Phenological response to climate change in China - Wiley Online Library

Global Change Biology Global Change Biology (2015) 21, 265–274, doi: 10.1111/gcb.12648

Phenological response to climate change in China: a meta-analysis Q U A N S H E N G G E 1 , H U A N J I O N G W A N G 1 , 2 , T H I S R U T I S H A U S E R 3 and J U N H U D A I 1 1 Institute of Geographical Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing, China, 2University of Chinese Academy of Sciences, Beijing, China, 3Oeschger Centre for Climate Change Research (OCCR) and Institute of Geography, University of Bern, Bern, Switzerland

Abstract The change in the phenology of plants or animals reflects the response of living systems to climate change. Numerous studies have reported a consistent earlier spring phenophases in many parts of middle and high latitudes reflecting increasing temperatures with the exception of China. A systematic analysis of Chinese phenological response could complement the assessment of climate change impact for the whole Northern Hemisphere. Here, we analyze 1263 phenological time series (1960–2011, with 20+ years data) of 112 species extracted from 48 studies across 145 sites in China. Taxonomic groups include trees, shrubs, herbs, birds, amphibians and insects. Results demonstrate that 90.8% of the spring/summer phenophases time series show earlier trends and 69.0% of the autumn phenophases records show later trends. For spring/summer phenophases, the mean advance across all the taxonomic groups was 2.75 days decade1 ranging between 2.11 and 6.11 days decade1 for insects and amphibians, respectively. Herbs and amphibians show significantly stronger advancement than trees, shrubs and insect. The response of phenophases of different taxonomic groups in autumn is more complex: trees, shrubs, herbs and insects show a delay between 1.93 and 4.84 days decade1, while other groups reveal an advancement ranging from 1.10 to 2.11 days decade1. For woody plants (including trees and shrubs), the stronger shifts toward earlier spring/summer were detected from the data series starting from more recent decades (1980s–2000s). The geographic factors (latitude, longitude and altitude) could only explain 9% and 3% of the overall variance in spring/summer and autumn phenological trends, respectively. The rate of change in spring/summer phenophase of woody plants (1960s–2000s) generally matches measured local warming across 49 sites in China (R = 0.33, P < 0.05). Keywords: China, climate change, meta-analysis, phenology, season, trend Received 21 March 2014 and accepted 27 May 2014

Introduction Global mean surface temperature has increased by 0.85 °C over the period 1880–2012 (IPCC, 2013). Especially over the period 1951–2012, the global combined land and ocean instrumental temperature records show an increasing trend of 0.12 °C decade1 (IPCC, 2013). However, the rate of warming over the past 15 years (1998–2012; 0.05 °C decade1) is significantly smaller than the rate calculated since 1951 (IPCC, 2013). During this so-called warming hiatus (Kosaka & Xie, 2013), observational data show a continued increase in hot extremes over land (Seneviratne et al., 2014). The changes both in extreme temperatures and mean temperature have already affected the physical and biological systems on all continents (Rosenzweig et al., 2007a) with very high confidence that living systems have been influenced by anthropogenic warming over the Correspondence: Huanjiong Wang, tel. 86-10-64889831, fax 86-10-64854230, e-mail: [email protected]; Junhu Dai, tel. 86-10-64889066, fax 86-10-64872274, e-mail: [email protected]

© 2014 John Wiley & Sons Ltd

last three decades (Rosenzweig et al., 2008). Regarding terrestrial ecosystems, the overwhelming majority of studies reveal consistent responses to warming trends, e.g. poleward and elevational range shifts of flora and fauna, the earlier onset of spring events and changes in abundance of certain species (Parmesan & Yohe, 2003; Rosenzweig et al., 2007a). A growing body of literature reviewed the response to warming of phenological change across the northern hemisphere (Walther, 2003; Badeck et al., 2004; Linderholm, 2006; Richardson et al., 2013). Several meta-analyses have summarized the coherent phenological response to climate change. Parmesan & Yohe (2003) analyzed phenological change in 677 species reported in literature regardless of the magnitude of change, while Root et al. (2003) restricted their metaanalysis to the species showing significant phenological changes. Consequently, the mean magnitude of advance in timing of spring events estimated by Parmesan & Yohe (2003) of 2.3 days decade1 was weaker than the estimated 5.1 days decade1 of Root et al. (2003). This discrepancy is primarily due to the 265

