Abstract

Numerous studies have been conducted on the impact of soil moisture on the climate, but few studies have attempted to diagnose the linkage between soil moisture and climate variability using observational data. Here, using both observed and reanalysis data, the spring (April–May) soil moisture is found to have a significant impact on the summer (June–August) monsoon circulation over East Asia and precipitation in east China by changing surface thermal conditions. In particular, the spring soil moisture over a vast region from the lower and middle reaches of the Yangtze River valley to north China (the YRNC region) is significantly correlated to the summer precipitation in east China. When the YRNC region has a wetter soil in spring, northeast China and the lower and middle reaches of the Yangtze River valley would have abnormally higher precipitation in summer, while the region south of the Yangtze River valley would have abnormally lower precipitation. An analysis of the physical processes linking the spring soil moisture to the summer precipitation indicates that the soil moisture anomaly across the YRNC region has a major impact on the surface energy balance. Abnormally wet soil would increase surface evaporation and hence decrease surface air temperature (Ta). The reduced Ta in late spring would narrow the land–sea temperature difference, resulting in the weakened East Asian monsoon in an abnormally strengthened western Pacific subtropical high that is also located farther south than its normal position. This would then enhance precipitation in the Yangtze River valley. Conversely, the abnormally weakened East Asian summer monsoon allows the western Pacific subtropical high to wander to south of the Yangtze River Valley, resulting in an abnormally reduced precipitation in the southern part of the country in east China.

1. Introduction

Numerous previous studies have examined impacts of abnormal soil moisture on climate. Walker and Rowntree (1977) found that precipitation was sensitive to the initial soil moisture of the numerical model in Africa. Douville et al. (2001) and Douville (2002) conducted idealized soil moisture sensitivity experiments. They found that the soil moisture anomaly had a positive (negative) feedback on the precipitation in the rainy season across the African continent (over the Indian subcontinent). Yeh et al. (1984) discussed the model atmospheric response to prescribed initial soil moisture anomalies. They showed that the climate effect depends significantly on the latitudinal zones where the soil moisture is located. In particular, the effect of the soil moisture anomaly in middle latitudes on rainfall can last for 2–3 months. Shukla and Mintz (1982) found, through numerical modeling, that the soil moisture anomaly could affect air temperature and precipitation significantly for up to several months.

The long-term soil moisture anomaly could be explained by the land surface reaction to the stochastic precipitation forcing, allowing rainwater to be stored in soil before being slowly released into the atmosphere (Delworth and Manabe 1988). Meehl (1994) derived two possible evaporation–precipitation feedback mechanisms from the modeling results, showing that the increased soil moisture could intensify surface evaporation, resulting in increased water vapor in the atmosphere on the one hand and a reduced surface air temperature on the other. The reduced surface air temperature, as a result, may lead to reduced precipitation. In the context of the monsoon precipitation in South Asia, the reduced surface air temperature could lead to narrowing of land–sea temperature differences, which in turn would dampen the monsoon and reduce precipitation. All these numerical studies confirmed that soil moisture affected atmospheric precipitation mainly through the “evaporation–precipitation” feedback mechanism. This mechanism, quite effective on inland heavy precipitation recycling, could differ greatly from one region to the other (Serafini 1990). Through numerical modeling, Wu and Dickinson (2004) further showed that the time scales of layered soil moisture memory depend on geography, season, and depth and are related to precipitation, runoff, and evapotranspiration. All these studies demonstrated that soil moisture can produce a large impact on the nonlocal climate in a prolonged persistence of several months, in addition to the impact on local climate in the same period.

East Asia, a major monsoon region in the world, has sophisticated dynamic and thermodynamic processes in its monsoon climate system and is marked with extremely active interactions among the atmosphere, hydrosphere, and biosphere (Zhang and Liu 2005). Land processes play an important role in shaping and building the regional monsoon climate. You et al. (2000) found, by using an atmospheric general circulation model (AGCM), that the wetter surface in spring (April) over Asian midlatitudes could lead to decreases of both geopotential heights and temperatures in the whole troposphere over most midlatitudes of Asia in the following four months. Based also on an AGCM, Wang (1991) showed that the higher soil moisture in early summer (June) in north China could enhance the precipitation and lower the surface air temperature not only over the initial soil moisture anomaly region but also over surrounding regions. Using a regional climate model, Kim and Hong (2007) investigated the role of soil moisture in the relatively wet summer (June–August) of 1998 and the relatively dry summer of 1997 in East Asia. They found that the initial soil moisture could not significantly affect the precipitation over East Asia.

Although many studies have been done on the impact of soil moisture on climate, few have attempted to diagnose the linkage between soil moisture and climate variability by observational data because of soil moisture data shortage and associated reliability problems. Most current studies investigating the effects of soil moisture on climate variability were made mainly through numerical modeling. Furthermore, most of these numerical modeling studies started with an initial soil moisture field specified subjectively, and modeling results were rarely verified by the observed results. Even inconsistent results were obtained when the impact of the soil moisture on the East Asian summer monsoon precipitation was studied (Wang 1991; You et al. 2000; Kim and Hong 2007). As a matter of fact, the linkage between soil moisture and climate variability has to be examined based on the observed data, a ground for all modeling efforts.

Zuo and Zhang (2007), through diagnosing reanalysis data, showed that there exists a relationship between the spring soil moisture and the summer precipitation anomalies in east China. However, how spring soil moisture affects summer precipitation was not discussed there, and the relationship was not evaluated by observations either. In this paper we will investigate the ties between the spring soil moisture and summer monsoon precipitation in east China and circulation over East Asia by both observed and reanalysis data, in an attempt to reveal the physical linkage between soil moisture and East Asian climate variability on the one hand and provide observational basis for numerical modeling on the other. This will also help to understand the physical mechanism behind the East Asian summer monsoon variability and associated hydrological cycles. Section 2 presents the data and methodologies employed in the study, while section 3 discusses the relation of the spring soil moisture with the summer precipitation in the eastern part of China. Section 4 examines the physical processes by which the spring soil moisture affects the summer precipitation in China, through diagnoses of the associated surface energy balance and atmospheric circulations. The paper ends with a summary and a discussion in section 5.

