Spatial and temporal variation of phenological growing season and climate change impacts in
See discussions, stats, and author profiles for this publication at: https://www.sodocs.net/doc/4c3627448.html,/publication/227728535 Spatial and temporal variation of phenological growing season and climate change impacts in temperate eastern China. Glob Chang Biol
ARTICLE in GLOBAL CHANGE BIOLOGY · JUNE 2005
Impact Factor: 8.22 · DOI: 10.1111/j.1365-2486.2005.00974.x
CITATIONS 105DOWNLOADS
24
VIEWS
82
3 AUTHORS, INCLUDING:
Xiaoqiu Chen
Peking University
16 PUBLICATIONS 343 CITATIONS
SEE PROFILE
Available from: Xiaoqiu Chen
Retrieved on: 11 August 2015
Spatial and temporal variation of phenological growing season and climate change impacts in temperate
eastern China
X I A O Q I U C H E N,B I N G H U and R O N G Y U
Department of Geography,College of Environmental Sciences,MOE Laboratory for Earth Surface Processes,Peking University, Beijing100871,China
Abstract
Using phenological and normalized difference vegetation index(NDVI)data from1982
to1993at seven sample stations in temperate eastern China,we calculated the
cumulative frequency of leaf unfolding and leaf coloration dates for deciduous species
every5days throughout the study period.Then,we determined the growing season
beginning and end dates by computing times when50%of the species had undergone
leaf unfolding and leaf coloration for each station year.Next,we used these beginning
and end dates of the growing season as time markers to determine corresponding
threshold NDVI values on NDVI curves for the pixels overlaying phenological stations.
Based on a cluster analysis,we determined extrapolation areas for each phenological
station in every year,and then implemented the spatial extrapolation of growing season
parameters from the seven sample stations to all possible meteorological stations in the
study area.
Results show that spatial patterns of growing season beginning and end dates
correlate signi?cantly with spatial patterns of mean air temperatures in spring and
autumn,respectively.Contrasting with results from similar studies in Europe and North
America,our study suggests that there is a signi?cant delay in leaf coloration dates,
along with a less pronounced advance of leaf unfolding dates in different latitudinal
zones and the whole area from1982to1993.The growing season has been extended by
1.4–3.6days per year in the northern zones and by1.4days per year across the entire
study area on average.The apparent delay in growing season end dates is associated
with regional cooling from late spring to summer,while the insigni?cant advancement
in beginning dates corresponds to inconsistent temperature trend changes from late
winter to spring.On an interannual basis,growing season beginning and end dates
correlate negatively with mean air temperatures from February to April and from May to
June,respectively.
Keywords:air temperature,climate change,interannual variability,linear trend,normalized difference
vegetation index,phenological growing season,spatial extrapolation,spatial pattern,temperate
eastern China
Received3April2004;revised version received18October2004;accepted23February2005
Introduction
Detecting growing season variability of terrestrial vegetation is crucial for identifying responses of ecosystems to recent climate change at seasonal and interannual time scales(Chen et al.,2001;Walther et al., 2002).A lengthening of the growing season in northern vegetation over the past decades is speculated based on observed changes in the seasonal signal in atmospheric CO2(Keeling et al.,1996)and satellite observations (Myneni et al.,1997;Zhou et al.,2001).These atmo-spheric and satellite data are supported by?eld-based phenological observation of plants(Ahas,1999;Bradley et al.,1999;Menzel&Fabian,1999;Beaubien& Freeland,2000;Chmielewski&Roetzer,2001;Fitter& Fitter,2002)and simulations from a dynamic vegetation model(Lucht et al.,2002),especially the signi?cant
Correspondence:Xiaoqiu Chen,tel.1861062875174,
fax1861062751187,e-mail:cxq@https://www.sodocs.net/doc/4c3627448.html,
Global Change Biology(2005)11,1118–1130,doi:10.1111/j.1365-2486.2005.00974.x
1118r2005Blackwell Publishing Ltd
advancement of phenological events in spring and the less pronounced delay of phenological events in autumn across Europe and North America.Recent global meta-analyses strongly suggest that these pheno-logical?ngerprints implicate climate change as an im-portant driving force on natural systems(Parmesan& Yohe,2003;Root et al.,2003).
In China,however,phenological stations and con-ventional phenological data are comparatively scarce; therefore,the options for detecting growing season trends are to estimate the growing season of land vegetation using limited station phenological and climate data,or with satellite data(Chen&Pan, 2002).In the former case,a simple phenological model, driven by daily maximum–minimum temperatures can be employed as a surrogate measure of the onset of spring.Contrasting with the result from a similar study in North America(Schwartz&Reiter,2000),the onset of spring plant growth derived from modelled lilac?rst leaf and?rst bloom dates in China showed no apparent linear trend from1959to1993(Schwartz&Chen,2002). In the latter case,various methods have been developed to determine a satellite-sensor-derived growing season at different spatial scales using the normalized differ-ence vegetation index(NDVI)data derived from the advanced very high-resolution radiometer(AVHRR)or the enhanced vegetation index data derived from the moderate-resolution imaging spectroradiometer(Lloyd, 1990;Fisher,1994;Reed et al.,1994;Markon et al.,1995; Moulin et al.,1997;White et al.,1997;Botta et al.,2000; Zhang et al.,2004).However,because metrics and thresholds of vegetation indices may not directly correspond to conventional,ground-based phenologi-cal events,but rather provide indicators of vegetation dynamics(Reed et al.,1994),a detailed comparison of these satellite measures with ground-based phenologi-cal events is needed(Schwartz,1998;Chen et al.,2000). In recent years,some studies have been carried out to compare satellite-sensor-derived onset and offset of greenness with surface phenological stages of indivi-dual plant species,mono-speci?c forests,and mixed forests for selected biomes(White et al.,1997;Duchemin et al.,1999;Schwartz et al.,2002;Badeck et al.,2004). Other than the above top-down method,namely determining the satellite-sensor-derived growing sea-son at a regional scale?rst,and then validating it using conventional phenological data at local scales,Chen et al.(2000,2001)developed a bottom-up method, namely determining the phenological growing season at sample stations?rst and then?nding out the corresponding threshold NDVI values at pixels over-laying the sample stations in order to extrapolate the phenological growing season at a regional scale.To implement the above goal,the objectives of this study were to(1)extrapolate the phenological growing season of plant community in temperate eastern China,using threshold NDVI values obtained by phenology-satellite analyses at sample stations,(2)identify spatial patterns and trends of growing season parameters(beginning, end,and length)at local,zonal,and regional scales;and (3)assess the relationship between growing season parameters and seasonal air temperatures with respect to spatial and temporal variations.
