1. Introduction
North China is one of the most densely populated regions in the world and a major agricultural production zone (Liu et al. 2015). This region is facing a heightened risk of water shortages due to reduced runoff, groundwater overexploitation, and increased water demand (Xia et al. 2007; Yu et al. 2018). Vulnerability to fluctuations in summer precipitation, which accounts for 50%–78% of the total annual precipitation, has fueled concern over an apparent rise in the frequency of intense summer drought events in recent years (Zhang et al. 2018). For example, consecutive severe summer droughts in 2014 and 2015 had devastating impacts on agricultural production, ecological security, and socioeconomic development (Wang and He 2015; Wang et al. 2017). Drought frequency and severity in North China are projected to increase even further under global warming (Dai 2011). Improved understanding of the physical mechanisms behind summer drought in North China is thus of both scientific and societal importance, and is an essential step toward improved drought prediction and mitigation.
The interannual variability of summer precipitation in North China is dominated by variations in the East Asian summer monsoon (EASM; Ding and Chan 2005; Huang et al. 2011), with summer drought empirically linked to a weak EASM and southward retreat of the western Pacific subtropical high (WPSH). This situation is characterized by an anomalous low-level cyclone over the midlatitude North Pacific, which favors weaker southwesterly winds and reduced water vapor transport to North China (Stephan et al. 2018; Wang and He 2015). These changes displace the subtropical front and monsoon rainband southward, resulting in the canonical “north drought–south flood” pattern (Yu et al. 2004). Anomalous strong subsidence due to upper-tropospheric cooling above North China also acts to suppress convection and promote the development of summer drought (Zhang et al. 2018, L. Zhang et al. 2019). A number of studies have argued that these drought-prone circulation anomalies are linked to atmospheric teleconnections, including the circumglobal teleconnection pattern (CGT; also known as the Silk Road pattern), the Eurasian (EU) teleconnection pattern, and the Pacific–Japan (PJ; also known as East Asia–Pacific or EAP) teleconnection pattern (Wallace and Gutzler 1981; Nitta 1987; Huang 1992; Lu et al. 2002; Ding and Wang 2005; Ding et al. 2011). The forcings underlying these patterns have been extensively explored. Transitions of El Niño–Southern Oscillation (ENSO) modulate the CGT along the subtropical jet via relationships with anomalous diabatic heating in the Indian monsoon region (X. Li et al. 2018; Zhang et al. 2018). Anomalous sea ice cover in the Arctic ocean and snow cover over the Eurasian continent are significant forcing sources for the summertime EU pattern (H. Li et al. 2018; Wang and He 2015; Wang et al. 2017; J. Zhang et al. 2019), while anomalous convective activity over the western Pacific warm pool is an important contributor to the PJ/EAP pattern (Huang 1992; Kosaka and Nakamura 2010; Wang and He 2015).
Despite extensive exploration of the dynamical pathways linking remote forcing to large-scale circulation anomalies associated with summer drought in North China, the detailed physical mechanisms that link these anomalies to drought development remain unclear. Understanding the mechanisms behind precipitation anomalies requires detailed diagnosis of the moisture budget and its relationships to changes in the energy budget and large-scale circulation (Seager and Henderson 2013). Moisture budget analysis, which relates regional precipitation and evaporation to moisture flux convergence, has been demonstrated to be useful in determining the causes of extreme hydroclimatic events, as precipitation anomalies are directly affected by changes in moisture sources and transport (Seager et al. 2010). Further decomposition of the contributors to moisture flux convergence then provides additional insight into the roles of changes in surface temperature and the atmospheric circulation (Akinsanola and Zhou 2019; Emori and Brown 2005; Peng and Zhou 2017). Energy budget analysis has also been used to identify and evaluate constraints on precipitation changes and can provide a complementary context to moisture budget analysis (O’Gorman et al. 2012). The regional energy budget involves a balance between diabatic heating and dry static energy (DSE) flux convergence, with precipitation treated as an energy flux rather than a water flux (Muller and O’Gorman 2011). Both moisture and energetic approaches have been widely used to address global or regional precipitation changes in response to climate change, but have rarely been applied to analyze drought mechanisms. In this study, we diagnose the local moisture and energy budgets associated with summer drought events in North China to provide new insight into the detailed physical mechanisms behind the development of summer drought in this region.
The paper is organized as follows. In section 2, we introduce the data and methods used to construct the moisture and energy budget analyses. In section 3, we examine summer precipitation and evaporation anomalies in North China for five recent severe summer droughts. In section 4, we describe the thermodynamic and dynamic processes behind summer drought in North China from the perspective of the regional moisture and energy budgets. We further explore how these budgets are linked to anomalous large-scale circulation patterns in section 5. We then discuss the results in the context of previous work in section 6 and summarize our conclusions in section 7.
2. Data and methods
a. Data
Diagnostic analyses in this study are based mainly on the fifth major global reanalysis produced by European Centre for Medium-Range Weather Forecasts (ERA5; Hersbach et al. 2020) over 1979–2017. Drought composites are constructed using precipitation and evaporation products at a horizontal resolution of 0.25° × 0.25°. To calculate budget components, we use 3-hourly fields of specific humidity, temperature, zonal and meridional winds, vertical pressure velocity, geopotential height, and diabatic heating rates on 23 pressure levels from 1000 to 200 hPa at 1° × 1° resolution. For context, we use observed monthly precipitation data from the Global Precipitation Climatology Project (GPCP; Adler et al. 2003) at a horizontal resolution of 2.5° × 2.5° during 1979–2017. We also use monthly estimates of evaporation from the Global Land Evaporation Amsterdam Model (GLEAM, version 3.3; Miralles et al. 2011; Martens et al. 2017) at 0.25° × 0.25° resolution during 1980–2017. July–August (JA) mean precipitation anomalies are used to identify summer drought, as JA precipitation dominates the total summer (June–August) precipitation in North China (see section 3). All anomalies in ERA5 are calculated relative to the 1979–2017 climatology. Student’s t test is used to calculate confidence levels against the null hypothesis that anomalies averaged over all nondrought years are zero. For comparison purposes, results based on the ECMWF interim reanalysis (ERA-Interim; Dee et al. 2011), the Japanese 55-year Reanalysis (JRA-55; Kobayashi et al. 2015; Harada et al. 2016), the Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2; Gelaro et al. 2017), and several additional observation-based precipitation analyses are included in online supplemental information.
b. Moisture budget equation
The vertically integrated moisture budget equation is given by
where the prime indicates a departure from the climatology (Trenberth and Guillemot 1995; Chou et al. 2009; Seager et al. 2010; Wills and Schneider 2015; Peng and Zhou 2017; Akinsanola and Zhou 2019). Here, P is precipitation, E is evaporation, q is specific humidity, V = (u, υ) is the horizontal wind vector, ω is the vertical pressure velocity, and ∇ = ∂x(⋅)i + ∂y(⋅)j denotes the horizontal gradient operator on an isobaric surface. The subscripts x, y, and p indicate partial derivatives in the zonal, meridional, and vertical directions, respectively. Vertical integrals are calculated via density-weighted integration through the atmospheric column:
where g = 9.81 m s−2 is the acceleration due to gravity, ps is the surface pressure, and pt = 200 hPa. The term δ in Eq. (1) represents the budget residual, which includes the effects of surface processes and model biases. The first two terms on the right-hand side of Eq. (1) are the horizontal moisture advection (−⟨V ⋅ ∇q⟩) and vertical moisture advection (−⟨ω∂pq⟩). These two terms combined represent the moisture flux convergence (−⟨∇ ⋅ Vq⟩) as, according to mass continuity, vertical moisture advection approximates the horizontal convergence of moisture (−⟨q∇ ⋅ V⟩).