2. Data and methodology

Limited soil moisture observations make a bottleneck, restricting the previous studies on the relation of soil moisture to climate variability. As a result, one has to employ proxy soil moisture data derived from models. Of the proxy data, reanalysis data of soil moisture have been widely used thanks to its global and long time coverages. The 40-yr European Centre for Medium-Range Weather Forecasts Re-Analysis (ERA-40) (Douville et al. 2000) and the National Centers for Environmental Prediction (NCEP) (Mintz and Serafini 1992) reanalysis data are the proxy data most extensively used. Li et al. (2005) made a detailed assessment of both ERA-40 and NCEP soil moisture reanalysis data using observations collected in Chinese meteorological stations. Results showed that the ERA-40 data are closer to observations, either in terms of interannual and seasonal variations or in terms of memory range. Additionally, the spatial distribution of the soil moisture from ERA-40 data was closer to that from in situ observations. The ERA-40 data contain four soil moisture layers—7, 21, 82, and 189 cm in thickness—from the surface downward at a 2.5° × 2.5° spatial resolution from September 1957 to August 2001.

Observed monthly soil moisture data from the China Meteorological Administration (CMA) started from 1981. During the winter season (December–February), most of China, except the southern part, has frozen soil with little surface evaporation, implying a very limited soil moisture variation. Therefore, there are no observed soil moisture data for boreal winters. A consistent observation starts from each April in 10 layers with 10 cm in thickness. In this study, the observed monthly soil moisture data collected from 61 stations during the period of 1981–2002 were employed. As shown in Fig. 1, these stations are located in the region east of 100°E. Zuo and Zhang (2009) made a comparison between the monthly ERA-40 data and observations in boreal spring, showing a fine consistency between the two in temporal variations and spatial distribution. As a result, in the present study the monthly soil moisture data of the ERA-40 was employed, in addition to in situ observations. Wu and Dickinson (2004) pointed out that the persistence of the soil moisture depends on its depth. Because of missing soil moisture observations in deeper layers, only the observed monthly soil moisture data in the top 30 cm and the monthly ERA-40 soil moisture in the top 28 cm in the period of 1982–2001 were applied in this study, which are deeper than the ERA-40 soil moisture in the top layer (7 cm) used in Zuo and Zhang (2007).

Fig. 1.

Distribution of 61 soil moisture stations in China. The rectangle indicates the YRNC region, which covers the region from the lower and middle reaches of the Yangtze River valley to north China.

Fig. 1.

Distribution of 61 soil moisture stations in China. The rectangle indicates the YRNC region, which covers the region from the lower and middle reaches of the Yangtze River valley to north China.

We also used the observed monthly precipitation and 2-m air temperature data from 160 regular meteorological stations of the CMA, whose locations can be found in Zhang et al. (1999). There are 120 observational stations located in the domain of the eastern part of China, shown in Fig. 1. The monthly sea surface temperature (SST) is from the National Oceanic and Atmospheric Administration (NOAA)’s Optimum Interpolation Sea Surface Temperature (OISST) analysis with a spatial resolution of 1° × 1° from 1982 to 2001 (Reynolds et al. 2002). Monthly winds and geopotential heights are from the ERA-40 for the period of 1982–2001 with a spatial resolution of 2.5° × 2.5°. To verify the reliability of the surface energy flux (latent, sensible, and radiative fluxes), we utilized three sets of the monthly surface energy flux data: those from the ERA-40 and NCEP reanalyses and those estimated by the observed routine meteorological data using the method of Xu and Haginoya (2001) and Xu et al. (2005).

A range of statistical analysis techniques is utilized in our present study. The singular value decomposition (SVD) (Prohaska 1976; Wallace et al. 1992) is applied to investigate the relationship between spring soil moisture and summer precipitation in east China. The correlation analysis is used to further confirm the results obtained from the SVD method. To verify the relationship between spring soil moisture and summer precipitation revealed in our study being independent of El Niño–Southern Oscillation (ENSO), the partial correlation analysis is applied to remove the parts linearly related to the ENSO signal. The composite analysis is employed to compare differences of summer precipitation and circulation between high and low spring soil moisture (HSM and LSM, respectively) years.

In our study we also utilized the correlation vector to investigate the correlation between circulation and a certain time series. The correlation vector R is defined as

 
formula

where Ru&t and Rv&t are correlation coefficients of the time series of a variable t with the zonal and meridional winds, respectively. Here i and j indicate the zonal and meridional unit vectors, respectively.

3. Relation of spring soil moisture to summer precipitation

Taking into account that the observed soil moisture data applied were collected across the eastern part of China, as shown in Fig. 1, the investigation of the linkage between spring (April–May) soil moisture and summer (June–August) precipitation is desirably confined to the geographic region to the east of 105°E. The SVD analysis was applied to the spring soil moisture and observed summer precipitation for the period 1982–2001 to obtain the correlation between the two fields. Figure 2a shows the first SVD mode of spring soil moisture of ERA-40 and summer precipitation with a variance contribution of 30.0% and a correlation coefficient of 0.87, respectively. The corresponding results using observed soil moisture and precipitation are shown in Fig. 2b with a variance contribution of 31.9% and a correlation coefficient of 0.91. The soil moisture depicted in Fig. 2a shows that the eastern part of China, except the northeastern portion, basically had a positive anomaly, with a significant anomaly area over a vast region running of approximately 30°–40°N, 105°–120°E from the lower and middle reaches of the Yangtze River valley to north China (the YRNC region shown in Fig. 1). Corresponding to the distribution of the spring soil moisture anomaly, a positive precipitation anomaly appeared in the northeast (approximately north of 42°N, 105°–120°E) and a negative anomaly across north China around 35°–42°N, 105°–120°E. Meanwhile, a positive anomaly in the lower and middle reaches of the Yangtze River valley over the area about 29°–33°N, 110°–120°E and a negative anomaly to the south of the Yangtze River (approximately south of 29°N) are seen.

Fig. 2.

