% and 53% higher than the annual accumulated degree-days of PHS25 and PHS30 occurrence in the 1960s, respectively (Table S2). The ratio of the annual accumulated degreedays of PHS30 occurrence to the annual accumulated degree-days of PHS25 occurrence within the distribution range of giant panda increased from 25.1% to 31.1% from the 1960s to the 2000s, suggesting a trend of intensified potential heat stress associated with climate warming (Table S2). The monthly accumulated degree-days of PHS25 and PHS30 occurrence within the nature reserves increased by 58–137% and 93–693%, respectively during the 40 years from the 1960s to the 2000s (Table S2). The annual accumulated degree-days of PHS25 and PHS30 occurrence within the reserves increased by 39% and 140%, respectively, from the 1960s to the 2000s. Additionally, the ratio of the annual accumulated degree-days of PHS30 occurrence to the annual accumulated degree-days of PHS25 occurrence within the reserves increased from 8.9% to 15.3% from the 1960s to the 2000s (Table S2).

each month. Based on the 214 maps of the decadal daily Tmax for the 1960s and the 2000s, we quantified the accumulated degree-days of PHS25 and PHS30 occurrence for each day and month and the entire year, respectively. We calculated the sum of the differences between the decadal daily Tmax and 25 °C (for PHS25) or 30 °C (for PHS30) in each grid of the map with the decadal daily Tmax above 25 °C (or 30 °C) for each day. As all spatial analyses were performed at a resolution of 500 m × 500 m, we then multiplied the daily summed differences by 0.25 to present the daily accumulated degree-days of PHS25 or PHS30 occurrence at a spatial level of 1 km2. Finally, we calculated the monthly and annual accumulated degree-days of PHS25 or PHS30 occurrence by summing the daily accumulated degree-days of PHS25 or PHS30 occurrence for each month and the entire year, respectively. We performed all of these spatial analyses in ArcGIS 10.1. 3. Results 3.1. The frequency and duration of potential heat stress

4. Discussion

The daily Tmax, Tmin, and Tmean of the distribution range of giant panda increased significantly (p < 0.001) at the rate of 1.6 °C/ 100 years, 2.0 °C/100 years, and 1.5 °C/100 years, respectively, between 1960 and 2010. The daily Tmax increased significantly (p < 0.05) at the rate of 0.3–4.0 °C/100 years at 38 of the 42 stations analysed (Fig. S2 a); the daily Tmin increased significantly (p < 0.05) at the rate of 0.3–5.6 °C/100 years at 36 stations (Fig. S2 b); and the daily Tmean increased significantly (p < 0.05) at the rate of 0.7–3.7 °C/ 100 years at 35 stations (Fig. S2 c). Among the 38 stations where PHS25 occurred between 1960 and 2010, the frequency of PHS25 occurrence increased significantly (p < 0.05) at the rate of 1.1–5.5 days/decade at 28 stations (Fig. 1a). Among the 31 stations where PHS25 occurred every year, the start date of PHS25 occurrence advanced significantly (p < 0.05) at the rate of −1.2 to − 4.6 days/decade at 16 stations (Fig. 1b), and the end date of PHS25 occurrence was delayed significantly (p < 0.05) at the rate of 0.8–3.0 days/decade at 13 stations (Fig. 1c).