Materials and methods
Study area
Situated in the southeastern part of the Eurasian Continent on the west coast of the Paci?c Ocean,China is world famous for its monsoonal climate caused by the difference in heat reserves between the world’s largest continent and the world’s biggest ocean. Monsoonal winds bring the most striking seasonal changes to the vast eastern areas of China,especially in the temperate zone,which is cold and dry in winter and warm and humid in summer.Because seasonal biome dynamics are controlled by recurrent variations of environmental conditions,especially climate,seasonal biome features in temperate areas with distinct climatic rhythmicity are rich and colorful(Chen,2003a).There-fore,we selected a study area that is located from321N to the northern border(531310N)of China and from 1061E to the eastern border(1351030E)or the coastline of China.The parallel321N and the meridian1061E represent approximately the southern border of the temperate zone in eastern China(Zhang&Lin,1985) and the west edge of the in?uence of the summer monsoon,respectively(Gao et al.,1962).The study area covers the cold temperate,middle temperate,and warm temperate zones.The dominant vegetation types include conifer forest in the cold temperate zone, deciduous broad-leaved and coniferous mixed forest and steppe in the middle temperate zone,deciduous broad-leaved forest in the warm temperate zone,and various interzonal crops(Compilation Committee on the Vegetation of China,1980).In order to determine the ground-based growing season and assess the relationship between growing season parameters and climate factors,we selected seven phenological stations and125meteorological stations altogether as sample sites(Fig.1).The selection of the seven phenological stations,namely Harbin(451450N,1261400E,146m), Mudanjiang(441260N,1291400E,300m),Gaixian (401260N,1221200E,45m),Beijing(401010N,1161200E, 50m),Xingtai(371040N,1141300E,77m),Luoyang (341400N,1121250E,155m),and Xi’an(341130N, 1081580E,438m),was based on diversity of the plant
VA R I A T I O N O F P H E N O L O G I C A L G R O W I N G S E A S O N I N C H I N A1119 r2005Blackwell Publishing Ltd,Global Change Biology,11,1118–1130
species(430)and accuracy of the observations.As there are parallel phenological and meteorological observations at Harbin,Mudanjiang,Beijing,Xingtai, and Xi’an,the total number of individual sample stations is127.Most of the phenological and meteoro-logical stations are located within suburbs of cities with less signi?cant urban heat island effects.
Phenology and climate data
Because phenological observation and study have important applications in directing agricultural produc-tion,the Chinese Academy of Sciences(CAS)estab-lished a countrywide phenological network in the early 1960s.Observations began in1963and continued until 1996.In2003,phenological observations were resumed. The observation program of the CAS network included a total of173observed species.Of these,33species of woody plants,two species of herbaceous plants,and 11species of fauna were observed across the network (Chen,2003b).The observed trees and shrubs were selected according to spatial comparability and local representativeness and the observations were carried out mainly by botanical gardens,research institutes, and universities according to uniform observation criteria(Institute of Geography at Chinese Academy of Sciences,1965).The phenophases of woody plants included bud burst,?rst leaf unfolding,50%leaf unfolding,?ower bud or in?orescence appearance,?rst bloom,50%bloom,end of blooming,fruit or seed maturing,fruit or seed shedding,?rst leaf coloration, full leaf coloration,?rst defoliation,and end of defoliation.Phenological data used in this study were acquired from Chinese Yearbooks of Animal and Plant Phenological Observation for the period of1982–1988 compiled by the Institute of Geography at Chinese
N
Fig.1Location of phenological and meteorological stations.
1120X.C H E N et al.
r2005Blackwell Publishing Ltd,Global Change Biology,11,1118–1130
Academy of Sciences(1988,1989a,b,1992)and from an unpublished data set for the period of1989–1993 provided by the Institute of Geography at CAS.The study period ends in1993because the data in the last3 years(1994–1996)were not available.In order to conduct joint analyses of plant phenology and satellite data,we chose leaf unfolding and leaf coloration as indicator events to show the start and end of the growing season of local plant communities.The data of leaf unfolding and leaf coloration of all observed deciduous trees and shrubs at the sample sites were veri?ed and revised according to the inherent sequence of occurrence dates of phenophases and the linear correlation between time series of phenophases at each site(Yang&Chen,1995).
As the occurrence dates of phenological events of the observed woody plants correlated closely with those of forest trees,herbs,and crops(Schnelle,1955;Brandtner, 1958;Chu,1964;Pfau,1964),and served as a measure of equal physiological development correlated with cur-rent and accumulative climatic factors(Newman& Beard,1962;Chu,1964),the observed trees and shrubs were usually applied as indicator species to deduce the development of forest trees,herbs,and crops,and to determine the correct time for forest and farm tasks (Hopkins&Murray,1933;Chu&Wan,1973).Extend-ing this approach,we can also use the observed plants and phenophases of the CAS network as indicator species and indicator events to represent the local plant community and its seasonality in temperate eastern China.