Reynolds decomposition is used to further separate specific humidity and winds into their climatological means and corresponding departures, namely
We refer to the three terms on the right-hand sides of Eqs. (3)–(5) as the thermodynamic, dynamic, and nonlinear terms, respectively. The thermodynamic terms involve changes in specific humidity assuming a fixed circulation, while the dynamic terms are related to changes in winds assuming fixed specific humidity. The nonlinear terms associated with fluctuations in both specific humidity and circulation are often negligibly small (Chou et al. 2009; Peng and Zhou 2017). Our results are consistent with this expectation, so we neglect variations in the nonlinear terms below.
c. Energy budget equation
Following Muller and O’Gorman (2011), we use a regional DSE budget to evaluate energetic constraints on summer drought in North China. The vertically integrated energy budget is
where Lc = 2.501 × 106 J kg−1 is the latent heat of condensation, s = cpT + gz denotes the DSE (with cp = 1005 J kg−1 K−1 the specific heat of air at constant pressure), T is temperature, and z is geopotential height. The term Q represents diabatic heating (excluding latent heating), which can be broken down into three parts:
where LW and SW are longwave and shortwave radiative heating (positive for absorption by the atmosphere), and SH is the upward sensible heat flux at the surface. The term ε′ in Eq. (6) represents the budget residual, which reflects imbalances between model-generated diabatic heating and analysis-derived DSE flux divergence. A decomposition similar to that used for the moisture budget is made by replacing specific humidity with DSE in the form of Eqs. (3)–(5). Quantities in units of energy flux (W m−2) are divided by Lc and multiplied by 86 400 s day−1 to convert them to units of equivalent precipitation flux (mm day−1).
d. Wave activity flux
The wave activity flux (WAF) as proposed by Takaya and Nakamura (2001) is calculated to describe Rossby wave propagation associated with atmospheric teleconnection patterns. The two-dimensional WAF is calculated as
where ψ′ represents the perturbation quasigeostrophic streamfunction and the subscripts denote the dimensions in which partial derivatives are taken.
3. Characteristic precipitation and evaporation anomalies
We first evaluate the performance of summer precipitation in ERA5 relative to the observationally based GPCP precipitation analysis. Time series of JA mean precipitation anomalies averaged over North China (32°–42°N, 105°–120°E) for 1979–2017 from ERA5 and GPCP are shown in Fig. 1. The climatological JA mean precipitation averaged over North China is 4.78 mm day−1 in ERA5, larger than that based on GPCP (3.87 mm day−1). The standard deviation of the precipitation anomaly in ERA5 (0.87 mm day−1) likewise exceeds that indicated by GPCP (0.66 mm day−1), although standard deviations in both datasets represent 17%–18% of the mean value. The linear correlation coefficient between ERA5 and GPCP precipitation anomalies during JA is 0.93, which is statistically significant at the 99.9% confidence level. Anomalies based on ERA5 and GPCP are very consistent with those based on other reanalysis and observation-based precipitation products, especially for the five drought years (Fig. S1 and Table S1 in the online supplemental material). Observed negative precipitation anomalies in GPCP are in most cases well captured by ERA5. We define summer drought events as years with negative precipitation anomalies exceeding one standard deviation of the time series in both ERA5 and GPCP. This criterion selects five severe summer drought years over the period 1979–2017: 1997 (−1.17 mm day−1), 1999 (−1.42 mm day−1), 2002 (−1.73 mm day−1), 2014 (−1.73 mm day−1), and 2015 (−2.04 mm day−1). These five years are marked by red stars in Figs. 1a and 1b. The four most recent drought years (1999, 2002, 2014, and 2015) have been investigated individually by previous studies (Wei et al. 2004; Wu and Zhang 2013; Wang and He 2015; Xu et al. 2017; Wang et al. 2017).
(left) Time series of July–August (JA) precipitation anomalies (units: mm day−1) averaged over North China (32°–42°N, 105°–120°E) from (a) ERA5 and (b) GPCP. The red stars indicate the five selected summer drought years. (right) Spatial patterns of composite-mean JA precipitation anomalies for these five drought years are shown for (c) ERA5 and (d) GPCP, with the North China domain bounded by the red box. Stippling indicates anomalies that are statistically significant at the 95% confidence level based on Student’s t test.
Citation: Journal of Climate 33, 23; 10.1175/JCLI-D-20-0176.1
All five of the selected summer drought events occurred after the mid-1990s, suggesting an increase in drought frequency relative to the early part of the analysis period. However, the ERA5 time series contains a statistically significant negative trend in JA precipitation over North China during this period that does not appear in GPCP or other observationally based estimates. The most important implication of this difference is that 1986 and 1991 also qualify as drought years when this decreasing trend is removed from the ERA5 time series. Full attribution of biases in ERA5 precipitation over North China is beyond the scope of this paper, but we discuss some limitations of this dataset as well as the impacts of including 1986 and 1991 in the composite analysis in section 6.
The spatial patterns of composite-mean JA precipitation anomalies for the five drought years (Figs. 1c,d) show negative precipitation anomalies stretching across North China. These precipitation deficits in North China are accompanied by excess precipitation in South China, evoking the well-known “north drought–south flood” pattern (Yu et al. 2004; Wang and He 2015). The large-scale spatial pattern of precipitation anomalies based on ERA5 (Fig. 1c) matches well with that based on GPCP (Fig. 1d), although the anomalies have larger magnitudes in ERA5.
Figure 2 shows annual cycles of precipitation and evaporation averaged over North China based on ERA5, GPCP, and GLEAM. Area-mean annual cycles of precipitation and evaporation based on ERA5 (Figs. 2a–c) are qualitatively consistent with the observationally based estimates (Figs. 2d–f), despite ERA5 producing larger fluxes of both precipitation and evaporation. The annual precipitation amount is dominated by rainfall during JA (Figs. 2b,e), which coincides with the summer monsoon in North China (Ding and Chan 2005). Values of precipitation minus evaporation (P − E) during JA of the five drought years (Figs. 2a,d) were 1–2 mm day−1 less than climatological values, approaching zero and even veering negative in some years. Reductions of P − E during drought years were dominated by anomalies in precipitation (Figs. 2b,e), as both ERA5 and GLEAM indicate that area-mean evaporation in drought years was statistically indistinguishable from the climatology (Figs. 2c,f). The spatial pattern of composite P − E anomalies averaged over drought years (Fig. S2a) was thus very similar to the spatial pattern of composite precipitation anomalies (Fig. 1).