The population size and habitat of the giant panda have rebounded over the past decades (State Forestry Administration, 2015). Due to substantial uncertainty resulting from lack of geographic closure and genotyping errors, it was unwarranted that giant panda numbers based on DNA estimation was double that obtained using the BSF technique, and that there could be 3000 giant pandas in the wild based on DNA estimation (Garshelis et al., 2008). The BSF technique may substantially underestimate giant panda numbers in dense populations (Garshelis et al., 2008), but based on the BSF method, the fourth nationwide census counted 1864 giant pandas in the wild, which means that there has been a 17% rise in its population since the third national survey (State Forestry Administration, 2015). The giant panda is not a species at an “evolutionary dead end”, however, it has suffered demographically at the hands of human pressure (B.W. Zhang et al., 2007a), and is still facing high extinction risk from habitat fragmentation and climate warming (Shen et al., 2015). In this study, we have detected accelerated warming over the past half century throughout the distribution range of giant panda, and have shown the thermal habitat of giant panda is being encroached by the expansion and intensification of potential heat stress occurrence, suggesting that the giant panda is facing increasing extinction risk from climate warming. Prolonged and intensified potential heat stress occurrence could cause metabolic disorders and dehydration, and ultimately decrease the fitness of giant pandas. When endotherms are exposed to heat stress, they must maintain their body functions at the expense of increased energy and water consumption (McNab and Morrison, 1963). Species exposed to increased duration of heat stress are more likely to experience fitness reductions (Oswald et al., 2011). And extreme hot weather can cause catastrophic mortality (McKechnie and Wolf, 2010; Welbergen et al., 2008). Our findings show that the frequency of PHS25 occurrence increased significantly at the rate of 1.1–5.5 days/decade from 1960 to 2010, and that the annual accumulated degree-days of PHS25 and PHS30 occurrence increased by 23% and 53%, respectively from the 1960s to the 2010s. Giant pandas maintain a low energy expenditure to match their specialized diet of bamboo (Nie et al., 2015a), and this behaviour might make them more vulnerable to the prolongation and intensification of potential heat stress occurrence. When giant pandas are exposed to prolonged and intensified potential heat stress, they must increase energy and water consumption, which might result in their metabolic disorders and dehydration. The increased duration and intensity of potential heat stress occurrence may threaten reproductive activities and offspring survival of giant pandas. Previous studies have shown that increases in ambient temperature would force a reduction in foraging time of large mammals

3.2. Area and elevation shifts in potential heat stress The total area of PHS25–PHS30 occurrence within the distribution range of giant panda was 332–54,529 km2 in the 1960s and 1532–62,332 km2 in the 2000s (Table 1). The total area of PHS25–PHS30 occurrence expanded by 14–1250% from the 1960s to the 2000s (Fig. 2a, Table 1). The total area of PHS25–PHS30 occurrence within the giant panda nature reserves was 0–4368 km2 in the 1960s and 18–6106 km2 in the 2000s (Fig. 3, Table 1). And the total area of PHS25–PHS30 occurrence within the reserves expanded by 32–317% from the 1960s to the 2000s (Table 1). Furthermore, the total area of PHS25–PHS28 occurrence within the densely populated giant panda habitat expanded by 38–218% in June and July from the 1960s to the 2000s (Table 1). The upper elevation of PHS25–PHS30 occurrence was 752–2106 m in the 1960s and 1058–2270 m in the 2000s (Fig. S3, Table S1). The median elevation of PHS25–PHS30 occurrence was 461–1181 m in the 1960s and 471–1274 m in the 2000s (Fig. S3, Table S1). The upper elevation of PHS25–PHS30 occurrence shifted upwards significantly (p < 0.05) by 21–389 m, while the median elevation shifted upwards significantly (p < 0.05) by 10–118 m from the 1960s to the 2000s (Fig. 2b and c, Table S1). 3.3. The intensity of potential heat stress The monthly accumulated degree-days of PHS25 occurrence within the distribution range of giant panda increased by 29–99%, while the monthly accumulated degree-days of PHS30 occurrence increased by 129

Biological Conservation 211 (2017) 125–133

Z. Zang et al.

Table 1 Total area of potential heat stress (PHS) occurrence from the 1960s to the 2000s. Month

May

Jun.

Jul.

Aug.

Sep.

PHS

PHS25 PHS26 PHS27 PHS25 PHS26 PHS27 PHS28 PHS29 PHS30 PHS25 PHS26 PHS27 PHS28 PHS29 PHS30 PHS25 PHS26 PHS27 PHS28 PHS29 PHS30 PHS25 PHS26

Area in the distribution range (km2)

Area in the nature reserves (km2)

Area in the densely populated habitat (km2)

1960s

2000s

shift

1960s

2000s

shift

1960s

2000s

shift

12,237 3892 332 35,875 27,018 18,252 9236 3947 1199 54,589 44,268 34,194 25,039 16,667 8387 48,302 38,063 28,517 20,102 12,125 3957 6457 670

