Reduced Soil Moisture Contributes to More Intense and More Frequent Heat Waves in Northern China

ZHANG Jie,LIU Zhenyuan,and CHEN Li

Key Laboratory of Meteorological Disasters of Ministry of Education, Nanjing University of Information Science&Technology,Nanjing 210044

Reduced Soil Moisture Contributes to More Intense and More Frequent Heat Waves in Northern China

ZHANG Jie∗,LIU Zhenyuan,and CHEN Li

Key Laboratory of Meteorological Disasters of Ministry of Education, Nanjing University of Information Science&Technology,Nanjing 210044

Heat waves have attracted increasing attention in recent years due to their frequent occurrence.The present study investigates the heat wave intensity and duration in China using daily maximum temperature from 753 weather stations from 1960 to 2010.In addition,its relationships with soil moisture local forcing on the ten-day period and monthly scales in spring and summer are analyzed using soil moisture data from weather stations and ERA40 reanalysis data.And finally,a mechanistic analysis is carried out using CAM5.1(Community Atmosphere Model,version 5.1)coupled with CLM2(Community Land Model,version 2).It is found that the heat wave frequency and duration show a sandwich distribution across China,with high occurrence rates in Southeast China and Northwest China,where the maximum frequency and duration exceeded 2.1 times and 9 days per year,respectively.The increasing trends in both duration and intensity occurred to the north of 35◦N.The relationships between heat wave frequency in northern China in July(having peak distribution)and soil moisture in the earlier stage(from March to June)and corresponding period(July)are further analyzed,revealing a strong negative correlation in March,June and July,and thus showing that soil moisture in spring and early summer could be an important contributor to heat waves in July via positive subtropical high anomalies.However,the time scales of influence were relatively short in the semi-humid and humid regions,and longer in the arid region.The contribution in the corresponding period took place via positive subtropical high anomalies and positive surface skin temperature and sensible heat flux anomalies.

heat wave,soil moisture,multiple time scales,heat wave frequency,heat wave duration

1.Introduction

Heat waves are natural disasters that impact upon human mortality,regional economies,and ecosystems(WMO, 2003;Retalis et al.,2010;Kotroni et al.,2011),and therefore have attracted increasing attention in the past two decades due to their frequent occurrence(Meehl and Tebaldi,2004). In 2003,a heat wave process covering the whole of Europe increased the summer mortality rate in France by up to 54%, and caused EUR 12.3 m of agricultural losses and EUR 1.6 m of forest losses in central,western and southern Europe (Heck et al.,2001;H´emon and Jougla,2004).Elsewhere,a heat wave event in the U.S in 2012 resulted in 42 casualties; and in Eastern China,the most serious heat wave on record occurred in 2013,which lasted for 38 days—more than double the average duration in Zhejiang province over the past 65 years(http://zj.weather.com.cn/tqyw/01/2048435.shtml). Due to global warming,heat waves are becoming more frequent and more intense in the U.S.(Meehl et al.,2001; Meehl and Tebaldi,2004).Moreover,model simulations suggest that heat waves will increase most in the western,upper mid-western,northeastern,and southern U.S.in the future(Dai et al.,2001;Ebi and Meehl,2007).Therefore,frequent heat waves have already influenced society.However, at present,there is no universal definition of a heat wave.It is well known that heat waves are associated with particularly hot and sustained temperatures,and on these grounds, a heat wave has been commonly defined.However,different standards exist in different regions,such as:a period of at least three consecutive days above 32.2◦C in the U.S.(Tamrazian et al.,2008);at least five consecutive days above 25◦C, with three of these above 30◦C,in Europe(Baldi et al.,2006); and a period of at least three consecutive days above 35◦C in China(Deng et al.,2009;Zhang et al.,2011).The World Meteorological Organization(WMO)suggests a period of at least five consecutive days when the daily maximum temperature(Tmax)exceeds its climatology(Tmax,clim)by 5◦C(Frich et al.,2002).