It is well known that air temperature is the most important controlling factor related to trees’phenology in the temperate zone(Romberger,1963;Chen,1994; Kramer,1996;Chmielewski&Roetzer,2001;Zhang et al.,2004),whereas precipitation and photoperiod play a less pronounced role in phenological develop-ment(Chen,1994;Chen&Pan,2002).Dealing with the selection of temperature parameters,the heat unit, expressed in growing degree days(GDD)is frequently used to describe the timing of phenological stages and simulate the phenological development of woody plants(McMaster&Wilhelm,1997;Murray et al., 1989;Haenninen,1990;Kramer,1996),while monthly and seasonal mean temperatures are usually applied to analyze statistical relationships between plant pheno-logy and climate conditions(Beaubien&Freeland, 2000;Sparks et al.,2000;Chmielewski&Roetzer,2001; Chen&Pan,2002;Fitter&Fitter,2002;Menzel,2003).In the present study,daily mean air temperatures at the 125meteorological stations from January1982to December1993were processed into monthly and annual mean air temperatures for each station,which then served as the driving parameter for carrying out a correlation analysis between the beginning and end dates of the growing season and mean air temperatures prior to and during the beginning and end dates. Satellite data
Satellite data were derived from the AVHRR on the ‘afternoon’NOAA operational meteorological satellite. The NDVI was obtained from the NOAA/NASA Path?nder AVHRR Land data set for1982–1993at 8km spatial resolution.The NDVI composites were generated by selecting the highest NDVI value over each10-day period in order to reduce the effect of cloud https://www.sodocs.net/doc/4c3627448.html,ing Geographic Information Systems software(ArcInfo),we produced an NDVI data set corresponding to pixels overlaying the phenological and meteorological stations during1982–1993.Because of the inherent nature in AVHRR data acquisition and processing because of satellite viewing geometry(Go-ward et al.,1991),atmospheric haze and cloud,and temporal data composites(Holben,1986),the10-day NDVI composites can be biased.These effects result in reduced NDVI values.To compensate for the effects,we used a smoothing method suggested by Chen et al. (2000)to modify the maximum value composite NDVI data at pixels overlaying the sample stations.For the NDVI value of every‘n th’10-day period within a year, we calculated the mean value of the(nà1)th and (n11)th10-day periods and then compared this mean with the value for the original n th10-day period.If the mean value was greater than the original value,we replaced the original with the mean as the NDVI value of the n th10-day period;if the mean NDVI value was smaller than the original NDVI value,we retained the original.As the maximum value composite technique has eliminated some high-frequency‘noises’in the NDVI pro?les,the smoothing method we proposed would probably be better appropriate for reducing residual‘noises’than methods for extracting NDVI pro?les from daily data(e.g.the best index slope extraction)(Viovy et al.,1992).
Determining the growing season and threshold NDVI values at sample sites
Many studies have shown that NDVI is sensitive to variations in vegetation coverage or density and provides an ef?cient tool to document the biophysical state of the continental surfaces(Justice et al.,1985; Malingreau,1986;Viovy&Saint,1994;Moulin et al., 1997;White et al.,1997;Botta et al.,2000).Among the surface parameters that can be related to satellite sensor-derived greenness development,plant pheno-logy is a key measure related to seasonal vegetation
VA R I A T I O N O F P H E N O L O G I C A L G R O W I N G S E A S O N I N C H I N A1121 r2005Blackwell Publishing Ltd,Global Change Biology,11,1118–1130
coverage and density in the temperate zone (Viovy &Saint,1994;Chen &Pan,2002).As NDVI measurements integrate observations of different plants and tend to provide descriptive characteristics of phenological landscape events,rather than direct associations with the phenological performance of speci?c plants during the growing season (Achard &Blasco,1990;Reed et al .,1994),phenological data of local plant communities are more suitable for surface satellite analyses than those of individual plant species.Previous comparisons be-tween temporal NDVI pro?les and frequency curves of occurrence dates of all plant phenophases at the biome level show that they have similar patterns in spring and autumn (Chen et al .,2001).This relationship validates that the NDVI curve represents ‘green wave’development in spring and ‘brown wave’development in autumn.For this current study,we have rede?ned the growing season as the time interval between the date on which 50%of all observed trees and shrubs enter leaf unfolding in spring and the date on which 50%of all observed trees and shrubs enter leaf coloration in autumn.Thus,the growing season will approximate the photosynthetic period of vegetation,and in turn,should have a closer link with satellite-sensor-derived greenness development than its early forms.
First,we calculated the case,frequency,and cumu-lative frequency of leaf unfolding and leaf coloration of all plants within every 5-day period throughout each year (from January 1to December 31)at each phenological station,and then drew the cumulative frequency curves of leaf unfolding and leaf coloration,respectively.Next,we determined growing season
beginning and end dates on the cumulative frequency curve by computing interpolation dates when the cumulative frequency of leaf unfolding and leaf coloration reaches 50%for each station and each year (Fig.2).As the phenological data at Harbin in 1992and 1993and at Xi’an in 1992were missing,we estimated the growing season beginning and end dates for both stations and the above years using a spatial interpola-tion technique (Yang &Chen,1995).The reliability of interpolation can be partially documented by spatial sequences of growing season beginning and end dates at sample sites in the corresponding years (not showed).Moreover,we used growing season beginning and end dates as time markers to determine the corresponding threshold NDVI values on the NDVI curves at pixels overlaying phenological stations from year to year (Fig.3).Table 1shows that beginning of the growing season is earlier at southern stations and later at northern stations,but end of the growing season is earlier at northern stations and later at southern stations.Generally,the threshold NDVI values at the beginning of the growing season in the south are larger than in the north,whereas the threshold NDVI values at the end of the growing season in the north are larger than in the south.