21,246 13,064 4482 43,124 33,180 24,125 15,394 6290 1532 62,332 52,559 42,326 32,422 23,497 15,141 46,485 36,112 26,703 18,466 10,726 1595 10,644 4011

74% 236% 1250% 20% 23% 32% 67% 59% 28% 14% 19% 24% 29% 41% 81% −4% −5% −6% −8% −12% −60% 65% 499%

269 69 18 2276 1234 585 206 43 0 4368 2749 1586 842 356 100 3369 2002 1089 516 177 35 56 19

703 288 69 3005 1778 942 410 133 18 6106 4199 2671 1580 850 368 3000 1802 1011 484 162 29 141 46

161% 317% 283% 32% 44% 61% 99% 209% / 40% 53% 68% 88% 139% 268% −11% −10% −7% −6% −8% −17% 152% 142%

0 0 0 402 158 42 0 0 0 911 452 172 49 0 0 627 280 90 17 0 0 0 0

16 0 0 555 227 64 0 0 0 1447 837 406 156 40 0 439 172 47 7 0 0 0 0

/ / / 38% 44% 52% / / / 59% 85% 136% 218% / / − 30% − 39% − 48% − 59% / / / /

Percent of shift = (shift in area of PHS occurrence from the 1960s to the 2000s) / area of PHS occurrence in the 1960s.

thermal habitat is available (Kearney et al., 2009). We found that the area of PHS25–PHS30 occurrence within the distribution range of giant panda increased by 14–1250% from the 1960s to the 2000s, and the area of PHS25–PHS28 occurrence is penetrating into the areas inhabited by dense giant panda populations. Increases in the area of potential heat stress occurrence would reduce the area and suitability of thermal habitat for giant pandas. Furthermore, it has been predicted that the area of giant panda habitat would decrease in the future under climate warming (Li et al., 2015a; Shen et al., 2015). Therefore giant pandas are forced to withstand potential heat stress encroachment due to a lack of alternative suitable habitat. The upward shift in elevation of potential heat stress occurrence might increase the fragmentation of the giant panda habitat. Many species have shifted their distributions to higher elevations in response

(Aublet et al., 2009; Bourgoin et al., 2011), and enhance their requirements of selection for cool habitat (Pigeon et al., 2016; van Beest et al., 2012). We found that the duration and intensity of potential heat stress occurrence increased over time, and this increase has the potential to limit the time that giant pandas can safely forage and perform other activities during their mating season (spring) and birth season (autumn). Heat stress can lead to disruptions in reproductive processes in mammals (Hansen, 2009). If giant pandas cannot find suitable microhabitats to escape from potential heat stress, they may forgo reproduction, abandon cubs, or suffer from higher cub mortality (Z.J. Zhang et al., 2007b). The expansion of potential heat stress occurrence would lead to thermal habitat loss and degradation for giant pandas. Species can escape environmental stresses through migration only if suitable

Fig. 2. Comparisons between a the total area; b the upper elevation; and c the median elevation of potential heat stress occurrence in the 1960s and the 2000s. The results for PHS25, PHS26, PHS27, PHS28, PHS29 and PHS30 occurrence are shown in black, green, blue, orange, red, and purple lines, respectively. Solid lines show results for the 1960s and dotted lines show results for the 2000s. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

130

Biological Conservation 211 (2017) 125–133

Z. Zang et al.

Fig. 3. Maps of the dynamics of the potential heat stress occurrence areas from the 1960s to the 2000s. The first and second rows of maps show the dynamics of PHS25 occurrence areas. Light grey patches show PHS25 occurrence areas that remained constant; coral pink patches show PHS25 occurrence areas that increased; green patches show PHS25 occurrence areas that decreased. The third row of maps shows the dynamics of PHS30 occurrence areas. The dark grey, purple, and verdelite green patches, respectively show PHS30 occurrence areas that remained constant, increased, and decreased. The density distribution map of giant panda and nature reserve boundaries are also shown. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