266 Q . G E et al. difference between the studies in criteria for incorporating data (Parmesan, 2007). For Europe, Menzel et al. (2006) conducted a systematic meta-analysis which includes phenological network data for 542 available species. Working Group II of the Fourth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC AR4), assessed observed changes and responses in natural and managed systems through synthesis of studies (Rosenzweig et al., 2007a). In addition, the multiple phenological changes among trophic levels may cause the temporal mismatch in key trophic interactions (Both et al., 2009). For addressing this problem, the difference in rate of phenological change among different trophic levels and functional groups has been assessed by using meta-analysis method (Thackeray et al., 2010). However, these meta-analyses mainly focused on the evidence from Europe and North America. Only one study about China’ phenological change (Schwartz & Chen, 2002) was included in Root et al. (2003). IPCC AR4 included no study about phenological change in China for global synthesis assessment (Rosenzweig et al., 2007b). A systematic analysis of Chinese phenological response is needed for a complete assessment of phenological change across the Northern Hemisphere. Even more so, China has broad geographical coverage and is one of the countries with the richest biodiversity in the world (Ministry of Environmental Protection of China, 2008). Moreover, China accumulated lots of phenological observation data in recent several decades through two nationwide observation networks (Chen, 2013). The data from these observation networks were published in lots of literature during the last decade mostly in Chinese (see Appendix S1) and are, thus, only accessible by native speakers. One previous meta-analysis investigated the change in Chinese spring phenology of 72 species covering the period 1980s–2000s (Ma & Zhou, 2012). However, they only examined data for herbaceous and woody plants. To the best of our knowledge, a systematic analysis of all available plant and animal phenological observations in China does not exist to date. Here, we present an exhaustive China-wide analysis of all observed phenological changes from 1960s to 2000s. We investigate whether the taxonomic groups, study periods and locations influence the magnitude of phenological response to climate warming.

Materials and methods The abundant data set about Chinese phenophases was systematically collected in two phenological observation networks (Chen, 2013). The first one is Chinese Phenological

Observation Network (CPON) established in 1963, administrated by Institute of Geographical Sciences and Natural Resources Research (IGSNRR), Chinese Academy of Sciences (CAS). In 1980, the China Meteorological Administration (CMA) established another countrywide phenological network affiliated with the national agrometeorological monitoring network. The results from 48 studies using the data from these observation networks were analyzed (Appendix S1). The selection criteria for inclusion in this meta-analysis were the following: 1. The studies had to cover at least 20 years of data between 1960 and 2011; 2. the studies must use the slope of regression between phenophase and years to measure the phenological trend; 3. the studies only focusing on one-phase, one-site and single species data were excluded to minimize expected positive publishing bias (studies showing no impact of climate change are unlikely to be published); 4. the studies reporting trends on averaged time series were excluded as they may replicate with studies that reported each species separately. For each study meeting the above criteria, we extract the information consisting of species name, phenophase (the observable life cycle stage of a specific plant or animal), the season then in occured, start year and end year of time series, location of observation site (latitude, longitude, altitude), the linear trend (standardized to days decade1) and the statistical significance of trend (Appendix S2). If there are more than one time series of same phenophase at a single site, we only retain the time series with longest records. As a result, a total of 112 species, including 65 trees, 22 shrubs, 17 herbs, 5 birds, 1 amphibian and 2 insects, were investigated (Table 1). The 1263 phenological time series (869 spring/summer phenophases and 394 autumn phenophases) of these species were distributed across 145 sites in China (Fig. 1, see Appendix S3 for the distribution of sites separately for each taxon). Most of the time series started from 1960 to 1965 or 1980 to 1985 (Fig. 2a). The end years of time series are all between 2000 and 2011 (Fig. 2a). As a result, the lengths of time series ranged from 20 to 50 years with a mean of 35.6 years (Fig. 2b). We used a boxplot method to depict differences in spring/ summer and autumn phenological trends among six taxonomic groups. In addition, we tested the mean trends for each group by using one-way analysis of variance (ANOVA) and Fisher’s LSD multiple comparison (The R Core Team, 2014). Second, for verifying the assumption that a stronger phenological trend would be expected in more recent decades, we firstly divided all time series of woody plants (including trees and shrubs, because there are no significant differences in trends between them through preanalysis) into three groups according to the start year (1960s, 1970s, 1980s). The end years are not considered as a basis for grouping because they are all distributed in recent decade (2001–2011, Fig. 2a). Subsequently, we compared the phenological trends across the above-mentioned three groups using boxplot and ANOVA methods. Third, we calculated the mean trend of all time series for

© 2014 John Wiley & Sons Ltd, Global Change Biology, 21, 265–274

P H E N O L O G I C A L R E S P O N S E I N C H I N A 267 Table 1 Summary of datasets in the meta-analyses of Chinese phenological changes Taxon

Phases

Season

Nts

Nspecies

Nsites

Start year

Tree

BB,FL,50L,FF,50F,EF,FM BLC,ELC, BLF, ELF BB,FL,50L,FF,50F,EF,FM BLC,ELC, BLF, ELF BB,FL,50L,FF,50F,EF,FM BTB,50TB, ETB AD,FC DD,LC FC LC FC LC

Sp/su Au Sp/su Au Sp/su Au Sp/su Au Sp/su Au Sp/su Au Sp/su Au

632 265 85 29 127 79 14 10 4 4 7 7 869 394

65 51 22 20 17 17 5 3 1 1 2 2 112 94

126 88 16 7 50 38 8 7 4 4 5 5 145 108

1971 1971 1968 1969 1983 1982 1981 1981 1984 1984 1982 1982 1973 1974

Shrub Herb Bird Amphibian Insect Overall

9 10 8 9 2 3 1 1 4 4 1 1 9 10

End year 2008 2007 2009 2009 2005 2005 2005 2005 2007 2007 2005 2005 2008 2007

3 2 2 2 2 2 3 3 3 3 3 3 3 3

Length (years) 38.4 36.9 43.1 41.2 23.8 24.2 24.9 24.7 24.0 24.0 23.7 23.7 36.3 34.0