Heat waves are influenced by many factors,including greenhouse gases,remote forcing such as sea surface skin temperature(SST)and the El Ni˜no–Southern Oscillation(ENSO),Arctic ice cover,and local responses(Dai et al., 2001;Meehl and Tebaldi,2004;Zhang et al.,2013,2014). SST is an important forcing of heat waves,and west coast cities in the U.S.in mid-summer are influenced by high Pacific SSTs,aided by the high-pressure systems in the subtropical Pacific,which tend to stall over the western U.S. (Meehl and Tebaldi,2004;Ziv et al.,2004).The Atlantic Multi-decadal Oscillation(AMO),related to Atlantic SST, is associated with heat waves in South America(Sutton and Hodson,2005).SST anomalies in the Indian and Mediterranean oceans resulted in heat waves in Europe and Africa in 2003,with the Mediterranean SST being attributed to the early stages of the heat wave and Indian Ocean SST contributing to the extension of the heat wave into August 2003 (Ziv et al.,2004;Sutton and Hodson,2005).All these findings demonstrate the remote forcing of SST on heat waves over continents.As a forcing from higher latitude,arctic ice loss also results in a positive geopotential height anomaly over Okhotsk,Lake Baikal,and most of Europe,which promotes extreme drought and heat wave conditions in Europe and northern China(Zhang et al.,2014).From the atmospheric circulation viewpoint,an anticyclone in the middle troposphere is the main factor.A simulation by Meehl and Tebaldi(2004)showed that the heat waves in France in 2003 and 1995 were directly influenced by an anticyclone at the 500 hPa level,while Gong et al.(2004)showed that a largescale,anticyclone enhancement is an important factor influencing heat waves in East Asia.

In addition to the remote forcing and direct association with atmospheric circulation anomalies,the local forcing of land surface characteristics are another set of factors affecting heat waves.In Weather Research and Forecasting (WRF)model simulations,land–atmosphere coupling suggests that local forcing contributes 30%–70%of the atmospheric temperature and heat wave intensity(Zhang and Wu, 2011).The surface changes affecting heat waves include vegetation degradation,urbanization,and decreasing soil moisture.The urban heat island effect due to urbanization was found to contribute to an increasing of the intensity of heat waves in Shanghai(Tai et al.,2008).In addition,soil moisture is a key surface factor influencing water-heat transfer in land–atmosphere processes.The feedback of soil moisture to the atmosphere has“memory”(Koster and Suarez,2001).Its perturbation may markedly modify the atmospheric pressure field,wind circulation system,and water vapor distribution, thus affecting atmospheric stability,and such effects may last for one season or longer(Khodayar et al.,2014).Soil moisture in northern China and the Tibetan Plateau(TP)has decreased in recent years(Zhang and Zuo,2011),and soil moisture on the TP could modify precipitation intensity and temperature not only over the TP itself,but also over East Asia as a whole through changing the intensity of monsoon circulation(Chow et al.,2008).Soil moisture simulations using CAM3.1(Community Atmosphere Model,version 3.1)and CLM(Community Land Model)have shown that the extent of heat wave regions decreases obviously when the annual anomalies of soil moisture are not considered in the model (Chen and Zhou,2013).Simulations of the 2003 heat wave have shown that spring soil moisture contributed to 40%of the probability of the heat wave events(Fischer et al.,2007). Under global warming,the soil moisture in spring and summerin the Northern Hemisphere ischanging(Chen and Zhou, 2013),which leads us to ask how moisture changes might affect the distribution of heat waves,their intensity,and frequency.This question remains open.

China lies in the humid,semi-humid,semi-arid,and arid climates of the Northern Hemisphere;and due to the common influence of monsoon and the westerly climate,soil moisture in China is sensitive to climate change(Koster et al.,2004; Wang et al.,2011),such that its spatiotemporal distribution could affect the distribution of heat waves,their intensity, and frequency.To assess the frequency and intensity of heat waves and the associated mechanisms,the numbers of heat wave events,their durations and trends are analyzed in this paper.The study also analyzes the soil moisture contributions to heat wave frequency.Data comprising daily temperature maxima and soil moisture from 58 weather stations are described in section 2.European Center for Medium-Range Weather Forecasts(ECMWF)soil moisture data are used for analyzing the spatial distribution of correlation coefficients between soil moisture and heat waves,and these data are also described in section 2.The spatiotemporal variability of heat wave factors and their relationships with soil moisture are described in section 3.To quantify the effects of soil moisture on heat waves in northern China,the results of simulation experiments using the NCAR(National Center for Atmospheric Research)Community Atmosphere Model, version 5.1(CAM5.1)are presented in section 3.Finally,a summary and conclusions are provided in section 4.

2.Data and methods

2.1.Datasets

Three datasets are used in the study:daily maximum temperature from 753 weather stations from 1960 to 2010;10 cm soil moisture from 58 weather stations from 1981 to 2001; and soil moisture data from the ECMWF(ERA40)dataset from 1960 to 2002.According to the definition applied by the China Meteorological Administration(CMA),a heat wave event in China is defined as a period of at least three consecutive days with temperatures above 35◦C(Deng et al.,2009). In order to further study the heat wave intensity,the durations of heat waves were calculated,where the duration is the number of heat wave days for a heat wave event.The heat wave frequency is the total number of heat wave events in a year. The heat wave duration and frequency were calculated using the daily maximum temperature data.To analyze the relationships between soil moisture and heat wave frequency and intensity,the 10 cm soil moisture data from the ERA40 dataset were used for investigating the climate zones and the spatial differences of the correlation coefficients between soil moisture and heat waves.The soil moisture from the ECMWF analysis was validated using observed soil moisture from 58weather stations before it is used in the study.These stations were regarded as representative stations.The data were first used for validating the soil moisture data from ERA40,and secondly for calculating the correlation coefficients between the heat wave frequency in July and the 10-day soil moisture content from March to July,in different typical climate zones.The climate zones were divided by empirical orthogonal functions(EOFs)and rotated empirical orthogonal functions(REOFs)of soil moisture from ERA40.The ERA40 grid and the distribution of weather stations are shown in Fig.1.