Spatial distributions of the first SVD mode between [left in (a)–(d)] spring soil moisture and [right in (a)–(d)] observed summer precipitation. The contour interval is 0.3. (a),(b) Spring soil moisture data are from ERA-40 and in situ observation, respectively. (c),(d) As in (a),(b) respectively, but the parts in soil moisture and precipitation linearly related to spring Niño-3 SST are removed by the partial correlation analysis. Dark and light shadings indicate positive and negative correlation, respectively, with a significance level >0.1.

Fig. 2.

Spatial distributions of the first SVD mode between [left in (a)–(d)] spring soil moisture and [right in (a)–(d)] observed summer precipitation. The contour interval is 0.3. (a),(b) Spring soil moisture data are from ERA-40 and in situ observation, respectively. (c),(d) As in (a),(b) respectively, but the parts in soil moisture and precipitation linearly related to spring Niño-3 SST are removed by the partial correlation analysis. Dark and light shadings indicate positive and negative correlation, respectively, with a significance level >0.1.

The SVD analysis using the observed soil moisture and precipitation data (Fig. 2b) also demonstrates a positive soil moisture anomaly in the northern part of the Yangtze River to north China, with a negative anomaly in the northeastern part. The corresponding summer precipitation distribution is consistent with that in Fig. 2a. It exhibits enhanced anomalies over northeast China and the Yangtze River valley and reduced anomalies over north China and to the south of the Yangtze River valley. Compared to Fig. 2a, Fig. 2b shows that the positive precipitation anomalies over northeast China become slightly weaker; both the positive anomalies over the Yangtze River valley and negative anomalies to the south become slightly stronger. Nonetheless, the distribution of soil moisture and precipitation using in situ soil moisture observations basically agrees with that using ERA-40 soil moisture data.

Zuo and Zhang (2007) discussed the relationship between summer precipitation and spring soil moisture in China based only on the ERA-40 soil moisture data in the top 7 cm layer. In the present study, we used both the observed soil moisture data in the top 30 cm and the ERA-40 soil moisture data in the top 28 cm. The results here further indicated that, either based on the ERA-40 or on the observed soil moisture data, the distributions of spring soil moisture and summer precipitation have a very high correlation. The first SVD mode can account for a relatively large variance contribution. The soil moisture anomaly in spring over the YRNC region has a significant correlation with summer precipitation in the eastern part of China. When the YRNC region registers an abnormally high (or low) soil moisture in spring, more (or less) summer precipitation appears in northeast China and the middle and lower reaches of the Yangtze River valley and less (or more) in north China and to the south of the Yangtze River valley.

Previous studies have shown that the summer rainfall in east China is influenced by ENSO (Zhang et al. 1999). To eliminate the effect of ENSO on the relationship between spring soil moisture and summer rainfall in east China, we applied the partial correlation analysis to both observed and ERA-40 spring soil moisture and summer precipitation to remove the parts linearly related to the Niño-3 (5°SW–5°N, 90°–150°E) SST in spring. The first SVD modes between summer precipitation and spring soil moisture of ERA-40 and observed soil moisture after this removal are shown in Figs. 2c and 2d, respectively. Clearly, the coupled modes after removing the effect of Niño-3 SST are similar to their original pattern (Figs. 2a and 2b). Wetter (drier) spring soil moisture in YRNC region is associated with more (less) summer rainfall in northeastern China and the middle and lower reaches of the Yangtze River valley and decreased (increased) rainfall in southeastern China and north China. That is to say, the correlation between spring soil moisture and summer rainfall in east China is independent of the spring Niño-3 SST. Based on the significant soil moisture anomaly shown in Fig. 2, we selected the region running from the lower and middle reaches of the Yangtze River to North China, the YRNC region, as shown in Fig. 1. The correlation coefficients of spring Niño-3 SST with averaged ERA-40 and observed spring soil moisture over the YRNC region are 0.26 and 0.15, respectively, which are also not statistically significant.

To further understand the linkage between the spring soil moisture and the summer precipitation, both the observed and ERA-40 soil moisture data were averaged over the YRNC region, and their correlation coefficients with summer precipitation were calculated, which are shown in Fig. 3. The spatial pattern of correlations in Fig. 3a is similar to the patterns of precipitation in Fig. 2 for using either the observed or ERA-40 soil moisture data. The positive correlation appears in the Yangtze River valley and in the northeastern part of the country, with the negative one sitting in the southern part of the country and north China, showing the similar results shared by SVD analysis. We also calculated the partial correlation coefficients between the spring soil moisture in YRNC region and summer rainfall in east China to remove the effects linearly related to spring Niño-3 SST. As expected, the distribution of correlation coefficients (Fig. 3b) for either the observed or ERA-40 soil moisture was quite similar to that using original data, as shown in Fig. 3a.

Fig. 3.

(a) Correlation coefficients between summer precipitation and spring (left) ERA-40 and (right) observed soil moisture averaged over the YRNC region. (b) As in (a), but for partial correlation coefficients after removing the parts linearly related to spring Niño-3 SST. Dark and light shadings indicate positive and negative correlations, respectively, with a significance level >0.1.

Fig. 3.

(a) Correlation coefficients between summer precipitation and spring (left) ERA-40 and (right) observed soil moisture averaged over the YRNC region. (b) As in (a), but for partial correlation coefficients after removing the parts linearly related to spring Niño-3 SST. Dark and light shadings indicate positive and negative correlations, respectively, with a significance level >0.1.

Figure 4 shows the time evolution of standardized spring soil moisture of the ERA-40 over the YRNC region, indicating a significant interannual variation in the YRNC region. The results in Fig. 4 are used to select six high (low) spring soil moisture years with an anomaly higher than 0.5 (lower than −0.5) standard deviation. The years of 1983, 1985, 1989, 1990, 1991, and 1998 are rated as HSM years, with those of 1982, 1984, 1986, 1995, 2000, and 2001 being the LSM years. Figure 5 depicts the difference between the composite of the summer precipitation in the six HSM years and that in the six LSM years. Apparently, when the spring soil in the YRNC region is wetter (dryer), more (less) summer precipitation appears over northeast China and the Yangtze River valley and less (more) over south China and north China. The distribution of precipitation anomalies agrees well with the results of SVD analysis shown in Fig. 2.

Fig. 4.

Time evolution of standardized spring soil moisture anomaly of ERA-40 averaged over the YRNC region.