Statistical extrapolations of the growing season in temperate eastern China
To extrapolate the growing season beginning and end dates to meteorological stations using threshold NDVI values obtained at the seven phenological stations,the spatial extrapolation area of each phenological station
00.10.20.30.40.50.60.70.80.91C u m u l a t i v e f r e q u e n c y
5-day period (pentad)
Fig.2Determining growing season beginning and end dates at a phenological station.
1122X.C H E N et al .
r 2005Blackwell Publishing Ltd,Global Change Biology ,11,1118–1130
was identi?ed.As one of the preconditions for spatial extrapolation is that the annual NDVI curves at the pixels overlaying the extrapolation sites should be coincident with that at the pixel overlaying a phenolo-gical station (Chen et al .,2001),a spatial extrapolation area can be determined by comparing the coincidence of the NDVI curves among the corresponding pixels.For this study,the NDVI data matrix consisted of 127rows (representing stations/pixels)and 36columns (representing 10-day periods).We measured the coin-cidence of the NDVI curves using the Euclidean distance (d ik ):
d ik ??????????????????????????????X
36j ?1
ex ij àx kj T2v u u t
;where x ij and x kj are the 10-day peak NDVI values at the
i th pixel and the k th pixel,and j is the ordinal number of the 10-day periods in a year.We carried out a hierarchical cluster analysis to de?ne extrapolation areas for phenological stations in each year.Considering the appropriate number of ?nal clusters and the signi?cant difference of d ik values between clustering stages,we assigned the maximum d ik for determining the ?nal clusters between 0.52and 0.78during 1982–1993.Generally,the smaller the maximum d ik of a cluster in a given year,the higher the coincidence of NDVI curves of the pixels within the cluster.As a result,the number of extrapolation areas changes between 4and 7.Here,we assume that the threshold NDVI values indicating the growing season beginning and end at the extrapola-tion sites are the same as those at the phenological station within an extrapolation area and in a given year.Therefore,we can directly estimate the growing season beginning and end dates at the extrapolation sites on their NDVI curves from 1982to 1993using the yearly threshold NDVI values at the phenological stations.In order to decrease the bias,the precise beginning and end dates were set at the center of the 10-day periods,namely the sixth day.The time series of growing season parameters at the seven phenological stations and the 82extrapolation sites (Fig.4)with more than 8years of data provide the basis for further analyses.
To detect the growing season variability in response to climate change,we carried out two types of analyses:spatial pattern analysis and temporal trend analysis.For the former analysis,we revealed the spatial relationship between average growing season parameters (beginning and end date)and mean air temperatures at the sites over the 1982–1993period.For the latter analysis,because the importance of nonclimatic mechanisms determining growing season changes should decrease with increasing scale (Parmesan &Yohe,2003),we focused on the trend analysis at both zonal and regional scales.First,we divided the research region into ?ve latitudinal zones:32–34.991N (zone 1),35–37.991N (zone 2),38–40.991N (zone 3),41–43.991N (zone 4),and !441N (zone 5).Secondly,we used the station time series to create the average zonal and regional time https://www.sodocs.net/doc/4c3627448.html,stly,we calculated simple linear trends for these time series and tested their signi?cance using Pearson’s correlation coef?cients and t -tests.
Results and discussion
Spatial relationship between growing season and air temperature
Generally,average beginning dates of the local growing season progress from south to north and from coastal
N D V I 10-day period
Fig.3Determining corresponding threshold NDVI values at a phenological station.
Table 1Mean (day of year)and standard deviation (days)of BGS and EGS and corresponding threshold normalized difference vegetation index (NDVI)values at phenological stations from 1982to 1993
Station Parameter BGS NDVI at BGS EGS NDVI at EGS Harbin Mean 1270.1712690.370SD 50.03150.042Mudanjiang Mean 1240.3812680.612SD 50.10130.044Gaixian Mean 1120.1532870.341SD 50.026130.103Beijing Mean 1050.1752980.256SD 50.06650.057Xingtai Mean 1030.3342920.258SD 70.08170.033Luoyang Mean 1000.4332950.289SD 30.10960.066Xi’an
Mean 980.4472990.278SD
4
0.093
5
0.089
BGS,beginning of the growing season;EGS,end of the growing season.
VA R I A T I O N O F P H E N O L O G I C A L G R O W I N G S E A S O N I N C H I N A
1123
r 2005Blackwell Publishing Ltd,Global Change Biology ,11,1118–1130
areas(lower altitudes)to inland areas(higher alti-tudes).In contrast,average end dates of the local growing season progress roughly from north to south and from inland areas(higher altitudes)to coastal areas (lower altitudes).Local heat and energy regimes are the cause of this spatial progression pattern.
A spatial correlation and regression analysis between average growing season parameters and mean air temperatures prior to and during the beginning and end dates of the growing season at all sites shows that beginning date of the growing season correlates closely with mean temperature from March to May (R250.6966,n587,P o0.001).The multinomial simu-lation(Fig.5)shows that the dependence of beginning of the growing season on mean temperature is much stronger at sites with March–May temperature above 81C(linear correlation coef?cient r5à0.8276,n551, P o0.001)than at sites with March–May temperature below81C(linear correlation coef?cient r5à0.0454, n536,P40.1).The negative correlation indicates that the higher the mean temperature during March and May at a site,the earlier the average growing season beginning time.In contrast,there is a highly linear correlation between end date of the growing season and mean temperature from August to October(r50.8204, n587,P o0.001).The positive correlation indicates that the higher the mean temperature during August and October at a site,the later the average growing season end time(Fig.6).