2.2.Methodology

The soil moisture data from ERA40 were in the form of volumetric water content(units:cm3cm−3);however,the soil moisture data from the 58 weather stations were in the form of the weight of the water content(units:%).Conversion of the soil moisture from weight of water content to volumetric water content was performed using the formula

whereθmis the weight of the water content(units:%),θvis the volumetric water content(units:cm3cm−3),andρis the dry density(g cm−3).The changes ofρwere very small compared withθm;therefore,it was regarded as a constant for,and obtained from,individual stations.

The 10 cm soil moisture data from ERA40 were interpolated to weather stations using a cubic method(tested using the 10 cm soil moisture of in-situ measurements;Fig.2)according to climate zones:17,19 and 12 soil stations in Northeast China(34◦–55◦N,100◦–123◦E),northern China(40◦–55◦N,123◦–135◦E),and Jianghuai basin(30◦–34◦N,110◦–125◦E)were used for the validations.The correlation coefficients with ERA40 in summer were 0.97,0.95,and 0.97 for Northeast China,northern China,and Jianghuai basin, respectively;and in spring they were 0.96,0.88,and 0.94, respectively.All correlations were statistically significant at the 95%confidence level.However,there were obvious differences between them,which should be validated.The validated coefficients in linear regression were 2.73,0.8,and 1.1 in spring,and 7.34,0.42,and 0.74 in summer,respectively; the biases were 0.0014,0.0026,and−0.0005 in spring,and−0.004,−0.002,and−0.002 in summer,respectively.All in all,in Northeast China,when soil moisture was larger than the average value(wet),soil moisture from ERA40 was less, but it was larger when soil moisture was less than the average value(dry).However,the opposite was the case in the other regions.

Because soil moisture showed a heterogeneous distribution,climatic sub-regions of soil moisture were defined using EOFs,REOFs,and the validated soil moisture from ERA40. In addition,correlation coefficients and trend analysis were used in the study.Reliability tests were also performed,using the Monte Carlo test.It has been proven that the Monte Carlo method is better than Pearson’s correlation.

2.3.The CAM5 model and experiments

The atmospheric module of the Community Earth System Model,i.e.CAM 5.1(Neale et al.,2011),was developed at the NCAR.This study applied the finite volume dynamic framework with a horizontal resolution of 1.9◦(lat)and 2.5◦(lon),with 30 vertical layers in theσ-p vertical coordinate.CAM5.1 was coupled with the CLM2 land process model so as to simulate the feedback of soil moisture on the atmosphere.In this study,several physical processes,including radiation processes,cloud effects,convection,boundary layer effects and other physical processes,were represented in the model according to the default options.A detailed description of the model is available at http://www.ccsm.ucar. edu/models/atm-cam/.The simulation ability of CAM3.1 in terms of the soil moisture effects on extreme temperature has been evaluated(Zhou and Chen,2012),and it was found that CAM3.1 could generally reproduce the basic features of the large-scale spatial patterns of annual-mean extreme climate indices over East Asia,although the bias in extreme temperature is large.CAM5.1 is an improved version based on version 3.Wang(2014)used and tested CAM5.1 for soil moisture simulation in East Asia and found the model shows a unanimously similar spatial distribution as observed.Therefore,it can be used for further soilmoisture sensitivity experiments in East Asia.To quantify the effects of soil moisture on atmospheric circulation,local temperature and heat waves in northern China,only the soil moisture data were changed in the sensitivity experiments when running CAM5.1.Because soil moisture showed a heterogeneous distribution,the sensi-tivity region of(34◦–55◦N,100◦–123◦E)was ensured based on the soil moisture sub-regions(see Fig.1 for the simulation region).The initial soil moisture and atmospheric data were the mean values from 1960 to 1985,regarded as the climate state,and three experimentswere performed through decreasing soil moisture by 20%in March,June and July.The experiments were started in those three months and ended at the end of August.