Annual cycle of area-averaged (a),(d) precipitation minus evaporation (P − E), (b),(e) precipitation, and (c),(f) evaporation over North China, all in units of mm day−1 based on (a)–(c) ERA5 and (d)–(f) the combination of GPCP precipitation and GLEAM evaporation. Blue solid lines indicate climatological means during 1979–2017 (1980–2017 for GLEAM), yellow solid lines indicate composite means for the five drought years, and magenta solid lines indicate composite anomalies for the five drought years. Blue shading indicates the 2σ uncertainty on the climatological mean annual cycle, with σ being the standard error of the mean. Yellow dashed lines show monthly mean values during the five individual drought years.
Citation: Journal of Climate 33, 23; 10.1175/JCLI-D-20-0176.1
Despite reasonable agreement in the mean annual cycle, ERA5 does not reliably reproduce the magnitude or spatial pattern of evaporation anomalies during drought years. Figure 3 shows time series of JA mean evaporation anomalies averaged over North China from ERA5 and GLEAM. The climatological mean evaporation during JA averaged over North China is 3.05 mm day−1 in ERA5, about 25% higher than that based on GLEAM (2.39 mm day−1), while the standard deviation in ERA5 (0.15 mm day−1) is about 20% smaller than that in GLEAM (0.19 mm day−1). The linear correlation coefficient between evaporation based on ERA5 and that based on GLEAM is 0.71 (99.9% confidence), but the two datasets show poor consistency between 1997 and 2002, when three of the five selected drought years occurred. Spatial patterns of composite-mean JA evaporation anomalies during the five drought years (Figs. 3c,d) show negative evaporation anomalies in South China in both datasets, suggesting an opposing relationship between summertime precipitation and evaporation anomalies in South China (cf. Figs. 1c,d). By contrast, the spatial pattern of evaporation anomalies in North China during drought years differs substantially between ERA5 and GLEAM: whereas GLEAM shows strong negative anomalies, anomalies in ERA5 are mostly weak or neutral. A positive anomaly located east of the Sichuan basin that appears in both datasets (but does not reach the significance threshold in GLEAM) extends farther northeastward in ERA5, canceling out negative anomalies in other parts of the domain. We briefly discuss possible origins of discrepancies between ERA5 and GLEAM in section 6.
As in Fig. 1, but for JA evaporation anomalies (units: mm day−1) averaged over North China from (a),(c) ERA5 and (b),(d) GLEAM.
Citation: Journal of Climate 33, 23; 10.1175/JCLI-D-20-0176.1
4. Thermodynamic and dynamic processes behind summer drought
a. Moisture budget
We conduct a moisture budget analysis to quantify how different physical processes contributed to precipitation deficits during recent summer drought events in North China. The small anomalies in evaporation during drought years indicate that negative precipitation anomalies must be roughly balanced by decreases in vertically integrated moisture flux convergence. The residual term is much smaller than P − E or vertically integrated moisture flux convergence (Fig. S2). We therefore proceed by decomposing changes in moisture flux convergence into thermodynamic and dynamic components according to Eqs. (3)–(5).
Figure 4 shows composite-mean vertical distributions of moisture advection terms averaged over North China from May to October during the five drought years. The dynamic terms (Figs. 4d–f) were the main contributors to negative precipitation anomalies during JA. The thermodynamic component of meridional moisture advection in the lower troposphere (
Evolution of composite vertical profiles of monthly anomalies in the (a) zonal, (b) meridional, and (c) vertical thermodynamic terms of the moisture budget (units: 10−4 kg kg−1 day−1) averaged over North China during the five drought years between 1000 and 200 hPa. (d)–(f) As in (a)–(c), but for the corresponding dynamic terms. The × symbols indicate anomalies that are statistically significant at the 95% confidence level based on Student’s t test.
Citation: Journal of Climate 33, 23; 10.1175/JCLI-D-20-0176.1
The contributions of negative anomalies in the dynamic components to reduced JA precipitation can be explained by circulation anomalies during drought years (Figs. 5a–c). Moisture supply to North China during summer is regulated by climatological southwesterly winds (Ding and Chan 2005). Zonal and meridional wind anomalies (Figs. 5a,b) show that prevailing southwesterlies during July and August were anomalously weak during the five drought years. Weaker southwesterlies correspond to reduced moisture transport from the Indian and Pacific Oceans, resulting in a drier atmosphere above North China (Fig. 5d). In addition to weaker southwesterlies, anomalous subsidence through most of the troposphere (Fig. 5c) suppressed vertical moisture advection, which in turn implies reduced horizontal moisture convergence.
Evolution of composite vertical profiles of monthly anomalies in area-mean (a) zonal wind (m s−1), (b) meridional wind (m s−1), (c) vertical pressure velocity (10−2 Pa s−1), (d) specific humidity (g kg−1), (e) temperature (K), and (f) geopotential height (gpm) based on ERA5 during the five selected summer drought years in North China. The × symbols indicate anomalies that are statistically significant at the 95% confidence level based on Student’s t test.
Citation: Journal of Climate 33, 23; 10.1175/JCLI-D-20-0176.1
Figure 6 shows time series of vertically integrated moisture advection anomalies during JA averaged over North China. Variations in the dynamic terms were typically much larger than those in the thermodynamic terms, affirming that these terms play the leading role in driving interannual variability of summer precipitation in this region. Correlation coefficients between P − E anomalies and the zonal and meridional dynamic components were 0.43 and 0.48, respectively, while that with vertical moisture advection was 0.94. All three correlations are significant at the 99% confidence level. The strong statistical link between P − E and vertical moisture advection illustrates the essential connection between precipitation anomalies and vertical motion.
(left) Time series and (right) cluster plots of the zonal, meridional, and vertical (a) thermodynamic and (b) dynamic terms in the moisture budget averaged over North China during JA in units of mm day−1. Bars in both panels show positive (green) and negative (brown) JA P − E anomalies from ERA5 (see also Figs. 1 and 3). Drought years are marked by red dots in the cluster plots.
Citation: Journal of Climate 33, 23; 10.1175/JCLI-D-20-0176.1
Cluster plots of different moisture advection terms during individual years are also shown in Fig. 6. The zonal and meridional thermodynamic terms varied substantially (up to ±1 mm day−1) from year to year, but with no clear distinctions between drought years and normal years. The vertical thermodynamic term exhibited very little interannual variability over the analysis period. By contrast, large spreads are evident in both the meridional and vertical dynamic terms, collectively accounting for much of the interannual variability in summer precipitation. The spread in the zonal dynamic term is much smaller than that in the meridional or vertical dynamic terms, but drought years cluster near the bottoms of the distributions for all three dynamic terms. The only exception is the meridional term in 1997. We discuss this exception further in section 6.