Growing season trends and interannual variability
An advance of0.7–1.7days per year in growing season beginning dates is detectable from zones2to5with a relatively low signi?cance level(Fig.7a),whereas a delay in end dates occurred in all?ve zones and a
N
Fig.4Location of phenological stations and extrapolation sites and excluded sites for extrapolation.
1124X.C H E N et al.
r2005Blackwell Publishing Ltd,Global Change Biology,11,1118–1130
signi?cant delay of0.9–1.9days per year appeared in zones2–4(Fig.7b).The growing season duration was, therefore,signi?cantly lengthened by1.4–3.6days per year from zones2to5(Fig.7c).A reversed trend in beginning dates and a slightly consistent trend in end dates were observed in zone1.The stronger linear trends in all higher latitudinal zones indicate that the growing season has a widespread and concordant pattern of change.
However,the northernmost zone did not show larger linear trends towards advanced spring events or delayed autumn events than southern zones did as global meta-analyses suggest(Root et al.,2003).The attribution of the weakened trends in zone5is complicated because physiological and environmental factors dominate local phenological changes.
Possible 50
70
90
110
130
150
170
190
Mean temperature (°C)
D
a
y
o
f
y
e
a
r
Fig.5Spatial relationships between March–May mean tem-
peratures and average growing season beginning dates from
1982to
1993.
220
240
260
280
300
320
Mean temperature (?C)
D
a
y
o
f
y
e
a
r
Fig.6Spatial relationships between August–October mean
temperatures and average growing season end dates from
1982to1993.
(A)
Fig.7(A)Linear trends of growing season beginning date.(a)
zone1;(b)zone2;(c)zone3;(d)zone4;(e)zone5;(f)whole
area;*P o0.1.(B)Linear trends of growing season end date.(a)
zone1;(b)zone2;(c)zone3;(d)zone4;(e)zone5;(f)whole
area;**P o0.05;***P o0.01;****P o0.001.(C)Linear trends of
growing season length.(a)zone1;(b)zone2;(c)zone3;(d)zone
4;(e)zone5;(f)whole area;**P o0.05;***P o0.01.
VA R I A T I O N O F P H E N O L O G I C A L G R O W I N G S E A S O N I N C H I N A1125 r2005Blackwell Publishing Ltd,Global Change Biology,11,1118–1130
(B)
Fig.7(Continued
)
(C)
1126X.C H E N et al.
r2005Blackwell Publishing Ltd,Global Change Biology,11,1118–1130
ecological explanations for the reduced trend of growing season beginning dates in the highest latitu-dinal zone are regional warming in winter(Table2)and physiological response of plants to the regional warm-ing.As certain cultivars show,woody plants have higher chilling requirements for resuming growth after winter dormancy in cold climates relative to warm climates(Powell,1986).If chilling cannot be completely satis?ed because of signi?cantly increasing tempera-tures during the dormancy period,then more heat units (GDD)are required for the onset of plant growth in spring(Couvillon&Erez,1985;Cannell&Smith,1986; Powell et al.,1986;Murray et al.,1989),which could slow the advancement of spring phenophases in the highest latitudinal zone under a less pronounced temperature increase in spring(Table2).
Linear trends were also observed at the regional scale:growing season beginning date advanced by0.4 days per year with low signi?cance(Fig.7a),whereas end date was signi?cantly delayed by1day per year and length was signi?cantly extended by1.4days per year(Fig.7b and c).The signi?cant extension of growing season duration by1.4days per year for our ground-based time series(1982–1993)is consistent with time series of satellite observations in the northern high latitudes(1981–1991,12days per decade,Myneni et al., 1997)and in Eurasia(1981–1999,10days per decade, Zhou et al.,2001).The trend in growing season beginning date in temperate eastern China is also consistent with the trend in growing season beginning date in Europe(Chmielewski&Roetzer,2001),but is not as strong as the trend of?rst-leaf spring index date in eastern North America over roughly the same period (1980–1990,9days per decade,Schwartz,1998).In contrast,the trend in growing season end date in temperate eastern China is much stronger than that in Europe(Chmielewski&Roetzer,2001).
In terms of interannual variability,the earliest growing season beginning date appeared in the end of1980s(1988or1989),whereas the latest end date occurred in the beginning of1990s(1991or1992)in most zones and the whole area(Fig.7a and b).A signi?cant correlation in growing season duration exists between adjacent zones,such as,between zones 2and3(Pearson’s correlation r50.754,P o0.01), between zones3and4(r50.839,P o0.001),and between zones4and5(r50.589,P o0.05).However, the spatial similarity of time series of growing season duration decreases rapidly as the distance between latitudinal zones increases in temperate eastern China (not shown).
Temporal relationship between growing season and air temperature
The growing season extension we detect in this study corresponds with an increasing trend in annual mean temperature in northeast China from1951to1994 (Wang&Gaffen,2001),and in different latitudinal zones and whole study area from1982to1993(Table2). Examination of monthly mean temperatures shows a warming from December to March and in July and
Table2Linear trends(1C/year)of monthly and seasonal and annual mean temperatures in different zones and for the whole area from1982to1993
Period Zone1Zone2Zone3Zone4Zone5Area December10.1010.1510.1410.1110.0410.11 Januaryà0.0110.0410.0910.1410.1110.07 February10.22w10.3810.26w10.32w10.35w10.31w Winter10.1010.1910.1710.1910.1710.16 Marchà0.0210.0410.1410.2110.3210.14 April10.01à0.03à0.04à0.10à0.06à0.04 Mayà0.18wà0.14wà0.06à0.0110.05à0.07* Springà0.06à0.0410.0110.0310.1010.01 June10.01à0.02à0.04à0.11wà0.07à0.05 July10.1310.0410.01à0.0210.0110.03 Augustà0.08à0.0410.01à0.03à0.05à0.04 Summer10.02à0.00à0.01à0.05à0.04à0.02 September10.1010.1110.0510.0310.0010.06 Octoberà0.04à0.09à0.08à0.0010.25w10.01 Novemberà0.10à0.08à0.07à0.0410.03à0.05 Autumnà0.01à0.02à0.03à0.0110.0910.00 Year10.0110.0210.0310.0410.0810.04 *P o0.1,w P o0.05.