tion and nutrient content of bamboo (Nie et al., 2015b; Schaller et al., 1985). Climate warming could result in a prolonged growing season for bamboo species, which would promote bamboo regeneration and ease the food shortages for giant pandas, especially in the spring (Li, 1997). However, many bamboo species are vulnerable to climate change, and the limited seed and vegetative dispersal ability of bamboo will impede their expansion to climatically suitable areas (Taylor and Qin, 1993; Tian, 1989). This impeded expansion would limit the responsiveness of bamboo distributions to the rapidly changing climate. In addition, the intensification of potential heat stress occurrence within the distribution range of giant panda may induce earlier maturation of bamboo, or even trigger bamboo flowering (Hu et al., 1997; Wang, 2006). Bamboo species die soon after flowering, and large-scale bamboo flowering has been disastrous to the giant panda (Feng, 1991). Substantial reductions to the distribution ranges of the three dominant bamboo species in the

to climate warming (Chen et al., 2011). The minimum elevation of giant panda habitat has been predicted to rise by 500 m in Qinling by the end of the 21st century (Fan et al., 2014). Our findings show that the upper elevation and median elevation of PHS25–PHS30 occurrence shifted upwards significantly by 21–389 m and 10–118 m, respectively from the 1960s to the 2000s. The rising elevation of potential heat stress occurrence are likely forcing giant pandas to move to higher altitudes (Pan et al., 2014), resulting in increased energy expenditure to forage in the more rough and steep terrain at high altitudes (Wei et al., 2015). Also, mountainous areas usually taper with increasing altitude (Körner, 2007), therefore, giant pandas are likely to experience severe range contractions and population isolation with their upward migration. Giant pandas are strict dietary specialists that feed only on bamboo, and their population dynamics are closely associated with the distribu-

131

Biological Conservation 211 (2017) 125–133

Z. Zang et al.

Fig. 4. Accumulated degree-days (ADD) of PHS25 (upper plot with solid lines) and PHS30 (lower plot with dashed lines) occurrence within the distribution range of giant panda in the 1960s (grey lines) and the 2000s (red lines). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

warming is already occurring throughout the distribution range of giant panda. Giant panda habitat is being encroached by potential heat stress, and the occurrence of potential heat stress is intensifying. The findings of this study are informative for management and conservation of giant pandas in an era of rapid warming, and can also be applied to other endangered species that are vulnerable to climate change.

Qinling Mountains have been predicted to occur by the end of the 21st century, which would severely limit food availability for the giant panda (Li et al., 2015b; Tuanmu et al., 2013). It is urgent to mitigate the potential adverse effects of climate warming on the giant panda. Appropriate microhabitats can efficiently buffer the detrimental effects of potential overheating (Hannah et al., 2014; Isaac et al., 2008). We found that the area of potential heat stress is encroaching into nature reserves and the densely populated giant panda habitat, such as in the Qinling, eastern Minshan, and eastern Liangshan Mountains (Fig. 3). Therefore, it is imperative to establish refuge of microhabitats such as cool retreats and maternity dens within the habitats encroached by potential heat stress. Refuge microhabitats for giant pandas should be constructed close to water with narrow entrances and roomy interior chambers, where giant pandas can shield themselves from potential heat stress and minimize the risks of heatrelated illnesses to cubs (Z.J. Zhang et al., 2007b). Simulated natural forests can alleviate the adverse effects of climate warming by providing suitable tree hollows and favorable canopy closure for giant pandas (Liu and Viña, 2014; Zhang et al., 2011). We found that the northwestern part of the Minshan and western Liangshan Mountains was far away from potential heat stress areas (Fig. 3). Studies have predicted that the northwestern Minshan and western Liangshan Mountains will continue to be climatically suitable for the giant panda (Li et al., 2015a; Li et al., 2015b; Shen et al., 2015). Therefore, it is necessary to restore forests within the potential suitable habitats of giant panda in the northwestern part of the Minshan and western Liangshan Mountains, and plant staple bamboo species under the forests where the canopy is closed. Protected areas form an important component of conservation strategies to cope with negative effects caused by climate warming (Hannah et al., 2007). New nature reserves are needed in the northwestern part of the Minshan and western Liangshan Mountains to maintain connectivity among suitable habitat patches, and facilitate giant panda migration and colonization among the isolated habitat patches (Li et al., 2015a; Shen et al., 2015). Due to a lack of long-term giant panda tracking data, we cannot validate our estimated climate-related range shifts of the giant panda over the past half century. Further long-term field monitoring of giant panda migration and bamboo phenology and growth are required to understand how climate change will impact giant panda and bamboo species. Despite these uncertainties, our results confirm that climate