Our moisture budget analysis thus indicates that summer drought in North China was typically associated with reduced moisture transport due to circulation changes. This result indicates that summer drought in North China was caused dynamically by anomalous circulation patterns, rather than thermodynamically by temperature-induced changes in specific humidity. Dynamic contributions to summer drought are dominated by suppressed vertical moisture advection due to anomalous subsidence, followed by reduced meridional and zonal moisture advection associated with anomalous northeasterly flow. Results based on other recent reanalysis products are consistent with those based on ERA5 (Figs. S3 and S4).
b. Energy budget
We further investigate the physical processes behind drought in North China by analyzing the regional DSE budget [Eq. (6)]. Precipitation deficits imply reduced latent heating in the atmosphere (LcP′ ≈ ⟨LH′⟩), which must be balanced at the regional scale by some combination of reduced radiative cooling, increased upward sensible heat flux, and reduced DSE flux divergence (Muller and O’Gorman 2011). Analysis of the DSE budget thus provides an alternative perspective on the relative contributions of local and nonlocal factors to the occurrence and persistence of drought.
Diabatic heating anomalies averaged over North China during drought years [i.e., the left-hand side of Eq. (6)] are shown in Fig. 7. Total diabatic heating (LW + SW + SH + LH; Fig. 7a) anomalies were dominated by the nonradiative component (SH + LH; Fig. 7c), which includes reduced latent heating in the middle-to-upper troposphere and increased sensible heating near the surface. The latter effect is also evident in positive surface sensible heat flux anomalies during drought years (Fig. 7d). Radiative heating also shows negative anomalies in the middle-to-upper troposphere and positive anomalies in the lower troposphere, especially during July (Fig. 7b). The former were associated with enhanced LW cooling and reduced SW heating and the latter with enhanced SW heating and reduced LW cooling, primarily due to changes in the vertical distribution of clouds (Fig. S5). However, anomalies in radiative heating were nearly ten times smaller than those in nonradiative heating. Reduced latent heating during drought years must therefore have been balanced mainly by reductions in vertically integrated DSE flux divergence.
Evolution of composite vertical profiles of monthly anomalies in area-mean (a) total diabatic heating, (b) radiative heating, and (c) nonradiative heating during the five selected summer drought years over North China. Heating rates are from ERA5 in units of K day−1. (d) Composite-mean spatial distribution of JA surface sensible heat flux anomalies during the five drought years in units of W m−2. The × symbols indicate anomalies that are statistically significant at the 95% confidence level based on Student’s t test.
Citation: Journal of Climate 33, 23; 10.1175/JCLI-D-20-0176.1
Figure 8 shows profiles of area-mean anomalies in the thermodynamic and dynamic components of DSE advection over North China during JA of drought years. In the context of Eq. (6), negative values in Fig. 8 relate to reduced export of DSE. This can be understood as an enhancement of downgradient transport, resulting in anomalous DSE advection from boundaries where it is relatively large (south, east, upper troposphere) into the region toward locations where it is relatively small. The thermodynamic contribution arises mainly from reduced zonal export of dry enthalpy (cpT) and geopotential (gz), as negative anomalies at upper levels and positive anomalies at lower levels largely offset each other in the meridional dry enthalpy term. However, as with the moisture budget, the decrease in DSE flux divergence was dominated by the dynamic contribution, as the largest anomalies in the thermodynamic terms (Figs. 8a–c) were much smaller than those in the dynamic terms (Figs. 8d–f). Anomalous vertical advection of DSE in the middle-to-upper troposphere played the dominant role, and was rooted mainly in anomalous downward advection of geopotential (i.e., decreased upper-level divergence of geopotential; red lines in Figs. 8d and 8f). Positive anomalies in vertical and meridional enthalpy advection (red and purple lines in Fig. 8e) partially offset negative anomalies in vertical geopotential advection. Dynamical contributions to the decrease in DSE flux divergence during drought years were thus dominated by decreased geopotential flux divergence in the middle-to-upper troposphere.
Profiles of zonal (blue), meridional (purple), and vertical (red) components of (a),(d) dry static energy (s), (b),(e) enthalpy (cpT), and (c),(f) potential energy (gz) export averaged over North China for the five selected summer drought years, for the (a)–(c) thermodynamic and (d)–(f) dynamic terms. The × symbols indicate anomalies that are statistically significant at the 95% confidence level based on Student’s t test.
Citation: Journal of Climate 33, 23; 10.1175/JCLI-D-20-0176.1
Figure 9 shows time series and cluster plots for the vertically integrated DSE budget components during JA. Vertically integrated total diabatic heating anomalies (blue and red bars in Figs. 9a and 9b) were significantly correlated with JA precipitation anomalies (r = 0.98; Fig. 1a), as expected since latent heating during precipitation formation is the main contributor to diabatic heating in the troposphere above North China. The zonal thermodynamic term exhibited strong interannual variations but with no clear distinction between drought years and normal years (Fig. 9a), while the meridional and vertical thermodynamic terms showed little interannual variability. Large variations in the meridional and vertical dynamic terms accounted for much of the interannual variability in total diabatic heating (Fig. 9b), with correlations of −0.49 and 0.92, respectively. Similar to the moisture budget, the drought years cluster near the extremes of the meridional and vertical dynamic distributions, with the exception of the meridional term in 1997. These results indicate that severe summer drought was typically associated with sharp decreases in DSE flux divergence by anomalous vertical motion, partially offset by meridional circulation anomalies. In 1997, by contrast, meridional circulation anomalies amplified the effects of anomalous vertical motion in reducing DSE flux divergence.
(left) Time series and (right) cluster plots of the zonal, meridional, and vertical (a) thermodynamic and (b) dynamic terms in the energy budget averaged over North China during JA in equivalent units of mm day−1. Bars in both figures show positive (red) and negative (blue) JA total diabatic heating anomalies from ERA5. Drought years are marked by red dots in the cluster plots.
Citation: Journal of Climate 33, 23; 10.1175/JCLI-D-20-0176.1
Our analysis of the energy budget thus links summer drought in North China to reduced latent heating in the middle-to-upper troposphere. This reduced latent heating was balanced dynamically by decreased DSE flux divergence associated with anomalous subsidence.
5. Physical interpretation
a. Upper-tropospheric cooling and surface heating
The moisture and energy budget analyses described in the previous section reveal that summer drought in North China during 1979–2017 was dominated by dynamic contributions. Here, we investigate the circulation anomalies associated with summer drought in North China and their relationships with the moisture and energy budgets. Figure 10 shows composite geopotential height anomalies on the 200-, 500-, and 850-hPa isobaric surfaces. Drought years were characterized by negative geopotential height and cyclonic circulation anomalies at 200 hPa over North China (Fig. 10a). The negative geopotential height anomaly implies upper-tropospheric cooling above North China, consistent with increasingly negative temperature anomalies in the upper troposphere during JA (Fig. 5e). The maximum negative temperature anomaly in the drought composite (approximately −1 K) was located in the upper troposphere, though negative anomalies extended downward to ~700 hPa. Upper-level cooling dynamically favored subsidence, as reflected in anomalous vertical motion over North China (Fig. 5c).
Composite-mean spatial distributions of geopotential height (shading; gpm) and wave activity flux (vectors; scale at upper right) anomalies at (a) 200 and (b) 500 hPa, and (c) wind anomalies (streamlines; m s−1) at 850 hPa during JA for the five selected summer drought years. (d) Zonal wind anomaly (shading) and climatology (contours) at 200 hPa in units of m s−1. Stippling indicates that anomalies are statistically significant at the 95% confidence level.