VA R I A T I O N O F P H E N O L O G I C A L G R O W I N G S E A S O N I N C H I N A1127 r2005Blackwell Publishing Ltd,Global Change Biology,11,1118–1130
September,with the most signi?cant temperature increase in February,and cooling from April to June and in August and November in most zones and the entire study area(Table2).The unsteady changes of temperature trends from February to April may result in the slight advancement of growing season beginning dates,whereas the dominant cooling from May to August may cause the apparent delay of end dates.Just as buds have a required heat accumulation for?ushing, leaf life and senescence are also mediated by heat accumulation(Worrall,1998).A later end of the growing season,as judged by later leaf coloration dates,would have been because of the leaves’heat sum not being met as early as in previous years,because of cool summer temperatures,rather than a mild autumn (Worrall,1999).Nevertheless,more detailed studies and explanations are needed to understand the physiological and ecological mechanisms of this rela-tionship.
On an interannual basis,correlation analysis shows that growing season beginning dates were mainly in?uenced by mean air temperatures from February to April,especially in zones2and3and the whole area (Table3).A negative correlation indicates that higher mean temperatures in late winter and spring trigger an earlier onset of the growing seasons of local plant communities.In contrast to beginning dates,a sig-ni?cant correlation between growing season end dates and mean air temperatures from May to June was detectable from zones2to4,as well as across the entire study area(Table3).The overall negative correlation indicates that lower mean temperatures during late spring and early summer induce a later end of the growing season.A similar result was also reported in Germany(Menzel,2003).Conclusions
The current study has employed spatial extrapolation of growing season parameters from several sample sites to an area.Spatial patterns of the growing season beginning and end dates correlate signi?cantly with spatial patterns of mean air temperatures in spring and autumn,respectively.Contrasting with results from similar studies in Europe and North America,our study suggests that(1)the growing season and the photosynthetic period of land vegetation in temperate eastern China has signi?cantly extended during a shorter period of time,(2)this extension is attributed mainly to delayed end dates instead of the advanced beginning dates,and(3)the apparently delayed trends in end dates are associated with dominant regional cooling during late spring and summer months,while the insigni?cantly advanced trends in beginning dates correspond to inconsistent regional warming during late winter and spring months.On an interannual basis, growing season beginning and end dates correlate closely with mean air temperatures from February to April and from May to June,respectively.These ?ndings imply that there are diverse seasonal response patterns by land vegetation to recent climate change in different parts of the world.Therefore,the results presented here provide new phenological evidence of climate change impacts from eastern Eurasia,and support the conclusion by the Intergovernmental Panel on Climate Change that climate change is already affecting living systems(McCarthy et al.,2001). Further research may explore the possibility of temporal extrapolation of growing season parameters for years after1993.In this way,we may carry out the spatial–temporal estimate of ground-based growing season at regional scales using plant phenology and satellite data.
Acknowledgements
The authors wish to thank Fuchun Zhang at CAS for providing the phenological data from1989to1993and Mark D.Schwartz at University of Wisconsin-Milwaukee for kindly revising the manuscript.Thanks to Weifeng Pan,Lin Ma,and Wenken Tan for processing phenology and satellite data.This research is funded by the National Natural Science Foundation of China under Grant No.40371042.
References
Achard F,Blasco F(1990)Analysis of vegetation seasonal evolution and mapping of forest cover in West Africa with the use of NOAA AVHRR HRPT data.Photogrammetric Engineering and Remote Sensing,56,1359–1365.
Table3Correlation coef?cients(r)between growing season
parameters and mean air temperatures in different zones and
for the whole area from1982to1993
Correlation
variables r(zone1)r(zone2)r(zone3)r(zone4)r(zone5)r(area)
BGS-T210.07à0.64wà0.47à0.45à0.33à0.44
BGS-T3à0.21à0.71zà0.54*10.0010.02à0.39
BGS-T4à0.47à0.18à0.47à0.12à0.26à0.48
BGS-T2–4à0.28à0.78zà0.63wà0.24à0.19à0.54*
EGS-T510.30à0.58wà0.44à0.03à0.00à0.60w
EGS-T610.1210.01à0.42à0.63wà0.34à0.34
EGS-T5,610.30à0.50*à0.53*à0.50*à0.32à0.61w
T2,T3,T4,and T2–4,mean air temperatures in February,March,
April,and from February to April.
T5,T6,and T5,6,mean air temperatures in May,June,and from
May to June.
*P o0.1,w P o0.05,z P o0.01.
1128X.C H E N et al.
r2005Blackwell Publishing Ltd,Global Change Biology,11,1118–1130
Ahas R(1999)Long-term phyto-,ornitho-and ichthyophenolo-gical time-series analyses in Estonia.International Journal of Biometeorology,42,119–123.
Badeck FW,Bondeau A,Boettcher K et al.(2004)Responses of spring phenology to climate change.New Phytologist,162, 295–309.