Acknowledgements This research was funded by the National Key Research and Development Program of China (2016YFC050330304), National Natural Science Foundation of China (31170500) and Basic Work of Science and Technology (2015FY1103002). We are grateful to the Research Groups on the Biodiversity and Ecological Conservation in the State Key Laboratory of Vegetation and Environmental Change. We thank the three anonymous referees for constructive suggestions to the manuscript. We would also like to thank Christine Verhille at the University of British Columbia for her assistance with English language and grammatical editing. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.biocon.2017.05.011. References Angilletta, M.J., Cooper, B.S., Schuler, M.S., Boyles, J.G., 2010. The evolution of thermal physiology in endotherms. Front. Biosci. E2, 861–881. Araújo, M.B., Ferri-Yáñez, F., Bozinovic, F., Marquet, P.A., Valladares, F., Chown, S.L., 2013. Heat freezes niche evolution. Ecol. Lett. 16, 1206–1219. Armbruster, P., Reed, D.H., 2005. Inbreeding depression in benign and stressful environments. Heredity 95, 235–242. Aublet, J.F., Festa-Bianchet, M., Bergero, D., Bassano, B., 2009. Temperature constraints on foraging behaviour of male Alpine ibex (Capra ibex) in summer. Oecologia 159, 237–247. van Beest, F.M., Van Moorter, B., Milner, J.M., 2012. Temperature-mediated habitat use and selection by a heat-sensitive northern ungulate. Anim. Behav. 84, 723–735. Bellard, C., Bertelsmeier, C., Leadley, P., Thuiller, W., Courchamp, F., 2012. Impacts of climate change on the future of biodiversity. Ecol. Lett. 15, 365–377. Blows, M.W., Hoffmann, A.A., 2005. A reassessment of genetic limits to evolutionary change. Ecology 86, 1371–1384. Bourgoin, G., Garel, M., Blanchard, P., Dubray, D., Maillard, D., Gaillard, J.M., 2011. Daily responses of mouflon (Ovis gmelini musimon × Ovis sp.) activity to summer