Citation: Journal of Climate 33, 23; 10.1175/JCLI-D-20-0176.1
The upper-level cyclone and associated subsidence can explain both the suppressed vertical moisture advection revealed by the moisture budget analysis and the reduced geopotential divergence revealed by the energy budget analysis. Anomalous subsidence also promoted the development of a weak but statistically significant anticyclonic anomaly below the upper-tropospheric cooling (Figs. 10a–c). This low-level anticyclone and the upper-level cyclone above it together comprise a tropospheric circulation anomaly in the first baroclinic mode over North China, as reflected in the vertical profile of geopotential height anomalies during JA (Fig. 5f). Local land surface feedbacks further reinforced the dynamically induced low-level anticyclone. Clear-sky conditions associated with anomalous subsidence increased the net solar radiation flux at the surface (mean over drought years: +4.3 W m−2), while precipitation deficits reduced the amount of moisture available for evaporation (Fig. 3). Both anomalies favored a more intense sensible heat flux from the surface to the atmosphere (Fig. 7d), which contributed to anomalous heating at low levels (p > 850 hPa; Figs. 5e and 7c). The vertical gradient of diabatic heating is an important contributor to the vorticity budget: decreases in heating with height generate negative (anticyclonic) vorticity and vice versa (e.g., Zeng et al. 2019). Positive sensible heat flux anomalies that diminished with height (Fig. 7c) thus contributed to maintaining the low-level anticyclone. The composite low-level high extended across much of Mongolia and North China, favoring northeasterly wind anomalies from high latitudes over the eastern part of North China (Fig. 7c). A cyclonic anomaly over the Yellow Sea and northwestern Pacific at middle and lower levels of the troposphere also favored anomalous northeasterly flow along the eastern coast of China and anomalous westerly flow over the North Pacific near the Luzon Strait. These circulation anomalies are characteristic of a weak EASM and an eastward retreat of the WPSH (Zhang et al. 2018). Northeasterly anomalies associated with these circulation patterns opposed the climatological southwesterlies, reducing moisture supplies to the lower and middle troposphere above North China (see also Fig. 5d).
b. Contributions of large-scale teleconnection patterns
Anomalous upper-tropospheric cooling during drought years was related to quasi-stationary wave patterns. Geopotential height anomalies in the middle and upper troposphere (Figs. 10a,b) reveal a well-organized wave train corresponding to the EU teleconnection pattern, with elements of the CGT along the subtropical jet (e.g., Wallace and Gutzler 1981; Ding and Wang 2005). Anomalous wave activity fluxes at 200 and 500 hPa during drought years (Figs. 10a,b) indicate a confluence of the eastward-propagating CGT wave train and the northwest–southeast-oriented EU wave train near 40°N, 110°E. The EU pattern upstream of North China is equivalent barotropic, with a slight northwestward tilt of the associated anomalies with height. By contrast, the CGT pattern is trapped along the upper-level westerly jet (around 40°N during JA; contours in Fig. 10d). Although less evident in the geopotential height fields, previous studies have also identified PJ-like patterns as influencing precipitation in East Asia (Wu and Wang 2002; Kosaka et al. 2011; Hirota and Takahashi 2012). The PJ is often considered as a “connector” linking tropical influences to East Asian climate (e.g., Xie et al. 2009; Tao et al. 2017).
The EU pattern index (EUI) is calculated following Wang and He (2015) as the linear combination of area-mean 500-hPa geopotential heights in three anomaly centers:
The CGT index (CGTI) is defined as the leading principal component of 200-hPa meridional wind within the domain 20°–60°N, 0°–150°E, consistent with the Silk Road pattern defined by Hong and Lu (2016). Similar to Kosaka et al. (2013), we define the PJ index (PJI) as the leading principal component of 850-hPa relative vorticity within 0°–45°N, 100°–160°E. These definitions for the CGTI, EUI, and PJI are adopted to emphasize distinctions among the teleconnections, respectively representing equivalent-barotropic forcing emanating from high latitudes (EU; 500-hPa geopotential height), low-level forcing emanating from the tropics (PJ; 850-hPa vorticity), and upper-level forcing propagating along the subtropical jet (CGT; 200-hPa meridional wind).
Figure 11 shows 200-hPa geopotential height anomalies regressed onto the EUI, CGTI, and PJI. The EU pattern extends from high-latitude Eurasia southeastward to North China, with cyclonic anomalies over the Ural Mountains and East Asia bracketing an anticyclonic anomaly over Mongolia (Fig. 11a). The CGT pattern consists of a series of alternating circulation anomalies in the upper troposphere, aligned east-to-west across the Eurasian continent near 40°N (Fig. 11b). The sign of the CGTI has been reversed in Fig. 11b, as dry conditions in North China are typically associated with the negative phase of the CGT (Fig. 11e). The PJ pattern extends northward from the tropical western Pacific in a series of zonally elongated anomalies in lower tropospheric vorticity. In the upper troposphere, the positive phase of the PJ pattern is associated with a widespread cyclonic anomaly over southern Asia (Fig. 11c), consistent with the 200-hPa zonal wind anomalies shown in Fig. 10d. The positive EU, positive PJ, and negative CGT patterns are thus all associated with cyclonic anomalies of varying magnitudes and distributions in the upper troposphere above eastern China. All five drought years occurred when the EUI and PJI were both positive (Figs. 11d,f), while four out of the five selected drought years (all except for 1997) occurred when the JA-mean CGTI was negative (Fig. 11e). Moreover, ERA5 indicates a transition in the CGT from predominantly positive to predominantly negative around the late 1990s (Fig. 11e), perhaps contributing to the greater abundance of drought years later in the analysis period.
(left) Regression maps of geopotential height anomalies (shading) at 200 hPa onto (a) the EU index (EUI = 1), (b) the negative CGT index (CGTI = −1), and (c) the PJ index (PJI = 1). Fits are calculated using ordinary least squares regression against JA-mean values of each index. Dotted areas denote statistical significance at the 95% confidence level based on Student’s t test. (right) Time series of indices for the (d) EU, (e) CGT, and (f) PJ indices (see text for details).
Citation: Journal of Climate 33, 23; 10.1175/JCLI-D-20-0176.1
Based on these results, we hypothesize that severe drought in North China results from co-occurrence of the positive EU pattern, which emanates from the high latitudes over Siberia and its Arctic coast (H. Li et al. 2018; J. Zhang et al. 2019), and the positive PJ pattern, which emanates from low latitudes in the western Pacific warm pool (Nitta 1987; Kosaka and Nakamura 2010). The former establishes an upper-level cyclone and associated upper-tropospheric cooling over North China, while the latter provides an additional source of low-level anticyclonic vorticity. This combination results in a deep layer of subsidence that dynamically suppresses precipitation. Support for this hypothesis arises from evidence implicating dynamical circulation anomalies as the primary driver of drought in this region (Figs. 4, 6, 8, and 9), as well as strong and significant correlations between both teleconnections and precipitation anomalies in North China (EU: r = −0.70; PJ: r = −0.49). Based on differences between 1997 and the other drought years, we further speculate that the CGT may modulate the mechanisms of drought development in North China. The CGTI and EUI were significantly anticorrelated during 1979–2017. Accordingly, four out of five drought years occurred when the EUI was positive and the CGTI was negative. In this combination, the EU, PJ, and CGT patterns are all associated with cyclonic anomalies in the upper troposphere above North China (Figs. 11a–c), thus favoring upper-tropospheric cooling. By contrast, a positive CGT combined with positive EU and positive PJ introduces an upper-level anomaly that opposes the EU-related cyclone over North China. The anticyclone then builds from below as negative low-level vorticity enters the region from both the northwest (along the EU wave train) and the southeast (along the PJ wave train).