11.2 11.6 9.1 10.0 2.2 3.4 2.5 2.8 2.4 2.4 2.3 2.3 11.6 11.6

Season: Sp/su, spring/summer; Au, autumn. Nts: number of time series. Nspecies: number of species. Nsites: number of sites. Phases: FF, First flowering; 50F, 50% of full flowering; EF, end of flowering; FL, first leaf; 50L, 50% of full leaf expansion; BB, bud burst; FM, fruit maturity; FD, fruit drop; BLC: beginning of leaf coloring; ELC: end of leaf coloring; BLF: beginning of leaf fall; ELF: end of leaf fall; BTB: beginning of turning brown; 50TB: 50% of the leaf turning brown; ETB: end of turning brown; FC: first calling date; LC: last calling date; AD: arrive date; DD: departure date. Mean Standard deviation (SD) of start year, end year and length of time series are shown.

each site to assess the spatial pattern of phenological changes across China. In view of the possible impact of taxonomic groups or time periods on trend estimates, we also calculated the mean trend for each site confining to time series of woody plants starting from 1960s. The impacts of geographic factors on phenological trends were examined by multiple linear regressions separately for spring/summer and autumn phenophases: Tp ¼ Slat x þ Slon y þ Salt z þ Sint

intercept; and Slat, Slon and Salt are the parameters that describe the impact of geographic factors on phenological trends.

(a)

ð1Þ

where Tp is the mean phenological trend at each site; x, y and z are latitude, longitude and altitude, respectively; Sint is the

(b)

Fig. 1 Spatial distribution of 145 observation sites in the metaanalyses of Chinese phenological changes. Center of circle: the locations of each site; circle size: number of time series observed at each site. © 2014 John Wiley & Sons Ltd, Global Change Biology, 21, 265–274

Fig. 2 Frequency distribution of (a) the start and end years of time series and (b) length of the data series.

P H E N O L O G I C A L R E S P O N S E I N C H I N A 273 phenological trends. Actually, phenological trends are determined by many factors. Other than the trends of temperature, the temperature sensitivity of phenophases (change in numbers of days per °C) is also a key factor which cannot be ignored. The existing evidence demonstrated temperature sensitivity of spring phenophases for woody plants is stronger at lower latitudes or warmer locations than at higher latitudes or colder locations (Menzel et al., 2006; Chen & Xu, 2012; Dai et al., 2014; Wang et al., 2014), suggesting the number of days changed in spring phenophases for every 1 °C increase in temperature is less at higher latitudes. Therefore, the greater phenological change at higher latitudes due to stronger warming would be partially offset by the relatively weaker temperature sensitivity. This is the reason why the latitude could only explain little spatial variance in phenological changes. In conclusion, our systematical meta-analysis of all observed phenological time series in China clearly confirmed that there is an evident signal of advancing spring/summer events in wild plants and animals across China. The mean advance of spring/summer of 2.75 days decade1 was stronger than the advance of vegetation green-up onset date (1.3 days decade1, 1982–2010) derived from satellite-derived normalized difference vegetation index (NDVI) in China (Cong et al., 2013). This difference is probably due to the fact that the magnitudes of advance derived by NDVI data are affected by their coarse temporal resolution and the NDVI data could not reflect the animals’ response. The response of autumn phenophases is more complex: the trees, shrubs, herbs and insects show later trend, while other taxonomical groups reveal an advancing trend. The regions of significant warming across China and the spatial pattern of observed changes in phenology of woody plants are generally consistent.

Acknowledgements This article was supported by the Key Project of the National Natural Science Foundation of China (NSFC, No. 41030101); the ‘Strategic Priority Research Program—Climate Change: Carbon Budget and Relevant Issues’ of the Chinese Academy of Sciences (No.XDA05090301), NSFC project (No. 41171043); and the National Basic Research Program of China (2012CB955304).

References Badeck FW, Bondeau A, B€ ottcher K, Doktor D, Lucht W, Schaber J, Sitch S (2004) Responses of spring phenology to climate change. New Phytologist, 162, 295–309. Both C, Van Asch M, Bijlsma RG, Van Den Burg AB, Visser ME (2009) Climate change and unequal phenological changes across four trophic levels: constraints or adaptations? Journal of Animal Ecology, 78, 73–83. Chen X (2013) East Asia. In: Phenology: An Integrative Environmental Science, 2nd edn (ed Schwartz MD), pp. 9–22. Springer, Dordrecht.