Fig. 4.

Time evolution of standardized spring soil moisture anomaly of ERA-40 averaged over the YRNC region.

Fig. 5.

Difference (mm) between the composite of the summer precipitation in the six HSM and LSM years.

Fig. 5.

Difference (mm) between the composite of the summer precipitation in the six HSM and LSM years.

4. Physical processes responsible for spring soil moisture effect on summer precipitation

As shown in the above analyses, the spring soil moisture in the YRNC region has a significant negative correlation with the summer precipitation in southern China, in addition to its positive link with the summer precipitation over the Yangtze River valley and northeastern China. In this section we will investigate how the spring soil moisture in the YRNC region affects the summer precipitation in the eastern part of China. The East Asian summer climate is mainly controlled by the East Asian summer monsoon system in the region. The thermal condition of land surface plays an important role in regulating monsoon activities (Webster 1987) and is able to change the distribution of summer precipitation in the East Asian monsoon region through affecting the development of the East Asian monsoon (Xue et al. 2004).

a. Effects on surface air temperature

As a major parameter regulating land processes, the soil moisture explains a combined effect of both precipitation and evaporation in the surface hydrological process. In this context, the soil moisture variation may result in a change in other surface parameters as well as water and energy exchanges between land and atmosphere. To analyze the effect of soil moisture on surface thermal condition, Fig. 6 exhibits the composite differences between HSM and LSM years in the spring surface shortwave radiation flux (Fig. 6a), longwave radiation flux (Fig. 6b), sensible heat flux (Fig. 6c), latent heat flux (Fig. 6d), specific humidity at a height of 2 m (Fig. 6e), and surface air temperature at a height of 2 m in May (Fig. 6f) using the ERA-40 data. It is apparent that the YRNC region would have an enhanced shortwave radiation in a spring featured with abnormally wet soil (Fig. 6a). Meanwhile, the net surface longwave radiation would decrease (Fig. 6b), with more latent heat (Fig. 6d) but less sensible heat (Fig. 6c) in the atmosphere. In the YRNC region, all these flux differences exceeded 10 W m−2, with the largest up to 20 W m−2.

Fig. 6.

Composite difference between the HSM and LSM years for (a) net surface shortwave radiation, (b) net surface longwave radiation, (c) sensible heat flux, (d) latent heat flux; all in W m−2. (e) The 2-m specific humidity (g kg−1) and (f) 2-m air temperature in May (K). The HSM and LSM years are same as those in Fig. 5. Dark and light shadings indicate positive and negative correlations, respectively, with a significance level >0.05.

Fig. 6.

Composite difference between the HSM and LSM years for (a) net surface shortwave radiation, (b) net surface longwave radiation, (c) sensible heat flux, (d) latent heat flux; all in W m−2. (e) The 2-m specific humidity (g kg−1) and (f) 2-m air temperature in May (K). The HSM and LSM years are same as those in Fig. 5. Dark and light shadings indicate positive and negative correlations, respectively, with a significance level >0.05.

Furthermore, all composite differences with statistical significance turned up in the eastern part of the country, sitting within the region 30°–42°N, 105°–120°E. The similarity to the spring soil moisture anomaly in YRNC region shown in Fig. 2 suggests a local energy balance being changed by the soil moisture anomaly. Figure 6e indicates that more soil moisture is accompanied by a positive anomaly of specific humidity, with a maximum 1.2 g kg−1 in YRNC region, which means that the wetter soil corresponded to increased evaporation. Meanwhile, Fig. 6f shows a negative temperature difference in May in the YRNC region and a low temperature center appeared, indicating that the abnormally wet soil is associated with an abnormally low surface temperature. Apparently, the abnormally wet soil builds up more evaporation, leading to the increase of the specific humidity in the near-surface atmosphere and the decrease of surface air temperature. This would eventually result in a reduced net surface longwave radiation flux and a similarly reduced sensible heat flux (Delworth and Manabe 1996). Idso et al. (1975) found that surface albedo is negatively correlated with average soil water content in the top soil layer. Therefore, the wetter soil would reduce the albedo and boost net shortwave radiation absorption.

Figure 7 exhibits time evolutions of the standardized spring soil moisture and surface fluxes averaged over the YRNC region for ERA-40 and observations, respectively. It is apparent that the observed soil moisture is finely consistent with the soil moisture of the ERA-40. The spring soil moisture shares a similar interannual variation with an upward latent heat flux and net surface shortwave radiation absorption (Fig. 7a). It indicates that wetter soil would result in a higher upward latent heat flux and an enhanced shortwave radiation, which is opposite of the dryer soil featured with both reduced upward latent heat flux and shortwave radiation absorption. The soil moisture has an out-of phase interannual variation pattern against sensible heat flux and net surface long wave radiation (Fig. 7b). It is featured with wetter (drier) soil corresponding to decreased (increased) upward sensible heat flux and weakened (stronger) net surface longwave radiation. These results in the interannual time scale agreed well with those of the composite analysis given in Figs. 6a–d.

Fig. 7.

Time evolution of standardized spring soil moisture and surface fluxes, averaged over the YRNC region. Black and gray solid lines represent the ERA-40 and observed soil moisture, respectively. Red and blue broken lines in (a) represent net surface shortwave radiation and latent heat fluxes, respectively; and those in (b) are net surface longwave radiation and sensible heat fluxes, respectively.

Fig. 7.

Time evolution of standardized spring soil moisture and surface fluxes, averaged over the YRNC region. Black and gray solid lines represent the ERA-40 and observed soil moisture, respectively. Red and blue broken lines in (a) represent net surface shortwave radiation and latent heat fluxes, respectively; and those in (b) are net surface longwave radiation and sensible heat fluxes, respectively.