Beaubien EG,Freeland HJ(2000)Spring phenology trends in Alberta,Canada:links to ocean temperature.International Journal of Biometeorology,44,53–59.
Botta A,Viovy N,Ciais P et al.(2000)A global prognostic scheme of leaf onset using satellite data.Global Change Biology,6, 709–725.
Bradley NL,Leopold AC,Ross J et al.(1999)Phenological changes re?ect climate change in Wisconsin.Proceedings of National Academy of Sciences USA:Ecology,96,9701–9704. Brandtner E(1958)Methodische Untersuchungen an phaenolo-gischen Beobachtungen unter besonderer Beruecksichtigung phyto-pathologischer Probleme.Berichte des Deutschen Wetterdienstes Nr.47,Offenbach,14pp.
Cannell MGR,Smith RI(1986)Climatic warming,spring budburst and frost damage on trees.Journal of Applied Ecology, 23,177–191.
Chen XQ(1994)Untersuchung zur zeitlich-raeumlichen Aehnlichkeit von phaenologischen und klimatologischen Parametern in West-deutschland und zum Ein?uss geooekologischer Faktoren auf die phaenologische Entwicklung im Gebiet des Taunus.Berichte des Deutschen Wetterdienstes Nr.189,Offenbach,116pp.
Chen XQ(2003a)Assessing phenology at the biome level.In: Phenology:An Integrative Environmental Science(ed.Schwartz MD),pp.285–300.Kluwer Academic Publishers,Dordrecht. Chen XQ(2003b)East Asia.In:Phenology:An Integrative Environmental Science(ed.Schwartz MD),pp.11–25.Kluwer Academic Publishers,Dordrecht.
Chen XQ,Pan WF(2002)Relationships among phenological growing season,time-integrated Normalized Difference Ve-getation Index and climate forcing in the temperate region of Eastern China.International Journal of Climatology,22,1781–1792.
Chen XQ,Tan ZJ,Schwartz MD et al.(2000)Determining the growing season of land vegetation on the basis of plant phenology and satellite data in Northern China.International Journal of Biometeorology,44,97–101.
Chen XQ,Xu CX,Tan ZJ(2001)An analysis of relationships among plant community phenology and seasonal metrics of Normalized Difference Vegetation Index in the northern part of the monsoon region of China.International Journal of Biometeorology,45,170–177.
Chmielewski FM,Roetzer T(2001)Response of tree phenology to climate change across Europe.Agricultural and Forest Meteorology,108,101–112.
Chu CC(1964)Phenology and agricultural production.New Construction,188–189,142–149(in Chinese).
Chu CC,Wan MW(1973)Phenology.Science Press,Beijing(in Chinese).
Compilation Committee on the Vegetation of China(1980)The Vegetation of China.Science Press,Beijing(in Chinese). Couvillon GA,Erez A(1985)In?uence of prolonged exposure to chilling temperatures on bud break and heat requirement for
bloom of several fruit species.Journal of the American Society for Horticultural Science,110,47–50.
Duchemin B,Goubier J,Courrier G(1999)Monitoring pheno-logical key stages and cycle duration of temperate deciduous forest ecosystems with NOAA/AVHRR data.Remote Sensing of Environment,67,68–82.
Fisher A(1994)A model for the seasonal variations of vegetation indices in coarse resolution data and its inversion to extract crop parameters.Remote Sensing of Environment,48,220–230. Fitter AH,Fitter RSR(2002)Rapid changes in?owering time in British plants.Science,296,1689–1691.
Gao YX,Xu SY,Guo QY,Zhang ML(1962)Some Aspects on the Monsoon over East Asia.Science Press,Beijing(in Chinese). Goward SN,Markham B,Dye DG et al.(1991)Normalized difference vegetation index measurements from the advanced very high resolution radiometer.Remote Sensing of Environ-ment,35,257–277.
Haenninen H(1990)Modelling bud dormancy release in trees from cool and temperate regions.Acta Forestalia Fennica,213,47. Holben BN(1986)Characteristics of maximum-value composite images from temporal AVHRR data.International Journal of Remote Sensing,7,1417–1434.
Hopkins AD,Murray MA(1933)Natural guides to the beginning,length and progress of the seasons.Acta Phaeno-logica,2,33–43.
Institute of Geography at Chinese Academy of Sciences(1965) Chinese Yearbook of Animal and Plant Phenological Observation No.1.Science Press,Beijing.
Institute of Geography at Chinese Academy of Sciences(1988) Chinese Yearbook of Animal and Plant Phenological Observation No.8.Geology Press,Beijing.
Institute of Geography at Chinese Academy of Sciences(1989a) Chinese Yearbook of Animal and Plant Phenological Observation No.9.Geology Press,Beijing.
Institute of Geography at Chinese Academy of Sciences(1989b) Chinese Yearbook of Animal and Plant Phenological Observation No.10.Survey and Drawing Press,Beijing.
Institute of Geography at Chinese Academy of Sciences(1992) Chinese Yearbook of Animal and Plant Phenological Observation No.11.Chinese Science and Technology Press,Beijing. Justice CO,Townshend JRG,Holben BN et al.(1985)Analysis of the phenology of global vegetation using meteorological satellite data.International Journal of Remote Sensing,6,1271–1318. Keeling CD,Chin JFS,Whorf TP(1996)Increased activity of northern vegetation inferred from atmospheric CO2measure-ments.Nature,382,146–149.
Kramer K(1996)Phenology and growth of European trees in relation to climate change.Thesis,Landbouw Universiteit Wageningen, 210pp.
Lloyd D(1990)A phenological classi?cation of terrestrial vegetation cover using shortwave vegetation index imagery. International Journal of Remote Sensing,11,2269–2279.