Chen X, Xu L (2012) Phenological responses of Ulmus pumila (Siberian Elm) to climate change in the temperate zone of China. International Journal of Biometeorology, 56, 695–706. Cong N, Wang T, Nan H, Ma Y, Wang X, Myneni RB, Piao S (2013) Changes in satellite-derived spring vegetation green-up date and its linkage to climate in China from 1982 to 2010: a multimethod analysis. Global Change Biology, 19, 881–891. Cotton PA (2003) Avian migration phenology and global climate change. Proceedings of the National Academy of Sciences of the United States of America, 100, 12219–12222. Dai J, Wang H, Ge Q (2014) The spatial pattern of leaf phenology and its response to climate change in China. International Journal of Biometeorology, 58, 521–528. Dawson A (2008) Control of the annual cycle in birds: endocrine constraints and plasticity in response to ecological variability. Philosophical Transactions of the Royal Society B: Biological Sciences, 363, 1621–1633. Dose V, Menzel A (2004) Bayesian analysis of climate change impacts in phenology. Global Change Biology, 10, 259–272. IPCC (2013) Summary for policymakers. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (eds Stocker TF, Qin D, Plattner GK, Tignor M, Allen SK, Boschung J, Nauels A, Xia Y, Bex V, Midgley PM), pp. 3–29. Cambridge University Press, Cambridge. Jones PD, Lister DH, Osborn TJ, Harpham C, Salmon M, Morice CP (2012) Hemispheric and large-scale land-surface air temperature variations: An extensive revision and an update to 2010. Journal of Geophysical Research, 117, D5127. Kosaka Y, Xie S (2013) Recent global-warming hiatus tied to equatorial Pacific surface cooling. Nature, 501, 403–407. Linderholm HW (2006) Growing season changes in the last century. Agricultural and Forest Meteorology, 137, 1–14. Ma T, Zhou C (2012) Climate-associated changes in spring plant phenology in China. International Journal of Biometeorology, 56, 269–275. Menzel A, Sparks TH, Estrella N et al. (2006) European phenological response to climate change matches the warming pattern. Global Change Biology, 12, 1969–1976. Ministry of Environmental Protection of China (2008) China’s Fourth National Report on Implementation of the Convention on Biological Diversity, pp. 1–99. China Environmental Science Press, Beijing. Parmesan C (2007) Influences of species, latitudes and methodologies on estimates of phenological response to global warming. Global Change Biology, 13, 1860–1872. Parmesan C, Yohe G (2003) A globally coherent fingerprint of climate change impacts across natural systems. Nature, 421, 37–42. Richardson AD, Keenan TF, Migliavacca M, Ryu Y, Sonnentag O, Toomey M (2013) Climate change, phenology, and phenological control of vegetation feedbacks to the climate system. Agricultural and Forest Meteorology, 169, 156–173. Root TL, Price JT, Hall KR, Schneider SH, Rosenzweig C, Pounds JA (2003) Fingerprints of global warming on wild animals and plants. Nature, 421, 57–60. Rosenzweig C, Casassa G, Karoly DJ et al. (2007a) Assessment of observed changes and responses in natural and managed systems. In: Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change (eds Parry ML, Canziani OF, Palutikof JP, van der Linden PJ, Hanson CE), pp. 79–131. Cambridge University Press, Cambridge. Rosenzweig C, Casassa G, Karoly DJ et al. (2007b) Supplementary material to chapter 1: assessment of observed changes and responses in natural and managed systems. In: Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change (eds Parry ML, Canziani OF, Palutikof JP, van der Linden PJ, Hanson CE), pp. 1–13. Cambridge University Press, Cambridge. Rosenzweig C, Karoly D, Vicarelli M et al. (2008) Attributing physical and biological impacts to anthropogenic climate change. Nature, 453, 353–357. Rutishauser T, Luterbacher J, Jeanneret F, Pfister C, Wanner H (2007) A phenologybased reconstruction of interannual changes in past spring seasons. Journal of Geophysical Research, 112, G4016. Schwartz MD, Chen X (2002) Examining the onset of spring in China. Climate Research, 21, 157–164. Seneviratne SI, Donat MG, Mueller B, Alexander LV (2014) No pause in the increase of hot temperature extremes. Nature Climate change, 4, 161–163. Sparks TH, Braslavska O (2001) The effects of temperature, altitude and latitude on the arrival and departure dates of the swallow Hirundo rustica in the Slovak Republic. International Journal of Biometeorology, 45, 212–216.

P H E N O L O G I C A L R E S P O N S E I N C H I N A 269 Table 2 Summary of phenological trends in China

(a) N Spring/summer phenophases 789 Negall Negsig 363 263 Negunk 80 Posall Possig 10 36 Posunk 869 Trmean Trsitemean 145 Autumn phenophases 122 Negall 24 Negsig Negunk 67 272 Posall 92 Possig Posunk 124 394 Trmean 108 Trsitemean

Proportion (%)

90.8 41.8 30.3 9.2 1.2 4.1 2.75 2.92

(b) 31.0 6.1 17.0 69.0 23.4 31.5 2.00 1.69

Temporal trends of spring/summer phenophases (top) and autumn phenophases (bottom) from 1960 to 2011 (time series 20+ years). Negall and Posall: proportions of negative and positive trends. Negsig and Possig: proportions of significantly negative and positive trends (P < 0.05). Negunk and Posunk: proportions of negative and positive trends whose significances were unknown. Trmean: mean trend of all time series; Trsitemean: mean trend for all sites (the trends were first averaged to each site).