Since the surface fluxes of ERA-40 shown in Figs. 6 and 7 are model derived, we utilize the data from two more sources to verify the reliability of these surface fluxes. One is from the NCEP reanalysis and the other from estimations by the observed routine meteorological data using the method of Xu and Haginoya (2001) and Xu et al. (2005). Figure 8 shows time evolutions of the standardized spring net surface radiation fluxes (shortwave plus longwave) (Fig. 8a), net sensible heat fluxes (Fig. 8b), and net latent heat fluxes (Fig. 8c) averaged over the YRNC region from ERA-40 and NCEP reanalyses and estimated surface energy fluxes. The intercomparison among them basically shows a good consistency. The correlation coefficients for net surface radiation fluxes of ERA-40 with those of NCEP reanalysis and estimations are 0.90 and 0.39, respectively. For sensible heat fluxes, the correlation coefficients are 0.86 and 0.82, respectively, and for latent heat fluxes, 0.84 and 0.73, respectively. All these positive correlations are statistically significant, exceeding the 0.05 confidence level, except the correlation for net surface radiation fluxes between ERA-40 and estimations exceeding the 0.1 confidence level. The similarity of surface energy fluxes from three different data sources verifies to a certain extent the reliability of our results.

Fig. 8.

Time evolutions of standardized spring (a) net surface radiation fluxes, (b) net sensible heat fluxes, and (c) net latent heat fluxes averaged over the YRNC region from ERA-40 (solid lines), NCEP reanalysis (dashed lines), and estimation by the observed routine meteorological data (dotted lines).

Fig. 8.

Time evolutions of standardized spring (a) net surface radiation fluxes, (b) net sensible heat fluxes, and (c) net latent heat fluxes averaged over the YRNC region from ERA-40 (solid lines), NCEP reanalysis (dashed lines), and estimation by the observed routine meteorological data (dotted lines).

Figure 9 presents time evolutions of standardized ERA-40 and observed soil moistures and surface air temperatures at a height of 2 m in May averaged over the YRNC region, respectively. Clearly, variations of ERA-40 and observed air temperatures are quite similar, with the correlation coefficient between them being 0.96, exceeding the 0.01 significance level. The surface air temperatures variations are almost out of phase with those of the spring soil moistures. The abnormally wet soil reduces surface air temperature, while the abnormally dry condition encourages the increase of surface air temperatures. Figure 9 shows that the relationship between soil moisture and surface air temperature in the interannual time scale is in line with the results of the composite analysis given in Fig. 6f, further implying that energy exchanges between land and atmosphere are very sensitive to the underlying soil moisture.

Fig. 9.

Time evolutions of standardized soil moistures from ERA-40 (black solid line) and observation (gray solid line), and ERA-40 (blue broken line) and observed (red broken line) 2-m air temperature in May, averaged over the YRNC region.

Fig. 9.

Time evolutions of standardized soil moistures from ERA-40 (black solid line) and observation (gray solid line), and ERA-40 (blue broken line) and observed (red broken line) 2-m air temperature in May, averaged over the YRNC region.

b. Impacts on precipitation and atmospheric circulations

From the above analysis, we have seen that the wetter (drier) spring soil moisture is correlated with a decreased (increased) surface air temperature in May over the YRNC region. We calculated the partial correlation of the summer precipitation with the ERA-40 surface air temperature in May averaged over the YRNC region, in which the parts linearly related to the spring Niño-3 SST are removed. In Fig. 10 the partial correlation coefficients are shown. It can be seen that the correlation coefficients are higher than 0.4 over southern China, lower than −0.4 in the Yangtze River valley, and −0.6 in northeastern China. The distribution of correlation coefficients generally resembles that in Fig. 3 except for the small differences over north China, indicating that the relation of the summer precipitation with spring soil moisture is also held with the surface air temperature in May. The lower surface air temperature in May over the YRNC region is correlated with more summer precipitation in northeast China and the middle and lower reaches of the Yangtze River valley and less to the south of the Yangtze River valley.

Fig. 10.

Partial correlation coefficients of the summer precipitation with the 2-m air temperature in May averaged over the YRNC region, in which the parts linearly related to spring Niño-3 SST are removed. Dark and light shadings indicate positive and negative correlations, respectively, with a significance level >0.1.

Fig. 10.

Partial correlation coefficients of the summer precipitation with the 2-m air temperature in May averaged over the YRNC region, in which the parts linearly related to spring Niño-3 SST are removed. Dark and light shadings indicate positive and negative correlations, respectively, with a significance level >0.1.

To show circulation features over East Asia associated with the precipitation anomalies, in Fig. 11a we show a composite difference of 850-hPa geopotential heights in summer between the HSM and LSM years. In east China positive (negative) anomalies of geopotential heights are located to the south (north) of the Yangtze River. Such distribution of the geopotential height anomalies is unfavorable for the northward movement of the western Pacific subtropical high (WPSH) and leads to the WPSH strengthening and locating to the south. This could correspond to a weak East Asian summer monsoon during which more precipitation occurs in the middle and lower reaches of the Yangtze River valley and less in south and north China (Ding 1994). From the composite differences between the HSM and LSM years for precipitation shown in Fig. 5 and geopotential heights in Fig. 11a, it is obvious that wetter spring soil moisture in the YRNC region in May corresponds to a weak East Asian summer monsoon.

Fig. 11.

(a) Composite difference of 850-hPa geopotential heights in summer between HSM and LSM years (gpm). Dark and light shadings indicate positive and negative differences, respectively, with a significance level >0.1. (b) Partial correlation vectors of the surface air temperature in May averaged over the YRNC region with 850-hPa winds and partial correlation coefficients with 850-hPa geopotential heights in summer (contours), in which the parts linearly related to the spring Niño-3 SST are removed. The HSM and LSM years are the same as those in Fig. 5. Only the partial correlation vectors exceeding a significant level of 0.1 are drawn. Dark and light shadings indicate positive and negative correlations, respectively, with a significance level >0.1.

Fig. 11.

(a) Composite difference of 850-hPa geopotential heights in summer between HSM and LSM years (gpm). Dark and light shadings indicate positive and negative differences, respectively, with a significance level >0.1. (b) Partial correlation vectors of the surface air temperature in May averaged over the YRNC region with 850-hPa winds and partial correlation coefficients with 850-hPa geopotential heights in summer (contours), in which the parts linearly related to the spring Niño-3 SST are removed. The HSM and LSM years are the same as those in Fig. 5. Only the partial correlation vectors exceeding a significant level of 0.1 are drawn. Dark and light shadings indicate positive and negative correlations, respectively, with a significance level >0.1.