3.Results and discussion

3.1.Spatial distribution of heat waves in China

Due to the inconsistent increase of temperature associated with global warming,the intensity and frequency of heat waves therefore show equivalently heterogeneous distributions(Kotroni et al.,2011).Figure 3 shows the spatial distributions of heat wave frequency(Fig.3a)and its trend (Fig.3b),as well as the heat wave duration(Fig.3c)and its trend(Fig.3d)in China,based on daily maximum temperature from 753 weather stations collected from 1960 to 2010.The heat wave frequencies show a sandwich distribution oriented from Southeast China to Northwest China(Fig. 3a).The spatial distributions of the durations of heat waves are the same as those of the frequencies of heat wave events (Fig.3c),showing high values in Southeast China and Northwest China,exceeding 2.1 times and 9 days per year,respectively;the respective minima are 0.8 times and 0.5 days in the semi-humid region.The distributions of the frequency and the duration of heat waves can be classified into clear climatic sub-regions,demonstrating that heat waves are influenced by differentcirculation systems,externalforcing,orlocal effects,as further indicated by the sandwich distributions. Previous work has shown that the western Pacific subtropical high(WPSH)controls Changjiang-Huaihe basin after the Mei-yu period,resulting in a summer drought period due to frequentadiabatic heating underdownward movements ofthe atmosphere(Gong et al.,2004),which is conducive to heat waves.However,heat waves in Northwest China are mainly influenced by stronger circulation systems over land in the mid-latitudes,such as continental high atmospheric pressure. Meanwhile,precipitation is low in the extreme arid regions of Northwest China,such that intense heat,a dry surface and low vegetation cover allow the surface to emit more longwave radiation resulting in a corresponding strong heat flux (Zhang et al.,2011);these factors contribute to the high air temperature and high heat wave frequency.The region of infrequent heat waves in the semi-humid region is under the combined influence of the WPSH and other circulation systems over land in the mid-latitude,because these circulation systemsdirectly influence the precipitation in the semi-humid region(Zhang et al.,2014).

Simulations have shown that North America and most of Europe are likely to face more intense and more frequent heat waves in the future under a global warming background. Similarly,we ask how the occurrence of heat waves might change in China as a result of clear temperature increases with global warming.To address this issue,the differences in heat wave occurrence under normal conditions and global warming conditions are analyzed.The temporal trend in the number of heat wave events reflects the change in heat wave frequency,and the durational trend of heat waves reflects the change in heat wave intensity.Figure 3 shows the trends in heat wave frequencies(Fig.3b)and heat wave durations(Fig. 3d),revealing that while both trends have the same spatial distribution,there are obvious differences between southern China and northern China.These distributions are therefore in contrast to the heat wave frequency(Fig.3a)and heat wave duration(Fig.3b).To the south of 35◦N,there are two high centers and two low centers,where the maximum rates of increase in the heat wave frequency and heat wave duration exceeded 2 times and 10 days per 100 years,respectively,while the minimum rates were less than−2 times and−10 days per 100 years,respectively.To the north of 35◦N,except for parts of the west,the rates of increase are obvious.The patterns of the heat wave frequency and heat wave duration follow that of the low rainfall regions(figure omitted).

3.2.More intense and more frequent heat waves in northern China

In order to address the heat wave intensity,frequency and the possible control factors,the variable trends in the region with increasing occurrence(northern China)are compared with those in the regions with decreasing rates(east of China).Considering that increasing-rate regions include the arid region and the semi-humid region,two centers of increase are selected according to the largest rate of increase in the arid region and semi-humid region,and a center of decrease in the humid region is also selected according to the largest rate of decrease.The representative stations are Ejin Banner in the western Inner Mongolia Autonomous Region (increasing rate,west region),Yan’an(increasing rate,semihumid region)and Xuchang(decreasing rate,humid region). Figure 4 shows the time series of annual heat wave duration and frequency at the three stations.In the arid region,heat wave duration and heat wave frequency have increased in the last 50 years,with peaks in the 1960s,1990s and 2000s when the highest duration and frequency were 20 days and 4 times,respectively.Both indicators were low in the 1970s and early 1990s.The duration and frequency of heat waves both increased steadily from 1960 to 2010 in the semi-humid region, reaching maximum values of8 days and 2 times,respectively. The heat wave duration and frequency show different trends in the arid region and semi-humid region,although both have increasing trends overall.The three regions show different trends in heat wave changes,especially in the humid region, indicating that heat waves could be influenced by different atmospheric circulation systems,external forcing,or other factors.Figure 4 also shows that the heat wave duration and frequency share similar trends in the same region,indicating that increasing frequency contributed to an increasing in the total of heat wave durations.