The large-scale teleconnections contributed to atmospheric moisture and energy budget anomalies above North China during drought years. Figure 12 shows linear regressions of JA-mean vertical moisture and energy advection anomalies onto the EU, CGT, and PJ indices for all years during 1979–2017. Composite values of the EUI (1.44), the CGTI (−0.38), and the PJI (0.85) averaged over the five drought years are then used to reconstruct expected anomalies based on these linear fits. Reductions in both horizontal moisture convergence (
Vertical profiles of JA area-mean anomalies in (a) horizontal moisture convergence and (b) horizontal geopotential divergence over North China onto the EU, PJ, and CGT indices. Relationships are based on linear regression over the entire 1979–2017 period, with reconstructed profiles based on composite values of the indices (EUI = 1.44, CGTI = −0.38, PJI = 0.85) averaged over the five selected drought years. The × symbols indicate anomalies that are statistically significant at the 95% confidence level based on Student’s t test.
Citation: Journal of Climate 33, 23; 10.1175/JCLI-D-20-0176.1
c. Two types of drought events: Cold core and warm core
Figure 13 summarizes relationships between summer precipitation anomalies over North China and the factors discussed above, including key terms in the moisture and energy budgets (Figs. 13a,b), local circulation anomalies (Fig. 13c), local land surface feedback (Fig. 13d), and large-scale teleconnections (Figs. 13e,f). All five drought years exhibited negative anomalies in vertical advection of moisture and DSE, indicating both reduced moisture convergence into North China (as this term peaks in the lower troposphere) and reduced DSE divergence out of North China (as this term peaks in the upper troposphere). Four of five drought years (i.e., 1999, 2002, 2014, and 2015) were characterized by negative anomalies in meridional moisture advection, positive anomalies in meridional DSE advection, upper-level cyclonic anomalies, positive land surface feedback, positive EUI and PJI, and negative CGTI. Among these drought years, 2002 and 2015 most clearly met all criteria: moisture budget anomalies, energy budget anomalies, and local circulation anomalies were consistent with the composite mean; local surface feedback was strong; the EUI and PJI were large and positive; the CGTI was negative; and ENSO was in its warm phase. The year 2014 also nominally met all criteria, but the large-scale teleconnection indices were relatively weak compared to 2002 and 2015. The year 1999 was consistent with other drought years in most aspects, but differed in that La Niña conditions were present in the tropical Pacific. Positive anomalies in sensible heat flux during JA of 1999 (+6.9 W m−2) and 2014 (+6.8 W m−2) were large relative to the mean over all other positive-EU years (+2.4 ± 1.5 W m−2), serving to reinforce the low-level anticyclone. The large magnitude of negative precipitation anomalies in 1999 and 2014 despite relatively small values of the EUI, PJI, and CGTI underscores the role played by local land surface feedback in supporting the development and persistence of summer drought in North China. The year 1997 was unique among the drought years, with a positive anomaly in meridional moisture advection, a negative anomaly in meridional DSE advection, an upper-level anticyclonic anomaly, and a positive CGTI.
Scatterplots of JA area-mean (a) dynamic components of meridional and vertical moisture advection, (b) dynamic components of meridional and vertical dry static energy advection, (c) upper-level cyclone and lower-level anticyclone (i.e., normalized geopotential height anomalies at 200 and 850 hPa), and (d) normalized surface sensible heat flux and radiation flux anomalies over North China, along with (e) EU and CGT indices and (f) PJ and Niño-3.4 indices. Anomalies are shown for all 39 individual years. Years with negative precipitation anomalies that do not meet the drought threshold are shown in brown and years with positive precipitation anomalies are shown in green. Drought years are marked in red. The size of each symbol indicates the absolute magnitude of the precipitation anomaly [see key at upper left of (a)].
Citation: Journal of Climate 33, 23; 10.1175/JCLI-D-20-0176.1
Closer examination of individual years raises the possibility of two types of dynamically driven drought events. Composite-mean geopotential height anomalies for drought years (Fig. 5f) show a low-level anticyclonic anomaly (Fig. 10c) that weakens with height, transitioning to a cyclonic anomaly in the upper troposphere (Fig. 10a). This composite picture qualitatively describes the vertical distribution of JA-mean geopotential height anomalies during all of the drought years except for 1997 (Figs. S6a–c and S7a–c). The heart of this circulation system is the upper-level cyclonic anomaly. The associated upper-tropospheric cooling (Figs. 5e and 7a) drives anomalous subsidence through the middle and upper troposphere (Fig. 5c), dynamically inducing an anomalous low-level high. These features are accompanied by a cold-core thermal anomaly centered over North China (Fig. 14a).
Spatial distributions of vertically integrated anomalies in thermal energy (cpT; shading) and 850-hPa geopotential (gz; contours) averaged over JA for (a) the drought years 1999, 2002, 2014, and 2015 and (b) the drought year 1997. The contour interval for 850-hPa geopotential is 25 m2 s−2, starting from ±25 m2 s−2. Dotted lines indicate negative anomalies. Stippling indicates anomalies that are statistically significant at the 95% confidence level based on Student’s t test.
Citation: Journal of Climate 33, 23; 10.1175/JCLI-D-20-0176.1
Conditions in 1997 differed, instead featuring an anticyclonic anomaly anchored in the lower troposphere that intensified with height (Figs. S6d–f and S7d–f), consistent with a warm-core system centered over North China (Fig. 14b). Differences in anomalous meridional advection of moisture and energy (Figs. 13a,b) offer a framework for understanding differences between this warm-core situation and the cold-core drought composite. In JA of 1997, meridional moisture inflow was larger than normal and meridional DSE export was smaller than normal. Both anomalies were opposite to those for the other four drought years. In the composite case, the reduction of meridional moisture advection reflects anomalous flow from the relatively dry northeast, which restricted the moisture supply for precipitation, while the increase in meridional DSE export helped to maintain negative anomalies in vertically integrated thermal energy over North China (i.e., the “cold core”; Fig. 14a). Collectively, these anomalies reduced dry static stability (cooling at upper levels overlying warming at lower levels) but enhanced moist static stability (strong drying at lower levels). In 1997, enhanced meridional moisture advection was associated with relatively moist near-surface southeasterlies and the reduction in DSE export helped to maintain the “warm” core of the anticyclone (Fig. 14b). These meridional advection anomalies thus enhanced dry static stability while maintaining moisture supply at low levels. These conditions permitted sporadic rainfall events that interrupted the prevailing dry conditions during JA of 1997 (Fig. S8), in contrast to more persistent negative precipitation anomalies during other drought years.