Fig. 4 Boxplot of phenological trends across taxonomic groups of species for (a) spring/summer phenophases and (b) autumn phenophases. The bottoms and tops of boxes are the 25th and 75th percentiles; the bands near the middle are the median. The ends of the whiskers represent 10th and 90th percentiles. The crosses designate the mean value.

Spatial patterns of phenological trends The spring/summer phenophases averaged across all time series at each site showed earlier trends in 135

showed advancing trends in autumn phenophases (Fig. 4b). Through multiple comparisons, no significant differences were found between any two groups (Table 3).

Table 3 Comparisons across taxonomic groups in phenological trends

Taxon

Differences among time periods Because all the time series ended between 2001 and 2011, the time series for woody plants (trees and shrubs) could be grouped by start year (1960s, 1970s and 1980s). On the average, the spring/summer phenophases (leafing, flowering, fruiting timings) advanced and autumn phenophases (leaf coloring or fall) delayed regardless of the start year of time series (Fig. 5). However, the mean trend starting from the 1980s (3.47 days decade1) for spring/summer phenophases was significantly stronger than the trend starting from the 1960s (1.81 days decade1) or 1970s (1.50 days decade1) (Fisher’s LSD test, P < 0.05, Table 4). With respect to autumn phenophases, the mean delaying trend from 1980s onwards (0.66 days decade1) was significantly weaker than the trend from 1960s (2.36 days decade1). © 2014 John Wiley & Sons Ltd, Global Change Biology, 21, 265–274

N

Mean SD (days decade1)

Spring/summer phenophases Tree 632 2.29 Shrub 85 2.17 Herb 127 5.71 Bird 14 0.56 Amphibian 4 6.11 Insect 7 2.11 Autumn phenophases Tree 265 1.93 Shrub 29 2.36 Herb 79 2.50 Bird 10 2.11 Amphibian 4 1.10 Insect 7 4.84

2.50 1.71 7.90 4.32 4.72 11.19

4.38 3.68 13.41 9.79 5.80 6.58

Fisher’s LSD test

a, c b, d, f a, b, e, h c, d, e, g f, g h

The trends in each group were analyzed by one-way ANOVA, with multiple comparisons by Fisher’s LSD. Significant differences between taxonomic groups are indicated by same letters (P < 0.05).

270 Q . G E et al. (a)

(b)

Fig. 5 Phenological trends grouped by start year of time series (1960s, 1970s, 1980s) for woody plants (trees and shrubs). (a) spring/summer phenophases and (b) autumn phenophases. See Fig. 4 for boxplot description.

Table 4 Comparisons in phenological trends across groups classified by start year of time series for woody plants (including trees and shrubs)

Start year

N

Mean SD (days decade1)

Spring/summer phenophases 1960s 393 1.81 1970s 99 1.50 1980s 223 3.47 Autumn phenophases 1960s 140 2.36 1970s 44 1.81 1980s 140 0.66

Fisher’s LSD test

1.22 1.25 3.58

a b a, b

2.51 3.30 7.81

a a

The trends in each group were analyzed by one-way ANOVA, with multiple comparisons by Fisher’s LSD. Significant differences between taxonomic groups are indicated by same letters (P < 0.05).

(93.1%) of 145 sites (Fig. 6a). The sites exhibiting earlier trends almost covered the whole country. The delay of spring/summer phenophases were found in only several sites located in Northeast China Plain, North China Plain and Yunnan–Guizhou Plateau. On the other hand, the changes in autumn phenophases were highly spatially heterogeneous (Fig. 6b). The trends toward later and earlier autumn phenophases were found in 71 (65.7%) and 37 (34.3%) of 108 sites, respectively (Fig. 6b).

By applying the multiple regression analysis method between phenological trends and geographical factors (Table 5), we found that the regression slopes of latitude (0.09 days decade1 °1) and altitude (0.11 days decade1 °1) were significant, suggesting that the advance of spring/summer phenophases was stronger in lower latitude and higher altitude. Furthermore, there are no significant geographical factors impacting the trends in autumn phenophases. Overall, geographical factors could only explain 9% (P < 0.01) and 3% (P = 0.98) of variation in phenological trends across sites for spring/summer and autumn phenophases, respectively. The multiple regression analysis was repeated using subdataset only consisting of the time series of woody plants from the 1960s to 2000s. The results showed that advance in leafing, flowering and fruiting timing of woody plants was stronger in higher latitude, more continent climate (western longitude) and lower altitude, but only the regression slopes of altitude was significant. Therefore, the spatial pattern of phenological trends was inversed when confining the data series to woody plants starting from 1960s. However, variation in spring/summer phenological trends explained by the model was improved (R2 = 0.14, P = 0.07, Table 5). Regarding autumn phenophases, no significant impacts of geographic factors on phenological trend were found and the variation explained by the model was still very limited (R2 = 0.04, P = 0.75).