To further explore the heat wave intensity,Fig.5 shows the heat wave durations for every event for the three representative stations in the last 50 years.The heat wave durations at the three stations have increased.Specifically,the frequency of events with a duration of longer than 4 days increased at the arid station,and that the frequency of events with a duration of longer than 5 days increased at the humid station, reaching a maximum in the 1990s and 2000s.Most of the heat wave durations were more than 4 days at the semi-humid station.Therefore,heat waves were found to have prolonged durations in both northern China and East China.Figure 6 shows the probability distribution of heat wave frequency in summer(from 152 to 242 days).In northern China,heat waves have two peaks,one in mid-June and the other during the last 10 days of July and the first 10 days of August. In the humid region,meanwhile,they occurred mostly before the first 10 days of July and decreased after the second 10 days of July,because the monsoon rain arrived in the latter period.Overall,the period of relatively high probability is July:the total probability in the arid region,semi-arid region, and humid region is 43%,53%,and 36%,respectively,which covers more than the monthly average of 33%.Therefore, heat waves in July can be selected to explore the relationships with local forcing.

3.3.Correlation between soil moisture and heat wave frequency in northern China

Both the heat wave frequency and intensity have increased in northern China,while the heat wave intensity in the east of China has also increased.Therefore,we ask if these increases were directly related to global warming. To address this question,the relationships between local forcing responding to global warming and heat waves are analyzed,with the local forcing including vegetation degradation,urbanization,temperature increase,and soil moisture anomalies.Soil moisture is an important parameter for describing local forcing,because of its influence on the surface water and energy balance.Studies have found that soil moisture feedback on the atmosphere can last for more than 6 months(Sun et al.,2005),indicating the ability of soilmoisture“memory”during the early stage should be considered.Because heat waves in northern China mainly occur in July,the relationships between the heat frequency in July and soil moisture in spring and summer are therefore analyzed.Figure 7 shows the lead–lag correlation coefficients for every 10-day period at the three stations.The results indicate that there was strong negative correlation between soil moisture during the early stage and the corresponding period and the heat wave frequency in July at the arid station.The highest correlation coefficients occurred in March and July,exceeding the 90%confidence level;however,the correlation was weaker from April to June due to precipitation adjustment,i.e.precipitation results in increasing soil moisture from April to June,which decreases the feedback of soil moisture conditions in March.Correlation coefficients at the semi-humid station were high in May,June and July, and those at the humid station were high in late May–early June and July,indicating negative correlation is earlier for two months in the humid station.But this is different from the correlation in the arid region.Therefore,early-season soil moisture could contribute to heat waves in July,but a notable difference is that soil moisture at the arid station could influence summer heat waves over a long time period;however, the influence was relatively short at the semi-humid and humid stations,showing that the“memory”ability of soil moisture content is related to soil moisture and soil properties, because the soil moisture content is the largest difference for the three stations corresponding to three climate belts.

Selecting the high correlation period(March,June and July),and using 10 cm soil moisture from ERA40,the spatial distribution of correlation coef ficients was calculated(Fig. 8).ERA40 data were interpolated to 753 weather stations using a cubic method.Correlation coef ficients in March were consistent with those in May,and were negative between 35◦N and 45◦N,and positive to the south of 35◦N and north of 45◦N.Correlation coef ficients in the other months were negative,except for a few areas in the northeastern and northern regions.The spatial distributions of the correlation coefficients further demonstrate the importance of soil moisture in contributing to heat waves,and soil moisture at the arid station could influence summer heat waves over longer time scales compared with the semi-humid and humid regions.

3.4.How does soil moisture contribute to heat waves?

CAM5.1 coupled with CLM2 was selected to explore the intensity of the effect of soil moisture on heat waves and the mechanisms involved.This was because the model has been previously assessed to be successful in East Asia,especially in terms of its ability to simulate the soil moisture feedback to extreme temperature(Zhou and Chen,2012).Due to the heterogeneous distribution of soil moisture in China,a sensitive region of soil moisture needed to be identified for carrying out the sensitivity experiments.Based on the soil moisture of (27◦–53◦N,73◦–132◦E)from May to July from ERA40,the climate zones of soil moisture were analyzed using EOF and REOF.The totalofthe firsteightload vectorsexplained about 80%of the soil moisture distribution,and the study area was divided into seven sub-regions(Fig.9)based on these firsteight load vectors.The northern region(A)was selected in the sensitivity experiments,because heat waves have become more intense and more frequent in this region in recent years (Fig.4).Considering that the west of the region is an arid region,soil moisture is influenced by human activity in the form of preserving irrigated oases.Moreover,region A is mainly located in eastern China;therefore,only the eastern region(34◦–55◦N,100◦–123◦E)was used for the present experiments,and this region also covered sub-region B.