6. Discussion
Our results provide new perspectives on previous resulting highlighting the influences of the EU, CGT, and PJ teleconnection patterns on precipitation anomalies in East Asia (e.g., Wu and Wang 2002; Kosaka et al. 2011; Hirota and Takahashi 2012; Wang and He 2015; Wang et al. 2017; Li et al. 2018a,b; Zhang et al. 2018; J. Zhang et al. 2019). Although our conclusions are broadly consistent with these prior studies, they are nonetheless limited by the small sample size: our criteria select only five drought years, within which we identify two distinct types. This small sample size is the main source of uncertainty regarding both our interpretations and the generalizability of our results. Further analysis based on free-running model simulations will be needed to address this limitation. In the meantime, we can test our understanding by examining those years on the margins; that is, years for which precipitation anomalies were large and negative but did not reach the drought threshold (e.g., 1986, 1991, and 2001), or years for which circulation patterns ostensibly favored drought but dry conditions did not develop (e.g., 2000 and 2011).
For years that almost met the drought threshold, differences in local land surface feedbacks are a distinguishing factor. Sensible heat flux anomalies in 1986 were less than a third of the average anomaly from the five selected drought years, despite large-scale circulation indices in line with those of the drought years. A similar argument can be made for 1991: despite large-scale circulation indices comparable to those in 2014, sensible heat flux anomalies were less than half of those in 2014. Perhaps for this reason, the low-level anticyclonic anomaly did not develop fully in either 1986 or 1991. Given the potential for model biases in land surface conditions or precipitation amounts (see below) to impact the reanalysis-derived land surface response we cannot judge whether these differences are realistic. However, it seems probable that weak land surface feedbacks during these years mitigated drought conditions at least in the context of ERA5. Both years did meet the drought threshold in GPCP; however, including these years in the composite has no noticeable impact on our conclusions beyond raising the number of cold-core-type events from 4 to 6 (see Fig. 13). The situation for 2001 is initially more difficult to parse, as both circulation anomalies and land surface feedbacks were consistent with those of drought years. However, this year featured strong dry anomalies immediately to the south of our analysis domain and would meet the drought threshold if the southern boundary were shifted south to 30°N. 2001 was also a typical cold-core year by the characteristics outlined in section 5c.
Among years when dry conditions did not develop, only 2000 and 2011 featured both positive EU and positive PJ. The CGTI was negative in both years, as in the typical cold-core case. However, the land surface response was weak, the lower tropospheric anticyclone did not develop, and JA precipitation anomalies in North China were within ±0.5 mm day−1 of the 1979–2017 mean (Fig. 1). In future work, it may be useful to examine the detailed seasonal evolution of the circulation patterns and surface feedbacks that favor drought in North China between drought years and years like 2000 and 2011. The timing when each factor appears may be critical for determining whether the land surface feedback is engaged or avoided, particularly given the stepwise northward progression of the EASM rainband (e.g., Ding and Chan 2005).
Our use of reanalysis products should also be considered when interpreting the results, as illustrated by uncertainty in whether 1986 and 1991 were drought years. On the one hand, detailed evaluation of historical variability in regional moisture and energy budgets is only possible through reanalyses or other observationally constrained model simulations. On the other hand, biases in the forecast model impact reanalysis performance, and changes in assimilated observations may lead to spurious trends or variability [see, e.g., Fujiwara et al. (2017) and references therein]. These issues can be particularly acute for variables like precipitation, evaporation, and diabatic heating that are only indirectly constrained by data assimilation. ERA5 represents the state of the art among atmospheric reanalyses, with an hourly analysis cycle, expanded capabilities for screening and assimilating observational data, and many improvements to moist physics in the forecast model. This reanalysis is thus particularly appealing for studying drought development. Overall, the annual cycles of precipitation and evaporation over North China and the spatial structure of precipitation anomalies during drought years are captured well by ERA5 (see section 3). However, comparison with GPCP and GLEAM reveals some potentially important discrepancies.
First, although both ERA5 and GPCP suggest that drought events were more frequent after the mid-1990s, ERA5 shows a much stronger decreasing trend in JA precipitation over 1979–2017 (−0.032 mm day−1 yr−1; statistically significant at the 99% confidence level) than supported by GPCP (−0.007 mm day−1 yr−1; not statistically significant). The climate shift detection algorithm proposed by Rodionov (2004) suggests that this negative trend arises from step-like reductions in the positive bias of ERA5 precipitation relative to observations around the years 1998, 2006, and 2013. Each of these “steps” corresponds to an introduction or expansion of satellite data assimilated by ERA5 that may have tightened constraints on the moist thermodynamic state of the atmosphere (Hersbach et al. 2018).
Second, ERA5 may not adequately capture the interannual variability of evaporation in North China. Agreement between ERA5 and GLEAM is particularly poor between 1997 and 2002, when three of the five drought years occurred. Such differences are particularly striking as atmospheric boundary conditions for GLEAM v3.3 are taken from ERA5. As discussed in section 3, disagreements between ERA5 and GLEAM largely concern the northeastward extent of positive anomalies centered in the Sichuan Basin (Fig. 3). ERA5 also underestimates negative anomalies in the eastern portion of the domain near Hebei and Shandong provinces. These biases may be rooted in errors in precipitation as discussed above, but may also involve issues in the land surface model. A separate ERA5-Land product using ERA5 fields to drive an updated, higher-resolution version of the ECMWF land surface model has recently been released as a companion to ERA5. Anomalies based on ERA5-Land during drought years (Fig. S9) are in better agreement with those based on GLEAM in the eastern part of the domain, suggesting that recent updates to the land surface model are impactful there. Additional work will be necessary to further clarify the reliability of ERA5 with respect to the climate of monsoonal East Asia and its variability in time.
7. Summary and conclusions
North China has experienced frequent summer drought events in recent decades, but the detailed physical mechanisms behind these drought events remain unclear. To help clarify these mechanisms, we have analyzed regional moisture and energy budgets during summer drought years in North China. Five severe summer drought years have been identified based on JA precipitation anomalies during 1979–2017 in ERA5 and GPCP. Summer precipitation minus evaporation anomalies during drought years were dominated by negative precipitation anomalies during JA, with area-mean evaporation anomalies statistically indistinguishable from zero. Moisture budget analysis indicates that precipitation deficits in drought years were driven by circulation changes that reduced both horizontal and vertical moisture advection. Anomalous northeasterlies inhibited the flow of moisture into North China, while anomalous subsidence indicates reduced horizontal moisture convergence. Energy budget analysis links reduced latent heat release during summer drought to decreased DSE flux divergence. The latter was contributed mainly by reduced geopotential flux divergence associated with anomalous subsidence in the middle-to-upper troposphere.