Relationships between phenological and temperature trends Figure 7(a) showed the relationship between trends in spring/summer phases of woody plants and March– August temperature. Sites with stronger warming trends had significantly earlier spring/summer phenophases (R = 0.33, P < 0.05). Furthermore, stronger warming trends were associated with more delay in leaf coloring and fall date of woody plants, but this relationship was not significant (R = 0.25, P = 0.17, Fig. 7b).

Discussion Our systematic meta-analysis of more than 1200 phenological time series in China comprised a huge selection of species and locations, and included various phases of plants and animals covering the all development stages. There is a coherent signal of advancing spring/ summer phenophases across China apparent in 90.8% of the records. The mean advance of spring/summer in China (2.75 days decade1) is weaker than the estimate (6.13 days decade1) of Ma & Zhou (2012). © 2014 John Wiley & Sons Ltd, Global Change Biology, 21, 265–274

P H E N O L O G I C A L R E S P O N S E I N C H I N A 271 (a)

(b)

Fig. 6 Mean phenological trends (days decade1) for all time series (with 20+ years of data between 1960 and 2011) at each observation site and the linear trends of surface temperature between 1961 and 2011 in China. (a) spring/summer phenophases and March–August temperature trends; (b): autumn phenophases and September–November temperature trends.

The reason is the huge difference in samples size. Ma & Zhou (2012) only used 384 phenological time series of woody and herbaceous plants from 1980s to 2000s, while we used an expanded dataset consisting of 1263 time series. In Europe, 78% of the spring/summer plant phenophases advanced, and the mean advance was 2.5 days decade1 (Menzel et al., 2006), thus the trends in Europe is a little bit lower than that in China. Parmesan (2007) estimated an overall spring advance across the northern hemisphere of 2.8 days decade1 using the combined dataset from Root et al. (2003) and Parmesan & Yohe (2003). This result is very close to our estimate. However, we noticed that the meta-analysis of Parmesan (2007) did not include the phenological time series in China, so our study provides an independent evidence of climate change impact on the Northern Hemisphere. With respect to autumn phenophases, we found a mean delay of 2.0 days decade1 in China. The weaker trend in autumn phenophases relative to spring phenophases results from the multiple response among © 2014 John Wiley & Sons Ltd, Global Change Biology, 21, 265–274

taxonomic groups. For woody plants (including trees and shrubs), the leaf coloring and leaf fall dates delayed by a mean trend of 1.98 days decade1. This result is Table 5 The spatial patterns of phenological trends in China N

Slat

Slon

Spring/summer phenophases Overall 145 0.09* 0.01 Woody 49 0.03 0.01 plants (1960s-) Autumn phenophases Overall 108 0.13 0.05 Woody 32 0.05 0.05 plants (1960s-)

Salt

Sint

R2

1.11** 1.01*

4.99 1.91

0.09** 0.14

1.30 0.85

0.22 4.85

0.03 0.04

N: number of sites. Slat, Slon, Salt and Sint: parameters in Eqn (1). Unit for Slat and Slon: days decade1 °1; Unit for Salt: days decade1 km1; Unit for Sint: days decade1. R2: variance of phenological trends explained by Eqn (1). *P < 0.05; **P < 0.01.

272 Q . G E et al. (a)

(b)

Fig. 7 Relationships between (a) trends in spring/summer phenophases of woody plants (1960s–2000s) and trends in March– August temperature over 1961–2011 and (b) trends in autumn phenophase of woody plants and in September–November temperature over the same time period among different sites. The dashed lines show the regression lines. For spring/summer, P = 0.02); for autumn, y = 2.41x 0.80 (R2 = 0.11, 2 y = 5.05x + 0.09 (R = 0.06, P = 0.17).

different with that in Europe, in which no clear trends of leaf coloring/fall were found (Menzel et al., 2006). Herbs show a later turning brown date of 2.5 days decade1 on the average. However, this value is not significantly different from zero (t-test, P = 0.23), since there were similar proportions of earlier and later trends. The last calling date of two insect species (cicada and cricket) also became later. The departure date of birds advanced by a mean trend of 2.11 days decade1, which was in accordance with the change in avian migration phenology in the UK (Cotton, 2003) and Slovak Republic (Sparks & Braslavsk a, 2001). The last calling date of frog (representative of amphibian) also became earlier at the mean rate of 1.1 days decade1, but this value was not significantly different from zero (t-test, P = 0.73). In general, different taxonomic groups did not show consistent earlier or later trends in autumn phenophases. In addition, different meta-analysis studies often give different relative magnitudes of phenological responses among taxonomic groups. For example, Root et al. (2003) found that the spring phenophases of trees show less advancement than all other taxonomic groups. In this study, however, the mean spring advance for trees