132

Biological Conservation 211 (2017) 125–133

Z. Zang et al.

endothermic species. J. Biogeogr. 38, 430–438. Pan, W.S., Lü, Z., Zhu, X.J., Wang, D.J., Wang, H., Long, Y., Fu, D.L., Zhou, X., 2014. In: McShea, W.J., Harris, R.B., Garshelis, D.L., Wang, D.J. (Eds.), A Chance for Lasting Survival. Smithsonian Institution Scholarly Press, Washington, DC. Parmesan, C., 2006. Ecological and evolutionary responses to recent climate change. Annu. Rev. Ecol. Evol. Syst. 37, 637–669. Parmesan, C., Yohe, G., 2003. A globally coherent fingerprint of climate change impacts across natural systems. Nature 421, 37–42. Parry, M.L., Canziani, O.F., Palutikof, J.P., Van der Linden, P.J., Hanson, C.E., 2007. Climate Change 2007: Impacts, Adaptation and Vulnerability. Cambridge University Press, Cambridge, UK. Pearson, R.G., Stanton, J.C., Shoemaker, K.T., Aiello-Lammens, M.E., Ersts, P.J., Horning, N., Fordham, D.A., Raxworthy, C.J., Ryu, H.Y., McNees, J., Akcakaya, H.R., 2014. Life history and spatial traits predict extinction risk due to climate change. Nat. Clim. Chang. 4, 217–221. Pigeon, K.E., Cardinal, E., Stenhouse, G.B., Côté, S.D., 2016. Staying cool in a changing landscape: the influence of maximum daily ambient temperature on grizzly bear habitat selection. Oecologia 181, 1101–1116. Porter, W.P., Kearney, M., 2009. Size, shape, and the thermal niche of endotherms. Proc. Natl. Acad. Sci. U. S. A. 106, 19666–19672. Reed, D.H., 2005. Relationship between population size and fitness. Conserv. Biol. 19, 563–568. Schaller, G.B., Hu, J.C., Pan, W.S., Zhu, J., 1985. The Giant Panda of Wolong. University of Chicago Press, Chicago, USA. Schweiger, O., Settele, J., Kudma, O., Klotz, S., Kühn, I., 2008. Climate change can cause spatial mismatch of trophically interacting species. Ecology 89, 3472–3479. Shen, G.Z., Pimm, S.L., Feng, C.Y., Ren, G.F., Liu, Y.P., Xu, W.T., Li, J.Q., Si, X.F., Xie, Z.Q., 2015. Climate change challenges the current conservation strategy for the giant panda. Biol. Conserv. 190, 43–50. Shen, M.G., Piao, S.L., Chen, X.Q., An, S., Fu, Y.S.H., Wang, S.P., Cong, N., Janssens, I.A., 2016. Strong impacts of daily minimum temperature on the green-up date and summer greenness of the Tibetan Plateau. Glob. Chang. Biol. 22, 3057–3066. State Forestry Administration, 2006. The Third National Survey Report on Giant Panda in China. China Science Press, Beijing, China. State Forestry Administration, 2015. Report of the fourth national survey on giant panda in China. http://www.wwfchina.org/specialdetail.php?pid=205&page=13 accessed November 18, 2015. Stocker, T.F., Qin, D., Plattner, G., Tignor, M., Allen, S., Boschung, J., Nauels, A., Xia, Y., Bex, V., Midgley, P., 2013. Climate Change 2013: The Physical Science Basis. Cambridge University Press, Cambridge, UK. Taylor, A.H., Qin, Z.S., 1993. Structure and dynamics of bamboos in the Wolong Natural