To understand the linkage between the surface air temperature over the YRNC region in May and East Asian summer monsoon circulations, we calculated the correlation vector as defined in Eq. (1) using the data in the period of 1981–2002 so as to have a direct view of the related circulations in summer with the surface air temperature in May. Here R in Eq. (1) represents the correlation vector between 850-hPa winds in summer and the surface air temperature averaged over the YRNC region in May. The Ru&t and Rv&t are correlation coefficients of the surface air temperature with the zonal and meridional winds at 850 hPa, respectively. For the convenience of discussion, correlation vectors will be described as wind directions. For example, the positive and negative correlation vectors between the surface air temperature and meridional wind will be described as southerly and northerly, respectively, and those between the surface air temperature and zonal winds as westerly and easterly, respectively.

Figure 11b shows the partial correlation vectors of the surface air temperature at a height of 2 m averaged over the YRNC region in May with 850-hPa winds and partial correlation coefficients with 850-hPa geopotential heights in summer, in which the parts linearly related to the spring Niño-3 SST are removed. To compare Fig. 11b to Fig. 11a, we multiply the correlation coefficients by −1 for consistency with the correlation vectors showed in Eq. (1). The positive correlation appears to the south of the Yangtze River and negative to its north. It indicates that the lower surface temperature over the YRNC region in May is correlated with increased geopotential heights in summer to the south of the Yangtze River and decreased to its north (Fig. 11b). The similarity between the distributions of the composite difference in Fig. 11a and the correlation coefficients in Fig. 11b illustrates that the summer circulations over East Asia are closely related to the surface air temperature in May, and this relationship is independent of ENSO. It is apparent that abnormally wet spring soil in the eastern part of China narrows the land–sea temperature difference through a reduced surface air temperature and weakens the East Asian summer monsoon. In a same manner, abnormally dry soil lifts surface air temperature, allowing an increased land–sea temperature difference, which is desirable for an abnormally strong summer monsoon.

When the surface air temperature in the YRNC region in May became abnormally low, anomalous northerlies appeared in the northern part of east China in summer, which weaken the East Asian summer monsoon (Fig. 11b). The anomalous southerlies are confined to the southern part of east China so that southerlies and northerlies join near the Yangtze River valley, forming a convergence zone. Therefore, the abnormally wet spring soil promotes an enhanced precipitation across the Yangtze River valley. It is also apparent in Fig. 11b that in association with the anomalous northerlies in the northern part of east China, a cyclonic circulation anomaly dominated over the northeastern part of the country, prompting a positive precipitation anomaly over this region. In addition, in association with the anomalous southerlies in the southern part of east China, an anticyclonic circulation anomaly prevailed over southeastern China, leading to a reduced precipitation there.

Figure 12 depicts the composite difference of a latitude–height cross section of winds in summer along 110°–120°E between HSM and LSM years. It shows that when the abnormally wet soil is in the YRNC region in spring, corresponding to the anticyclonic circulation anomaly over the southeastern part of the country in summer, as shown in Fig. 11b, the downstream appears around the area at about 15°–27.5°N. An abnormally reduced summer precipitation over the southern part of China is formed. Corresponding to the convergence between southerly and northerly anomalies shown in Fig. 11b, an ascending motion is around the Yangtze River valley at about 30°–32.5°N. This ascending motion classifies the region as having an abnormally wet summer. Meanwhile, the downstream at about 35°–40°N reduces the precipitation in north China. The updrafts to the north of 40°N, corresponding to the cyclonic circulation anomaly in the northeastern part of the country shown in Fig. 11b, make more summer precipitation in the northeastern region.

Fig. 12.

Composite difference of latitude–height cross section of winds in summer along 110°–120°E between HSM and LSM years. Units of the meridional wind and p-velocity are m s−1 and 100 × Pa s−1, respectively. The HSM and LSM years are the same as those in Fig. 5. Dark and light shadings indicate positive and negative differences of vertical winds, respectively, with a significance level >0.1.

Fig. 12.

Composite difference of latitude–height cross section of winds in summer along 110°–120°E between HSM and LSM years. Units of the meridional wind and p-velocity are m s−1 and 100 × Pa s−1, respectively. The HSM and LSM years are the same as those in Fig. 5. Dark and light shadings indicate positive and negative differences of vertical winds, respectively, with a significance level >0.1.

c. Relation to summer soil moisture

Our study here suggests that the impact of spring soil moisture on summertime circulations over East Asia and precipitation in east China would not be through the persistence of the soil moisture anomaly from spring to summer but through the changing of surface thermal conditions in May that affect the East Asian summer monsoon. Climatologically, the onset of the East Asian summer monsoon is usually within May; the surface thermal contrast between land and sea is the main cause for the monsoon onset. The later (earlier) onset corresponds to a weakened (enhanced) East Asian summer monsoon (Ding 1994). Therefore, the reduced (increased) land–sea thermal contrast in May caused by the wetter (drier) soil moisture is possibly responsible for a weakened (enhanced) East Asian summer monsoon.

To further illustrate that the summer precipitation anomalies related to spring soil moisture are not caused by the persistence of the soil moisture anomalies, in Fig. 13 we show the correlation coefficients between ERA-40 summer soil moisture and the time series of the summer precipitation obtained from SVD first mode, as shown in Fig. 2a. When comparing Fig. 13 to Fig. 2, the large difference of the distribution of the correlation coefficients in Fig. 13 with the soil moisture anomalies in Fig. 2 reveals that the soil moisture anomalies in spring do not last to summer. The distribution of the correlation coefficients is quite similar to that of the precipitation anomalies shown in Fig. 2, indicating the summer soil moisture anomalies are possibly a direct response to the precipitation anomalies related to the spring soil moisture anomalies.

Fig. 13.

Correlation coefficients between summer soil moisture and the time series of the summer precipitation obtained from SVD first mode as shown in Fig. 2a. Dark and light shadings indicate positive and negative correlations, respectively, with a significance level >0.1.

Fig. 13.

Correlation coefficients between summer soil moisture and the time series of the summer precipitation obtained from SVD first mode as shown in Fig. 2a. Dark and light shadings indicate positive and negative correlations, respectively, with a significance level >0.1.