Three sensitivity experiments were carried out to simulate the feedback between soil moisture and air temperature in July at different time scales;and because the high correlation coefficients(exceeding 90%confidence)were in March, late May–June and July,these months were selected for the simulations.The first two experiments for soil moisture in March and June represent the contributions of early-season soil moisture to air temperature in July,while the later experiment for soil moisture in July represents the soil moisture effects on air temperature during the same terms.The initial soil moisture and atmospheric data were the mean values from 1960 to 1985,regarded as the climate state,and the three experiments were performed through decreasing soil moisture by 20%.The model runs began from March,June and July,respectively,and ended at the end of August;the temporal resolution was one month.Sensitivity experiments and control experiments were both run.When soil moisture in March was decreased by 20%,long-term feedback between moisture in March and atmospheric circulation from March to July was established.Figure 10 shows the anomalies of variables such as geopotential height,air temperature at 2 m,surface skin temperature,and sensible heat flux,in which the anomaly value is the difference between the sensitivity experiment and control experiment results.Figure 10a shows the geopotential height anomaly at 900 hPa in July,displaying a positive anomaly at the China–Mongolia border,and across northern China and the Tibetan Plateau, and a negative anomaly in southern China and the Xinjiang autonomous region in western China.Coincident with the positive geopotential height anomaly,there was a positive air temperature anomaly on the China–Mongolia border,northern China,and over the Tibetan Plateau(Fig.10b),as well as a positive surface skin temperature anomaly(Fig.10c). Meanwhile,the sensible heat flux showed a zonal distribution (Fig.10d),with a positive zone between 40◦N and 45◦N,and another positive zone extending along the southern Tibetan Plateau.The increasing temperature in July shows good relations with the anticyclone anomaly,being mainly influenced by it.This was affected by decreased early-season soil moisture,indicating that decreasing soil moisture is conducive to height and anticyclone anomalies.An enhanced anticyclone contributes to increasing temperature across most of northern China because anticyclone could increase adiabatic heating with downward movement of the atmosphere.When soil moisture in June was decreased by 20%,the change in soil moisture fed back to atmospheric circulation over timescales exceeding one month,and resulted in an influence on atmospheric circulation persisting into July.Figure 10e shows the geopotential height anomaly at 900 hPa in July,revealing a positive anomaly across northern,northeast,and parts of western China,and a negative anomaly across southern China,the Tibetan Plateau,and most of western China.A positive air temperature anomaly(Fig.10f)and a positive surface skin temperature anomaly(Fig.10g)were coincident with the positive geopotential height anomaly.Meanwhile, sensible heat flux was not consistent with geopotential height (Fig.10h),showing a positive center over the north of the Tibetan Plateau and northern China.The distributions of increase air temperature in July were the same as those of decreased soil moisture,despite the one-month time duration.When soil moisture in July was decreased by 20%,it still influenced the atmosphere during this time.On the one hand,soil moisture feeds back to air temperature within the boundary layer,while on the other hand it feeds back to the atmospheric circulation.The atmospheric circulation then influences the air temperature,as can be seen from the distribution of the positive geopotential height anomaly(Fig.10i), positive air temperature anomaly(Fig.10j),and surface skin temperature(Fig.10k).At the same time,the distribution of sensible heat flux shows some consistency with air temperature in the west of 110◦E and south of 45◦N.Overall,soilmoisture in the corresponding period influences heat waves through two processes:(1)feedback to the atmosphere;(2) decreased soil moisture,resulting in less soil heat capacity, which is conducive to increasing soil temperature and further for increasing long wave radiance;it also favors an increase in the temperature difference between the surface and air,and finally for increasing the sensible heat flux in the west of 110◦E,all which contribute to increasing air temperature and heat waves.However,the sensible heat flux distribution is not in accordance with the air temperature distribution and the decreased soil moisture in East China and at the Chinese and Mongolian frontier,demonstrating the relationships between soil moisture and sensible heat flux are complicated.

4.Summary and conclusion

The heat wave frequencies and durations in recent five decades(1960–2010)showed a sandwich distribution with a southeast to northwest orientation.The highest values were found in southern and northwest China,where the maximum heat wave frequency and duration was 2.1 times and 9 days, respectively;the lowest values were found in the semi-humid region,with minima of 0.8 times and 0.5 days,respectively. The heat wave frequency and duration distributions displayed different trends and obvious climatic sub-regions,as further demonstrated by the sandwich distributions,indicating that heat waves are influenced by many factors such as different circulation systems,external forcing,or local effects.

The trends in heat wave frequency and duration shared the same spatial distribution,indicating that the frequency and duration are affected by the same mechanisms.However, there were obvious spatial differences between both southern and northern China.To the north of 35◦N,except forparts of western China,rates in heat wave events and duration matched the rainfall distribution.The trends of heat wave duration and frequency displayed differences between the arid region and semi-humid region.The heat wave frequency increased in northern China and decreased in eastern China,and the increased duration indicates a prolonging of heat waves in northern China.