Reserve, China. Am. J. Bot. 80, 375–384. Thackeray, S.J., Sparks, T.H., Frederiksen, M., Burthe, S., Bacon, P.J., Bell, J.R., Botham, M.S., Brereton, T.M., Bright, P.W., Carvalho, L., Clutton-Brock, T., Dawson, A., Edwards, M., Jones, I.D., Jones, J.T., Leech, D.I., Roy, D.B., Scott, W.A., Smith, M., Smithers, R.J., Winfield, I.J., Wanless, S., 2010. Trophic level asynchrony in rates of phenological change for marine, freshwater and terrestrial environments. Glob. Chang. Biol. 16, 3304–3313. Tian, X.Q., 1989. Shooting and growth of Bashania fargesii. J. Bamboo Res. 8, 400–407. Tuanmu, M.N., Viña, A., Winkler, J.A., Li, Y., Xu, W.H., Ouyang, Z.Y., Liu, J.G., 2013. Climate-change impacts on understorey bamboo species and giant pandas in China's Qinling Mountains. Nat. Clim. Chang. 3, 249–253. Urban, M.C., 2015. Accelerating extinction risk from climate change. Science 348, 571–573. Wang, X.H., 2006. Study on the influence of environmental factors on bamboo bloom. Sichuan J. Zool. 28, 618–621. Wang, Y.C., 2007. The Thermal Microclimate Effects on the Thermoregulation Behaviour and Physiology Bio-chemical Indexes of Captive Giant Pandas. Sichuan Agriculture University, Chengdu, China. Wei, W., Nie, Y.G., Zhang, Z.J., Hu, Y.B., Yan, L., Qi, D.W., Li, X.H., Wei, F.W., 2015. Hunting bamboo: foraging patch selection and utilization by giant pandas and implications for conservation. Biol. Conserv. 186, 260–267. Welbergen, J.A., Klose, S.M., Markus, N., Eby, P., 2008. Climate change and the effects of temperature extremes on Australian flying-foxes. Proc. R. Soc. B 275, 419–425. Willi, Y., Van Buskirk, J., Hoffmann, A.A., 2006. Limits to the adaptive potential of small populations. Annu. Rev. Ecol. Evol. Syst. 37, 433–458. Xu, Z.X., Gong, T.L., Li, J.Y., 2008. Decadal trend of climate in the Tibetan Plateau—regional temperature and precipitation. Hydrol. Process. 22, 3056–3065. Yin, Y.F., Wang, H., Chen, A., Liu, G.Q., 2005. Accuracy assessment of the bamboo stem fragment distinguishing method used in panda population surveys. Divers. Sci. 13, 439–444. Zhang, B.W., Li, M., Zhang, Z.J., Goosens, B., Zhu, L.F., Zhang, S.N., Hu, J.C., Bruford, M., Wei, F.W., 2007a. Genetic viability and population history of the giant panda, putting an end to the “evolutionary dead end”? Mol. Biol. Evol. 24, 1801–1810. Zhang, Z.J., Swaisgood, R.R., Wu, H., Li, M., Yong, Y.G., Hu, J.C., Wei, F.W., 2007b. Factors predicting den use by maternal giant pandas. J. Wildl. Manag. 71, 2694–2698. Zhang, Z.J., Swaisgood, R.R., Zhang, S.N., Nordstrom, L.A., Wang, H.J., Gu, X.D., Hu, J.C., Wei, F.W., 2011. Old-growth forest is what giant pandas really need. Biol. Lett. 7, 403–406. Zhang, Z.J., Sheppard, J.K., Swaisgood, R.R., Wang, G., Nie, Y.G., Wei, W., Zhao, N.X., Wei, F.W., 2014. Ecological scale and seasonal heterogeneity in the spatial behaviors of giant pandas. Integr. Zool. 9, 46–60.