5. Summary and discussion

In this study we investigated the relation of the spring soil moisture on the surface energy balance and summer monsoon circulation over East Asia as well as the summer precipitation in east China. Based on both observed and ERA-40 soil moisture data, the first SVD mode between spring soil moisture and summer precipitation shows that when the vast region running from the lower and middle reaches of the Yangtze River valley to north China (the YRNC region) had an abnormally wet soil, the northeastern part of the country and the lower and middle reaches of the Yangtze River valley would have an abnormally wet summer. Less precipitation appeared in east China to the south of the Yangtze River valley. Conversely, the northeastern part and the lower and middle reaches of the Yangtze River valley would be abnormally dry and the southern part of the country abnormally wet in the summer when the YRNC region had an abnormally dry soil. Correlation analysis agreed well with the SVD analysis, suggesting a close linkage between the spring soil moisture anomaly and summer precipitation in east China.

The abnormally wet soil increased evaporation and hence increased the specific humidity in the near-surface atmosphere and reduced the surface air temperature. This would result in reduced surface net longwave radiation flux and sensible heat flux. The wetter soil could reduce the albedo and boost net shortwave radiation absorption. Abnormally wet spring soil in east China in the YRNC region would result in a reduced surface air temperature in late spring, narrowing the land–sea temperature difference. This would weaken the East Asian summer monsoon with decreased geopotential heights to the north of the Yangtze River and an abnormally strengthened and western Pacific subtropical high (WPSH) to the south. A cyclonic circulation anomaly would prevail over the northeastern part of the country and an anticyclonic circulation anomaly would dominate over the southeastern region. The anomalous northerlies from the north associated with the cyclonic circulation and the anomalous southerlies from the south associated with the anticyclonic circulation would join near the Yangtze River valley, forming convergence and updrafts. This suggests that an abnormally wet spring soil would result in an abnormally wet Yangtze River valley in summer. The cyclonic circulation anomaly prevailing over the northeastern part encouraged the development of updrafts, resulting in a positive precipitation anomaly over the region. In contrast, the anticyclonic circulation in southeastern China left the southern part of east China with less precipitation.

Based on these results, we proposed the physical processes responsible for the impact of spring soil moisture on the summer precipitation in east China as follows: (i) wetter (drier) soil in the vast region running from the lower and middle reaches of the Yangtze River valley to north China in spring lowers (raises) the surface air temperature in late spring by changing the surface energy balance; (ii) the lowered (raised) surface air temperature leads to a weakened (enhanced) East Asian summer monsoon by altering the land–sea thermal contrast in late spring; and (iii) the circulations over East Asia associated with the weakened (enhanced) East Asian summer monsoon result in more (less) precipitation in the middle and lower reaches of the Yangtze River valley and in northeastern China and less (more) in southeastern China.

Since ENSO has a major effect on the variability of the East Asian climate, our study investigated the effect of ENSO on the relationship between spring soil moisture and summer climate in east China. The partial correlation analysis was used to eliminate the ENSO signal. All partial correlation analyses confirmed the relationship we proposed between spring soil moisture and summer precipitation and circulation. We illustrated that there is little effect of ENSO on the relationship revealed in our study. As for the reason why the relation of spring soil moisture to the summer climate over east China is independent of ENSO, it is possibly because the effect of ENSO follows a different mechanism. As revealed by Zhang et al. (1996) and Wang et al. (2000), the effect of ENSO on the East Asian summer monsoon is through the anomalous anticyclone in the lower troposphere over the northwestern Pacific, while the effect of spring soil moisture is through altering the surface temperature over east China in late spring by which the East Asian summer monsoon is affected.

Although some studies have investigated the effects of spring soil moisture on the summer climate variability over East Asia using numerical modeling (Wang 1991; You et al. 2000; Kim and Hong 2007), the initial soil moisture fields were specified differently. It is obvious that the inconsistent results of these numerical models should be mainly due to the difference of the locations of the initial soil moisture anomaly. Our study suggests that the spring soil moisture anomaly that affects the summer climate over East Asia should be in specific region, such as the YRNC region found in our present study. Therefore, our results provide the observational base for numerical study of the effects of spring soil moisture on the summer climate variability over East Asia. The numerical experiments are needed in our future study to verify our results of dada diagnoses. In addition, we have illustrated that the summer precipitation anomalies related to spring soil moisture are not caused by the persistence of the soil moisture anomalies. The spring soil moisture anomaly changes the surface temperature in late spring by altering the land surface energy balance, which in turn changes the land–sea thermal difference in late spring and eventually affects the East Asian summer monsoon. In our present study, we also demonstrated that the summer precipitation anomalies are associated with the summertime circulation anomalies, indicating the monsoon circulation anomaly is the main factor that determines the precipitation anomaly. Nonetheless, the role played by the summertime soil moisture in the monsoon precipitation is another important topic for future investigation.

Acknowledgments

We thank the anonymous reviewers’ constructive comments, which helped us to improve the manuscript a great deal. This study is supported by the National Basic Research Program of China (Grant 2007CB411505), the National Key Program (Grant 2007BAC29B02), the National Natural Science foundation of China (NSFC) under Grant 40921003, and the National Key Program of China for Developing Basic Sciences under Grant 2004CB418300.