Because heat waves in northern China mainly occurred in July,with probabilities of 43%,53%,and 36%in the arid region,semi-humid region and humid region,the heat waves in July were selected to analyze the correlation with soil moisture during early stages and the corresponding periods.There was obvious negative correlation between soil moisture and heat wave frequency in July at the arid station,while the correlation coefficient at the semi-humid station was high in May,June and July,and the correlation coefficient at the humid station was high in late May,early June,and July.Therefore,both early-season and July soil moisture contributed to heat waves in July;however,the difference is that soil moisture in the arid region influenced summer heat waves over a longer time scale,and only over relatively shorter time scales at the semi-humid and the humid stations.The spatial correlation results also showed obviously negative correlation. The mechanisms involved were elucidated using CAM5.1-based sensitivity experiments.The decreased soil moisture in March,June and July resulted in an anticyclonic anomaly and positive anomalies in geopotential height and air temperature.Meanwhile,soil moisture decreasing in July not only resulted in positive geopotential height and air temperature anomalies,but also contributed to a high surface skin temperature and sensible heat flux in the arid and semi-arid regions, which promoted high air temperatures and heat waves.However,sensible heat flux was not in accordance with the temperature distribution in the humid region in eastern China.In addition,from Fig.10,we can conclude that the feedback of early-season soil moisture on heat waves is complex,because positive temperature anomalies do not match well with decreased soil moisture.A possible reason is that soil moisture could influence temperature in remote regions through atmospheric circulation,while another reason might be the uncertainty and limited simulation ability of CAM5.1.Of course, there are likely to be many other factors involved,and thus further research in the future in this regard is necessary.

Acknowledgements.This research was jointly supported by the National Natural Science Foundation of China(Grant Nos. 41375155 and 91437107)and the Priority Academic Program Development of Jiangsu Higher Education Institutions(PAPD).We thank the two anonymous reviewers for their helpful comments and suggestions,which greatly improved the manuscript.

REFERENCES

Baldi,M.,G.Dalu,G.Maracchi,M.Pasqui,and F.Cesarone, 2006:Heat waves in the Mediterranean:A local feature or a larger-scale effect?Int.J.Climatol.,26,1477–1487.

Chen,H.S.,and J.Zhou,2013:Impactofinterannual soilmoisture anomaly on simulation of extreme climate events in China. Part II:Sensitivity experiment analysis.Chinese J.Atmos. Sci.,37(1),1–13.(in Chinese)

Chow,K.C.,J.C.L.Chan,X.Shi,Y.M.Liu,and Y.H.Ding, 2008:Time-lagged effects of spring Tibetan Plateau soil moisture on the monsoon over China in early summer.Int. J.of Climatol.,28,55–67.

Dai,A.,T.M.L.Wigley,G.A.Meehl,and W.M.Washington, 2001:Effects of stabilizing atmospheric CO2on global climate in the next two centuries.Geophys.Res.Lett.,28,4511–4514.

Deng,Z.Y.,Q.Zhang,J.F.Xu,M.Chen,S.J.Qin,and S.Y. Zhang,2009:Comparative studies of the harm characteristic of hot-day wind and high temperature heat waves.Advance in Earth Science,24(8),865–873.(in Chinese)

Ebi,K.L.,and G.A.Meehl,2007:The heat is on:Climate change &heat waves in the Midwest.Regional Impacts of Climate Change:Four Case Studies in the United States,K.L.Ebi, G.A.Meehl,D.Bachelet,R.R.Twilley,and D.Boesch,Eds., Pew Center on Global Climate Change.Arlington,14 pp.

Fischer,E.M.,S.I.Senviratne,P.L.Vidale,D.L¨uthi,and C.Sch¨ar, 2007:Soil moisture–atmosphere interactions during the 2003 European Summer Heat Wave.J.Climate,20,5081–5099.

Frich,P.,L.V.Alexander,P.Della-Marta,B.Gleason,M.Haylock,A.K.Tank,and T.Peterson,2002:Observed coherent changes in climatic extremes during the second half of the twentieth century.Climate Research,19,193–212.

Gong,D.Y.,Y.Z.Pan,and J.A.Wang,2004:Changes in extreme daily mean temperatures in summer in eastern China during 1961–2000.Theor.Appl.Climatol.,77,25–37.

Heck,P.,D.L¨uthi,H.Wernli,and C.Sch¨ar,2001:Climate impacts of European-scale anthropogenic vegetation changes:Astudy with a regional climate model.J.Geophys.Res.,106,7817–7835.

H´emon,D.,and E.Jougla,2004:Surmortalit´e li´ee`a la canicule daoˆut 2003.L’Institut National de la Sant´e et de la recherch´e M´edicale(INSERM),Paris,France,76 pp.(in French)

Khodayar,S.,A.Sehlinger,H.Feldmann,and C.Kottmeier,2014: Sensitivity of soil moisture initialization for decadal predictions under different regional climatic conditions in Europe. Int.J.Climatol.,doi:10.1002/joc.4096.