is only significantly less than herbs but not significantly different from other taxa. In Parmesan (2007), the birds’ arrival date show more than three times greater phenological advancement than does the flowering of herbs or shrubs. However, in this study, the birds’ migratory arrival shows a weakly later trend. Theoretically, birds would demonstrate constrained phenological plasticity due to photoperiodic induction of migration (Dawson, 2008). However, it should be noticed that the number of different species of birds is rather small and also the number of the locations and time series (Appendix S3). The limited number of taxonomic representation may generate a biased estimate of trends. Our results highlighted the advancement in leafing or flowering time of woody plants would be stronger if the observation series started from more recent decade. This result is consistent with the stronger warming trend from the middle 1980s in China (Wang & Gong, 2000). Thus, time period is a considerable factor affecting the estimate of phenological response. Besides China, the impact of time period is also found in other countries. For example, using the records of the statistical ‘spring plants’ in Switzerland and the beginning of flowering of red currant in Germany, Rutishauser et al. (2007) calculated the slopes of linear regression for each 30-year period. Their results showed that the spring phenophases advanced during the warming trend of 1940s, delayed during the cooling trends of the 1950s and 1960s, and again became earlier from 1970s (Rutishauser et al., 2007). The faster rate of change from mid1980s in the blossom onset of three tree species (Prunus avium, Galanthus nivalis and Tilia platyphyllos) was also detected by using Bayesian analysis method (Dose & Menzel, 2004). Instrumental temperature record indicated that the higher latitude in Northern hemisphere showed stronger warming trend than in lower latitude (Jones et al., 2012). Therefore, a stronger phenological trend at higher latitude is expected. Root et al. (2003) found the mean trend between 32° and 49.9° N latitude is weaker than between 50° and 72° N latitude. However, their data show no visible pattern with latitude when analyzed by linear regression between phenological trends and latitudes (Parmesan, 2007). In our study, the analysis of the dataset consisting of woody plants (1960s– 2000s) does show a increase in strength of either spring/summer advancement or autumn delay with increasing latitude (Table 5), but the geographic factors (latitude, longitude and altitude) could only explain 14% and 4% of the overall variance in spring/summer and autumn phenological changes, respectively. Hence, as suggested by Parmesan (2007), the latitude is not yet an important explanatory variable in estimate of © 2014 John Wiley & Sons Ltd, Global Change Biology, 21, 265–274

P H E N O L O G I C A L R E S P O N S E I N C H I N A 273 phenological trends. Actually, phenological trends are determined by many factors. Other than the trends of temperature, the temperature sensitivity of phenophases (change in numbers of days per °C) is also a key factor which cannot be ignored. The existing evidence demonstrated temperature sensitivity of spring phenophases for woody plants is stronger at lower latitudes or warmer locations than at higher latitudes or colder locations (Menzel et al., 2006; Chen & Xu, 2012; Dai et al., 2014; Wang et al., 2014), suggesting the number of days changed in spring phenophases for every 1 °C increase in temperature is less at higher latitudes. Therefore, the greater phenological change at higher latitudes due to stronger warming would be partially offset by the relatively weaker temperature sensitivity. This is the reason why the latitude could only explain little spatial variance in phenological changes. In conclusion, our systematical meta-analysis of all observed phenological time series in China clearly confirmed that there is an evident signal of advancing spring/summer events in wild plants and animals across China. The mean advance of spring/summer of 2.75 days decade1 was stronger than the advance of vegetation green-up onset date (1.3 days decade1, 1982–2010) derived from satellite-derived normalized difference vegetation index (NDVI) in China (Cong et al., 2013). This difference is probably due to the fact that the magnitudes of advance derived by NDVI data are affected by their coarse temporal resolution and the NDVI data could not reflect the animals’ response. The response of autumn phenophases is more complex: the trees, shrubs, herbs and insects show later trend, while other taxonomical groups reveal an advancing trend. The regions of significant warming across China and the spatial pattern of observed changes in phenology of woody plants are generally consistent.

Acknowledgements This article was supported by the Key Project of the National Natural Science Foundation of China (NSFC, No. 41030101); the ‘Strategic Priority Research Program—Climate Change: Carbon Budget and Relevant Issues’ of the Chinese Academy of Sciences (No.XDA05090301), NSFC project (No. 41171043); and the National Basic Research Program of China (2012CB955304).

References Badeck FW, Bondeau A, B€ ottcher K, Doktor D, Lucht W, Schaber J, Sitch S (2004) Responses of spring phenology to climate change. New Phytologist, 162, 295–309. Both C, Van Asch M, Bijlsma RG, Van Den Burg AB, Visser ME (2009) Climate change and unequal phenological changes across four trophic levels: constraints or adaptations? Journal of Animal Ecology, 78, 73–83. Chen X (2013) East Asia. In: Phenology: An Integrative Environmental Science, 2nd edn (ed Schwartz MD), pp. 9–22. Springer, Dordrecht.