REFERENCES

REFERENCES
Delworth
,
T. L.
, and
S.
Manabe
,
1988
:
The influence of potential evaporation on the variabilities of simulated soil wetness and climate
.
J. Climate
,
1
,
523
547
.
Delworth
,
T. L.
, and
S.
Manabe
,
1996
:
Climate variability and land surface process
.
From Atmospheric Circulation to Global Change, The Institute of Atmospheric Physics, Chinese Academy of Sciences, Ed., China Meteorological Press, 477–502
.
Ding
,
Y.
,
1994
:
Monsoons over China
.
Kluwer Academic Publishers, 419 pp
.
Douville
,
H.
,
2002
:
Influence of soil moisture on the Asian and African monsoons. Part II: Interannual variability
.
J. Climate
,
15
,
701
720
.
Douville
,
H.
,
P.
Viterbo
,
J.-F.
Mahfouf
, and
A. C. M.
Beljaars
,
2000
:
Evaluation of the optimal interpolation and nudging techniques for soil moisture analysis using FIFE data
.
Mon. Wea. Rev.
,
128
,
1733
1756
.
Douville
,
H.
,
F.
Chauvin
, and
H.
Broqua
,
2001
:
Influence of soil moisture on the Asian and African monsoons. Part I: Mean monsoon and daily precipitation
.
J. Climate
,
14
,
2381
2403
.
Idso
,
S. B.
,
R. D.
Jackson
,
R. J.
Reginato
,
B. A.
Kimball
, and
F. S.
Nakayama
,
1975
:
The dependence of bare soil albedo on soil water content
.
J. Appl. Meteor.
,
14
,
109
113
.
Kim
,
J.-E.
, and
S.-Y.
Hong
,
2007
:
Impact of soil moisture anomalies on summer rainfall over East Asia: A regional climate model study
.
J. Climate
,
20
,
5732
5743
.
Li
,
H.
,
A.
Robock
,
S.
Liu
,
K.
Mo
, and
P.
Viterbo
,
2005
:
Evaluation of reanalysis soil moisture simulations using updated Chinese soil moisture observations
.
J. Hydrometeor.
,
6
,
180
193
.
Meehl
,
G. D.
,
1994
:
Influence of the land surface in the Asian summer monsoon: External conditions versus internal feedback
.
J. Climate
,
7
,
1033
1049
.
Mintz
,
Y.
, and
Y. V.
Serafini
,
1992
:
A global monthly climatology of soil moisture and water balance
.
Climate Dyn.
,
8
,
238
240
.
Prohaska
,
J. A.
,
1976
:
A technique for analyzing the linear relationships between two meteorological fields
.
Mon. Wea. Rev.
,
104
,
1345
1353
.
Reynolds
,
R. W.
,
N. A.
Rayner
,
T. M.
Smith
,
D. C.
Stokes
, and
W.
Wang
,
2002
:
An improved in situ and satellite SST analysis
.
J. Climate
,
15
,
1609
1625
.
Serafini
,
Y. V.
,
1990
:
The time scale of land surface hydrology in response to initial soil moisture anomalies: A case study
.
Tellus
,
42A
,
390
400
.
Shukla
,
J.
, and
Y.
Mintz
,
1982
:
The influence of land-surface evaporation precipitation on earth’s climate
.
Science
,
215
,
1498
1501
.
Walker
,
J.
, and
P. R.
Rowntree
,
1977
:
The effect of soil moisture on circulation and rainfall in a tropical model
.
Quart. J. Roy. Meteor. Soc.
,
103
,
29
46
.
Wallace
,
J. M.
,
C.
Smith
, and
C. S.
Bretherton
,
1992
:
Singular value decomposition of wintertime sea surface temperature and 500-mb height anomalies
.
J. Climate
,
5
,
561
576
.
Wang
,
B.
,
R.
Wu
, and
X.
Fu
,
2000
:
Pacific–East Asian teleconnection: How does ENSO affect East Asian climate?
J. Climate
,
13
,
1517
1536
.
Wang
,
W. Q.
,
1991
:
Numerical experiments of the soil temperature and moisture anomalies’ effects on the short term climate (in Chinese)
.
Chin. J. Atmos. Sci.
,
15
,
115
123
.
Webster
,
P. J.
,
1987
:
The elementary monsoons
.
Monsoons, John Wiley, 3–32
.
Wu
,
W.
, and
R. E.
Dickinson
,
2004
:
Time scales of layered soil moisture memory in the context of land–atmosphere interaction
.
J. Climate
,
17
,
2752
2764
.
Xu
,
J.
, and
S.
Haginoya
,
2001
:
An estimation of heat and water balances in the Tibetan Plateau
.
J. Meteor. Soc. Japan
,
79
,
485
504
.
Xu
,
J.
,
S.
Haginoya
,
K.
Masuda
, and
R.
Suziki
,
2005
:
Heat and water balance estimates over the Tibetan Plateau in 1997-1998
.
J. Meteor. Soc. Japan
,
83
,
577
593
.
Xue
,
Y.
,
H.-M. H.
Juang
,
W.-P.
Li
,
S.
Prince
,
R.
DeFries
,
Y.
Jiao
, and
R.
Vasic
,
2004
:
Role of land surface processes in monsoon development: East Asia and West Africa
.
J. Geophys. Res.
,
109
,
1029
1053
.
Yeh
,
T. C.
,
R. T.
Weterald
, and
S.
Manabe
,
1984
:
The effect of soil moisture on the short-term climate and hydrology change—A numerical experiment
.
Mon. Wea. Rev.
,
112
,
474
490
.
You
,
X. T.
,
T. N.
Xiong
,
T.
Yasunari
, and
H. L.
Tanaka
,
2000
:
The impact of the ground wetness anomalies in spring on the climate of following months (in Chinese)
.
Chin. J. Atmos. Sci.
,
24
,
660
668
.
Zhang
,
R.
, and
J.
Liu
,
2005
:
Effect of the ocean and land process on the climate change in China (in Chinese)
.
Evolution of Climate and Environment in China, D. Qin, Y. Chen, and X. Li, Eds., Science Press, 319–357
.
Zhang
,
R.
,
A.
Sumi
, and
M.
Kimoto
,
1996
:
Impact of El Niño on the East Asian monsoon: A diagnostic study of the ’86/87 and ’91/92 events
.
J. Meteor. Soc. Japan
,
74
,
49
62
.
Zhang
,
R.
,
A.
Sumi
, and
M.
Kimoto
,
1999
:
A diagnostic study of the impact of El Niño on the precipitation in China
.
Adv. Atmos. Sci.
,
16
,
229
241
.
Zuo
,
Z. Y.
, and
R.
Zhang
,
2007
:
The spring soil moisture and the summer rainfall in eastern China
.
Chin. Sci. Bull.
,
52
,
3310
3312
.
Zuo
,
Z. Y.
, and
R.
Zhang
,
2009
:
Temporal and spatial features of the soil moisture in boreal spring in eastern China
.
Sci. China
,
52D
,
269
278
.