Koster,R.D.,and M.J.Suarez,2001:Soil moisture memory in climate models.J.Hydrometeorol.,2,558–570.

Koster,R.D.,and Coauthors,2004:Regions of strong coupling between soil moisture and precipitation.Science,305,1138–1140.

Kotroni,V.,K.Lagouvardos,and A.Retalis,2011:The heat wave of Jun.2007 in Athens,Greece—Part 2:Modeling study and sensitivity experiments.Atmos.Res.100,1–11.

Meehl,G.A.,and C.Tebaldi,2004:More intense,more frequent, and longer lasting heat waves in the 21st Century.Science,305,994–997.

Meehl,G.A.,F.Zwiers,J.Evans,T.Knutson,L.Mearns,and P.Whetton,2001:Trends in extreme weather and climate events:Issues related to modeling extremes in projections of future climate change.Bull.Amer.Meteor.Soc.,81,427–436.

Neale,R.B.,J.H.Richter,and A.J.Conley,2011:Description of the NCAR community atmosphere Model(CAM5).Technical Report NCAR/TN-486+STR,National Center for Atmospheric Research,Boulder,Colorado,268 pp.

Retalis,A.,D.Paronis,V.Kotroni,and K.Lagouvardos,2010: The heat wave of Jun.2007 in Athens,Greece-Part 1:study of satellite derived land surface temperature.Atmos.Res.,98, 458–467.

Sun,C.H.,W.J.Li,Z.Q.Zhang,J.H.He,and X.Q.Zhang,2005: Impact of Huaihe River basin soil temperature and humidity abnormality in pre-winter and spring time on the anomalous summer rainfall and its application.Acta Meteorologica Sinica,63(1),115–122.(in Chinese)

Sutton,R.T.,and D.L.R.Hodson,2005:Atlantic Ocean forcing of North American and European summer climate.Science,309,115–118.

Tai,J.G.,Y.F.Zhen,L.Peng,S.Q.Gu,and J.Shi,2008:Effect of urban heat island on heat waves in summer of Shanghai. Plateau Meteorology,27(S1),144–149.(in Chinese)

Tamrazian,A.,S.L.Dochy,J.Willis,and W.C.Patzert,2008: Heat waves in southern California:Are they becoming more frequent and longer lasting?Yearbook of the Association of Pacific Coast Geographers,70,59–69.

Wang,A.H.,D.P.Lettenmaier,and J.Sheffield,2011:Soil moisture in China 1950–2006.J.Climate,24,3257–3271.

Wang,L.,2014:The relationship between heat waves and soil moisture in north China.M.S.thesis,Nanjing University of Information Science&Technology,77 pp.(in Chinese)

WMO,2003:The health impacts of 2003 summer heat-waves. Briefing note for the Delegations of the fifty-third session of the WHO Regional Committee for Europe.12 pp.

Zhang,R.H.,and Z.Y.Zuo,2011:Impact of spring soil moisture on surface energy balance and summer monsoon circulation over East Asia and precipitation in East China.J.Climate,24, 3309–3322.

Zhang,J.,D.L.Li,L.Li,and W.T.Deng,2013:Decadal variability of droughts and floods in the Yellow River basin during the last five centuries and relations with the North Atlantic SST. Int.J.Climatol.,33,3217–3228.

Zhang,J.,L.Li,D.L.Li,and W.T.Deng,2014:Summer drought around the mid-latitude river in the northern China associated with the recent Arctic ice loss in summertime.Int.J.Climatol.,doi:10.1002/joc.4177.

Zhang,J.Y.,and L.Y.Wu,2011:Land-atmosphere coupling amplifies hot extremes over China.Chinese Science Bulletin,56, 3328–3332.

Zhang,Q.,J.Zhang,J.Qiao,and S.Wang,2011:Relationship of atmospheric boundary layer depth with thermodynamic processes at the land surface in arid regions of China.Science China Earth Sciences,54(10),1586–1594.

Ziv,B.,H.Saaroni,and P.Alpert,2004:The factors governing the summer regime of the Eastern Mediterranean.Int.J.Climatol.,24,1859–1871.

Zhou,J.,and H.S.Chen,2012:Impact ofinterannual soilmoisture anomaly on simulation of extreme climate events in China. Part I:Model evaluation based on CAM3.1.Chinese J.Atmos.Sci.,36(6),1077–1092.(in Chinese)

:Zhang,J.,Z.Y.Liu,and L.Chen,2015:Reduced soil moisture contributes to more intense and more frequent heat waves in northern China.Adv.Atmos Sci.,32(9),1197–1207,

10.1007/s00376-014-4175-3.

8 August 2014;revised 23 December 2014;accepted 23 December 2014)

∗Corresponding author:ZHANG Jie

Email:gs-zhangjie@163.com