Chen X, Xu L (2012) Phenological responses of Ulmus pumila (Siberian Elm) to climate change in the temperate zone of China. International Journal of Biometeorology, 56, 695–706. Cong N, Wang T, Nan H, Ma Y, Wang X, Myneni RB, Piao S (2013) Changes in satellite-derived spring vegetation green-up date and its linkage to climate in China from 1982 to 2010: a multimethod analysis. Global Change Biology, 19, 881–891. Cotton PA (2003) Avian migration phenology and global climate change. Proceedings of the National Academy of Sciences of the United States of America, 100, 12219–12222. Dai J, Wang H, Ge Q (2014) The spatial pattern of leaf phenology and its response to climate change in China. International Journal of Biometeorology, 58, 521–528. Dawson A (2008) Control of the annual cycle in birds: endocrine constraints and plasticity in response to ecological variability. Philosophical Transactions of the Royal Society B: Biological Sciences, 363, 1621–1633. Dose V, Menzel A (2004) Bayesian analysis of climate change impacts in phenology. Global Change Biology, 10, 259–272. IPCC (2013) Summary for policymakers. In: Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (eds Stocker TF, Qin D, Plattner GK, Tignor M, Allen SK, Boschung J, Nauels A, Xia Y, Bex V, Midgley PM), pp. 3–29. Cambridge University Press, Cambridge. Jones PD, Lister DH, Osborn TJ, Harpham C, Salmon M, Morice CP (2012) Hemispheric and large-scale land-surface air temperature variations: An extensive revision and an update to 2010. Journal of Geophysical Research, 117, D5127. Kosaka Y, Xie S (2013) Recent global-warming hiatus tied to equatorial Pacific surface cooling. Nature, 501, 403–407. Linderholm HW (2006) Growing season changes in the last century. Agricultural and Forest Meteorology, 137, 1–14. Ma T, Zhou C (2012) Climate-associated changes in spring plant phenology in China. International Journal of Biometeorology, 56, 269–275. Menzel A, Sparks TH, Estrella N et al. (2006) European phenological response to climate change matches the warming pattern. Global Change Biology, 12, 1969–1976. Ministry of Environmental Protection of China (2008) China’s Fourth National Report on Implementation of the Convention on Biological Diversity, pp. 1–99. China Environmental Science Press, Beijing. Parmesan C (2007) Influences of species, latitudes and methodologies on estimates of phenological response to global warming. Global Change Biology, 13, 1860–1872. Parmesan C, Yohe G (2003) A globally coherent fingerprint of climate change impacts across natural systems. Nature, 421, 37–42. Richardson AD, Keenan TF, Migliavacca M, Ryu Y, Sonnentag O, Toomey M (2013) Climate change, phenology, and phenological control of vegetation feedbacks to the climate system. Agricultural and Forest Meteorology, 169, 156–173. Root TL, Price JT, Hall KR, Schneider SH, Rosenzweig C, Pounds JA (2003) Fingerprints of global warming on wild animals and plants. Nature, 421, 57–60. Rosenzweig C, Casassa G, Karoly DJ et al. (2007a) Assessment of observed changes and responses in natural and managed systems. In: Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change (eds Parry ML, Canziani OF, Palutikof JP, van der Linden PJ, Hanson CE), pp. 79–131. Cambridge University Press, Cambridge. Rosenzweig C, Casassa G, Karoly DJ et al. (2007b) Supplementary material to chapter 1: assessment of observed changes and responses in natural and managed systems. In: Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change (eds Parry ML, Canziani OF, Palutikof JP, van der Linden PJ, Hanson CE), pp. 1–13. Cambridge University Press, Cambridge. Rosenzweig C, Karoly D, Vicarelli M et al. (2008) Attributing physical and biological impacts to anthropogenic climate change. Nature, 453, 353–357. Rutishauser T, Luterbacher J, Jeanneret F, Pfister C, Wanner H (2007) A phenologybased reconstruction of interannual changes in past spring seasons. Journal of Geophysical Research, 112, G4016. Schwartz MD, Chen X (2002) Examining the onset of spring in China. Climate Research, 21, 157–164. Seneviratne SI, Donat MG, Mueller B, Alexander LV (2014) No pause in the increase of hot temperature extremes. Nature Climate change, 4, 161–163. Sparks TH, Braslavska O (2001) The effects of temperature, altitude and latitude on the arrival and departure dates of the swallow Hirundo rustica in the Slovak Republic. International Journal of Biometeorology, 45, 212–216.

274 Q . G E et al. Thackeray SJ, Sparks TH, Frederiksen M et al. (2010) Trophic level asynchrony in rates of phenological change for marine, freshwater and terrestrial environments. Global Change Biology, 16, 3304–3313. The R Core Team (2014) R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna. Walther GR (2003) Plants in a warmer world. Perspectives in Plant Ecology, Evolution and Systematics, 6, 169–185. Wang S, Gong D (2000) Enhancement of the warming trend in China. Geophysical Research Letters, 27, 2581–2584. Wang H, Dai J, Zheng J, Ge Q (2014) Temperature sensitivity of plant phenology in temperate and subtropical regions of China from 1850 to 2009. International Journal of Climatology, doi: 10.1002/joc.4026. Xu Y, Gao X, Shen Y, Xu C, Shi Y, Giorgi F (2009) A daily temperature dataset over China and its application in validating a RCM simulation. Advances in Atmospheric sciences, 26, 763–772.

Supporting Information Additional Supporting Information may be found in the online version of this article: Appendix S1. References of studies that fit the criteria selected for meta-analysis. Appendix S2. Information about phenological change in China derived from 48 studies. Appendix S3. Spatial distribution of observation sites in the meta-analyses of Chinese phenological changes for each taxon.