climatic conditions. Can. J. Zool. 89, 765–773. Chen, I.C., Hill, J.K., Ohlemüller, R., Roy, D.B., Thomas, C.D., 2011. Rapid range shifts of species associated with high levels of climate warming. Science 333, 1024–1026. Childress, E.S., Letcher, B.H., 2017. Estimating thermal performance curves from repeated field observations. Ecology 98, 1377–1387. China Meteorological Administration, 2016. http://data.cma.cn (accessed June 25, 2016). Dodson, R., Marks, D., 1997. Daily air temperature interpolated at high spatial resolution over a large mountainous region. Clim. Res. 8, 1–20. Fan, J.T., Li, J.S., Xia, R., Hu, L.L., Wu, X.P., Li, G., 2014. Assessing the impact of climate change on the habitat distribution of the giant panda in the Qinling Mountains of China. Ecol. Model. 274, 12–20. Feng, W.H., 1991. Research papers on giant panda. J. Sichuan Univ. 3, 7–13. Fonseca, C.R., 2009. The silent mass extinction of insect herbivores in biodiversity hotspots. Conserv. Biol. 23, 1507–1515. Frishkoff, L.O., Hadly, E.A., Daily, G.C., 2015. Thermal niche predicts tolerance to habitat conversion in tropical amphibians and reptiles. Glob. Chang. Biol. 21, 3901–3916. Garshelis, D.L., Wang, H., Wang, D.J., Zhu, X.J., Li, S., McShea, W.J., 2008. Do revised giant panda population estimates aid in their conservation? Ursus 19, 168–176. Hannah, L., Midgley, G., Andelman, S., Araújo, M., Hughes, G., Martinez-Meyer, E., Pearson, R., Williams, P., 2007. Protected area needs in a changing climate. Front. Ecol. Environ. 5, 131–138. Hannah, L., Flint, L., Syphard, A.D., Moritz, M.A., Buckley, L.B., McCullough, I.M., 2014. Fine-grain modeling of species' response to climate change: holdouts, stepping-stones, and microrefugia. Trends Ecol. Evol. 29, 390–397. Hansen, P.J., 2009. Effects of heat stress on mammalian reproduction. Philos. Trans. R. Soc. B 365, 3341–3350. Hasanean, H., Basset, H.A., 2006. Variability of summer temperature over Egypt. Int. J. Climatol. 26, 1619–1634. Haynes, M.A., Kung, K.J.S., Brandt, J.S., Yang, Y.P., Waller, D.M., 2014. Accelerated climate change and its potential impact on Yak herding livelihoods in the eastern Tibetan plateau. Clim. Chang. 123, 147–160. Hu, J.C., 2001. Research on the Giant Panda. Shanghai Publishing House of Science and Technology, Shanghai, China. Hu, Y., Xiang, W.G., Zhu, K.Y., Lin, R.Z., Wang, J.X., Wang, W.J., 1997. Study of meteorological conditions of the bamboo growth and bloom. J. Chengdu Inst. Meteorol. 12, 225–233. Huey, R.B., Kearney, M.R., Krockenberger, A., Holtum, J.A., Jess, M., Williams, S.E., 2012. Predicting organismal vulnerability to climate warming: roles of behaviour, physiology and adaptation. Philos. Trans. R. Soc. B 367, 1665–1679. Isaac, J., De Gabriel, J., Goodman, B., 2008. Microclimate of daytime den sites in a tropical possum: implications for the conservation of tropical arboreal marsupials. Anim. Conserv. 11, 281–287. Kearney, M., Shine, R., Porter, W.P., 2009. The potential for behavioral thermoregulation to buffer “cold-blooded” animals against climate warming. Proc. Natl. Acad. Sci. U. S. A. 106, 3835–3840. Körner, C., 2007. The use of ‘altitude’ in ecological research. Trends Ecol. Evol. 22, 569–574. Lavergne, S., Mouquet, N., Thuiller, W., Ronce, O., 2010. Biodiversity and climate change: Integrating evolutionary and ecological responses of species and communities. Annu. Rev. Ecol. Evol. Syst. 41, 321–350. Li, C.B., 1997. The Research of Staple Food Bamboo for Giant Panda. Guizhou Science and Technology Publishing House, Guiyang, China. Li, R.Q., Xu, M., Wong, M.H.G., Qiu, S., Li, X.H., Ehrenfeld, D., Li, D.M., 2015a. Climate change threatens giant panda protection in the 21st century. Biol. Conserv. 182, 93–101. Li, R.Q., Xu, M., Wong, M.H.G., Qiu, S., Sheng, Q.K., Li, X.H., Song, Z.M., 2015b. Climate change-induced decline in bamboo habitats and species diversity: implications for giant panda conservation. Divers. Distrib. 21, 379–391. Liu, J.G., Viña, A., 2014. Pandas, plants, and people. Ann. Mo. Bot. Gard. 100, 108–125. Marsico, T.D., Hellmann, J.J., Romero-Severson, J., 2009. Patterns of seed dispersal and pollen flow in Quercus garryana (Fagaceae) following post-glacial climatic changes. J. Biogeogr. 36, 929–941. Mattila, N., Kotiaho, J.S., Kaitala, V., Komonen, A., 2008. The use of ecological traits in extinction risk assessments: a case study on geometrid moths. Biol. Conserv. 141, 2322–2328. McKechnie, A.E., Wolf, B.O., 2010. Climate change increases the likelihood of catastrophic avian mortality events during extreme heat waves. Biol. Lett. 6, 253–256. McNab, B.K., Morrison, P., 1963. Body temperature and metabolism in subspecies of Peromyscus from arid and mesic environments. Ecol. Monogr. 33, 63–82. Miller-Rushing, A.J., Primack, R.B., 2008. Global warming and flowering times in Thoreau's concord: a community perspective. Ecology 89, 332–341. National Geomatics Center of China, 2015. http://ngcc.sbsm.gov.cn (accessed Novenber 25, 2015). Nie, Y.G., Speakman, J.R., Wu, Q., Zhang, C.L., Hu, Y.B., Xia, M.H., Yan, L., Hambly, C., Wang, L., Wei, W., 2015a. Exceptionally low daily energy expenditure in the bambooeating giant panda. Science 349, 171–174. Nie, Y.G., Zhang, Z.J., Raubenheimer, D., Elser, J.J., Wei, W., Wei, F.W., 2015b. Obligate herbivory in an ancestrally carnivorous lineage: the giant panda and bamboo from the perspective of nutritional geometry. Funct. Ecol. 29, 26–34. Oswald, S.A., Huntley, B., Collingham, Y.C., Russell, D.J.F., Anderson, B.J., Arnold, J.M., Furness, R.W., Hamer, K.C., 2011. Physiological effects of climate on distributions of

133