Lucht W,Prentice IC,Myneni RB et al.(2002)Climatic control of the high-latitude vegetation greening trend and Pinatubo effect.Science,296,1687–1689.
Malingreau JP(1986)Global vegetation dynamics:satellite observation over Asia.International Journal of Remote Sensing, 7,1121–1146.
VA R I A T I O N O F P H E N O L O G I C A L G R O W I N G S E A S O N I N C H I N A1129 r2005Blackwell Publishing Ltd,Global Change Biology,11,1118–1130
Markon CJ,Fleming MD,Binnian EF(1995)Characteristics of vegetation phenology over the Alaskan landscape using AVHRR time-series data.Polar Record,31,179–190. McCarthy JJ,Canziani OF,Leary NA et al(eds)(2001)Climate Change2001:Impacts,Adaptations,and Vulnerability.Cambridge University Press,Cambridge.
McMaster GS,Wilhelm WW(1997)Growing degree-days:one equation,two interpretations.Agricultural and Forest Meteoro-logy,87,291–300.
Menzel A(2003)Plant phenological anomalies in Germany and their relation to air temperature and NAO.Climatic Change,57, 243–263.
Menzel A,Fabian P(1999)Growing season extended in Europe. Nature,397,659.
Moulin S,Kergoat L,Viovy N et al.(1997)Global-scale assessment of vegetation phenology using NOAA/AVHRR satellite measurements.Journal of Climate,10,1154–1170. Murray MB,Cannell MGR,Smith RI(1989)Date of budburst of ?fteen tree species in Britain following climatic warming. Journal of Applied Ecology,26,693–700.
Myneni RB,Keeling CD,Tucker CJ et al.(1997)Increased plant growth in the northern high latitudes from1981to1991. Nature,386,698–702.
Newman JE,Beard JB(1962)Phenological observations:the dependent variable in bioclimatic and agrometeorological studies.Agronomy Journal,54,399–403.
Parmesan C,Yohe G(2003)A globally coherent?ngerprint of climate change impacts across natural systems.Nature,421, 37–42.
Pfau R(1964)Varianz-und korrelationsanalytische Untersu-chungen an phaenologischen Phasen im Naturraum06 (Unterbayerisches Huegelland).Meteorologische Rundschau, 17,113–122.
Powell LE(1986)The chilling requirement in apple and its role in regulating time of?owering in spring in cold-winter climates.Acta Horticulturae,179,129–139.
Powell LE,Swartz HJ,Pasternak G et al.(1986)Time of?owering in spring:its regulation in temperate zone woody plants. Biologia Plantarum(PRAHA),28,81–84.
Reed BC,Brown JF,VanderZee D et al.(1994)Measuring phenological variability from satellite imagery.Journal of Vegetation Science,5,703–714.
Romberger JA(1963)Meristems,growth and development in woody https://www.sodocs.net/doc/4c3627448.html,DA Technical Bulletin,Nr.1292,US Government Printing Of?ce,214pp.
Root TL,Price JT,Hall KR et al.(2003)Fingerprints of global warming on wild animals and plants.Nature,421,57–60.Schnelle F(1955)P?anzen-Phaenologie.Akademische Verlagsge-sellschaft Geest&Portig K.-G,Leipzig.
Schwartz MD(1998)Green-wave phenology.Nature,394,839–840. Schwartz MD,Chen XQ(2002)Examining the onset of spring in China.Climate Research,21,157–164.
Schwartz MD,Reed BC,White MA(2002)Assessing satellite-derived start-of-season measures in the conterminous USA. International Journal of Climatology,22,1793–1805.
Schwartz MD,Reiter BE(2000)Changes in North American spring.International Journal of Climatology,20,929–932. Sparks TH,Jeffree EP,Jeffree CE(2000)An examination of the relationship between?owering times and temperature at the national scale using long-term phenological records from the UK.International Journal of Biometeorology,44,82–87.
Viovy N,Arino O,Belward AS(1992)The best index slope extraction(BISE):a method for reducing noise in NDVI time-series.International Journal of Remote Sensing,13,1585–1590.
Viovy N,Saint G(1994)Hidden Markov models applied to vegetation dynamics analysis using satellite remote sensing. IEEE Transactions on Geoscience and Remote Sensing,32,906–917. Walther GR,Post E,Convey P et al.(2002)Ecological responses to recent climate change.Nature,416,389–395.
Wang JXL,Gaffen DJ(2001)Late-twentieth-century climatology and trends of surface humidity and temperature in China. Journal of Climate,14,2833–2845.
White MA,Thornton PE,Running SW(1997)A continental phenology model for monitoring vegetation responses to interannual climatic variability.Global Biogeochemical Cycles, 11,217–234.
Worrall J(1998)Autumn leaf coloration.Forest Chronology,74, 668–669.
Worrall J(1999)Phenology and the changing seasons.Nature, 399,101.
Yang GD,Chen XQ(1995)Phenological Calendars and their Applications in the Beijing Area.Capital Normal University Press,Beijing(in Chinese).
Zhang JC,Lin ZG(1985)The Climate of China.Shanghai Science and Technology Publishers,Shanghai(in Chinese).
Zhang XY,Friedl MA,Schaaf CB et al.(2004)Climate controls on vegetation phenological patterns in northern mid-and high latitudes inferred from MODIS data.Global Change Biology,10, 1133–1145.
Zhou L,Tucker CJ,Kaufmann RK et al.(2001)Variations in northern vegetation activity inferred from satellite data of vegetation index during1981to1999.Journal of Geophysical Research–Atmospheres,106,20069–20083.
1130X.C H E N et al.
r2005Blackwell Publishing Ltd,Global Change Biology,11,1118–1130