Evidence of two types of drought events emerges on examination of the individual years. In the typical case, which matches four of the five drought years and three dry years that did not meet our drought threshold (section 6), an upper-level cyclonic anomaly developed where the EU wave train intersected a second wave train characteristic of the CGT. This upper-level cyclonic anomaly caused strong cooling in the middle-to-upper troposphere over North China, which dynamically forced descending motion and the development of an anticyclonic anomaly in the lower troposphere. This anticyclonic anomaly was reinforced by local land surface feedbacks. The latter are influential in determining the severity of dry anomalies associated with this type of event. We refer to this type of event as “cold core” because it features negative anomalies in vertically integrated thermal energy over North China. Positive anomalies in area-mean geopotential near the surface decreased with height, characteristic of cold-core highs. The second type of drought event occurred only once in our analysis period, in 1997. This event was distinct from the others in that the anticyclonic anomalies were anchored in the lower troposphere (rather than at the surface) and increased with height (rather than decreasing). This drought event was associated with positive anomalies in vertically integrated thermal energy over North China, and is therefore referred to as “warm core.”
Summer drought in North China was marked by reduced moisture flux convergence in the moisture budget framework and by reduced latent heat release and DSE flux divergence in the energy budget framework. These features are common to both cold-core and warm-core events and are intimately linked to anomalous subsidence in the troposphere. Differences between 1997 and the other drought years can be largely explained by differences in meridional advection. Cold-core events were characterized by reduced meridional moisture inflow and increased meridional export of DSE. The former suppressed precipitation by inhibiting moisture supply, while the latter reduced dry static stability and helped to maintain negative anomalies in thermal energy. By contrast, the 1997 warm-core event featured enhanced meridional moisture inflow and reduced meridional export of DSE. The latter acted to enhance dry static stability (maintaining the warm core); however, increased moisture inflow at low levels permitted a greater frequency of intermittent rainfall events relative to the cold-core drought years.
Cold-core events were exclusively associated with the negative phase of the CGT, the most common state of the CGT during positive-EU years. By contrast, the warm-core event occurred when the CGT and EU patterns were both positive, raising the possibility that the CGT may modulate the mechanisms of drought development. More analysis will be needed to evaluate this possibility. To leading order, however, summer drought in North China is linked to co-occurrence of the positive EU and positive PJ patterns. All five drought years occurred when both EUI and PJI were positive, and strong dry anomalies developed in 8 out of 10 years that met this criterion. The results provide a new perspective on the physical processes underlying drought in North China and raise several promising lines for further inquiry. In pursuing these, additional work will also be needed to establish the general applicability of the results, especially in light of the small sample size and other limitations intrinsic to our use of reanalysis products.
Acknowledgments
We thank Dr. Amir Erfanian for valuable suggestions that helped to guide the data analysis conducted in this work. This research was supported by a research grant from the Ministry of Science and Technology of the People’s Republic of China to Tsinghua University (2017YFA0603902). Part of the work was conducted at UCLA with the support of a Tsinghua University fellowship awarded to Lan Dai. Rong Fu was supported by the National Oceanic and Atmospheric Administration Climate Program Office, Modeling, Analysis, Predictions, Projections Program (Award NA170AR4310123).
Data availability statement
ERA5 and ERA5-Land products except for diabatic heating rates were acquired from the Copernicus Climate Change Services (C3S) Climate Data Store (https://cds.climate.copernicus.eu). Diabatic heating rates were acquired from the ECMWF public archive (https://apps.ecmwf.int/data-catalogues/era5). GPCP precipitation data were acquired from the NOAA/OAR/ESRL Physical Sciences Division (https://www.esrl.noaa.gov/psd). GLEAM data were acquired from the archive maintained by the GLEAM project (https://www.gleam.eu). Access information for datasets not appearing in the main text is provided in the supplemental information.
REFERENCES
Adler, R., and Coauthors, 2003: The version 2 Global Precipitation Climatology Project (GPCP) monthly precipitation analysis (1979–present). J. Hydrometeor., 4, 1147–1167, https://doi.org/10.1175/1525-7541(2003)004<1147:TVGPCP>2.0.CO;2.
Akinsanola, A., and W. Zhou, 2019: Dynamic and thermodynamic factors controlling increasing summer monsoon rainfall over the West African Sahel. Climate Dyn., 52, 4501–4514, https://doi.org/10.1007/s00382-018-4394-x.
Chou, C., J. Neelin, C. Chen, and J. Tu, 2009: Evaluating the “rich-get-richer” mechanism in tropical precipitation change under global warming. J. Climate, 22, 1982–2005, https://doi.org/10.1175/2008JCLI2471.1.
Dai, A., 2011: Drought under global warming: A review. Wiley Interdiscip. Rev.: Climate Change, 2, 45–65, https://doi.org/10.1002/wcc.81.
Dee, D., and Coauthors, 2011: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc., 137, 553–597, https://doi.org/10.1002/qj.828.
Ding, Q., and B. Wang, 2005: Circumglobal teleconnection in the Northern Hemisphere summer. J. Climate, 18, 3483–3505, https://doi.org/10.1175/JCLI3473.1.
Ding, Q., B. Wang, J. Wallace, and G. Branstator, 2011: Tropical–extratropical teleconnections in boreal summer: Observed interannual variability. J. Climate, 24, 1878–1896, https://doi.org/10.1175/2011JCLI3621.1.
Ding, Y., and J. Chan, 2005: The East Asian summer monsoon: An overview. Meteor. Atmos. Phys., 89, 117–142, https://doi.org/10.1007/s00703-005-0125-z.
Emori, S., and S. Brown, 2005: Dynamic and thermodynamic changes in mean and extreme precipitation under changed climate. Geophys. Res. Lett., 32, L17706, https://doi.org/10.1029/2005GL023272.
Fujiwara, M., and Coauthors, 2017: Introduction to the SPARC Reanalysis Intercomparison Project (S-RIP) and overview of the reanalysis systems. Atmos. Chem. Phys., 17, 1417–1452, https://doi.org/10.5194/acp-17-1417-2017.
Gelaro, R., and Coauthors, 2017: The Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2). J. Climate, 30, 5419–5454, https://doi.org/10.1175/JCLI-D-16-0758.1.
Harada, Y., and Coauthors, 2016: The JRA-55 Reanalysis: Representation of atmospheric circulation and climate variability. J. Meteor. Soc. Japan, 94, 269–302, https://doi.org/10.2151/jmsj.2016-015.
Hersbach, H., and Coauthors, 2018: Operational global reanalysis: Progress, future directions and synergies with NWP. ERA Rep. Series 27, 63 pp., https://www.ecmwf.int/sites/default/files/elibrary/2018/18765-operational-global-reanalysis-progress-future-directions-and-synergies-nwp.pdf.
Hersbach, H., and Coauthors, 2020: The ERA5 global reanalysis. Quart. J. Roy. Meteor. Soc., 146, 1999–2049, https://doi.org/10.1002/qj.3803.
Hirota, N., and M. Takahashi, 2012: A tripolar pattern as an internal mode of the East Asian summer monsoon. Climate Dyn., 39, 2219–2238, https://doi.org/10.1007/s00382-012-1416-y.
Hong, X., and R. Lu, 2016: The meridional displacement of the summer Asian jet, Silk Road pattern, and tropical SST anomalies. J. Climate, 29, 3753–3766, https://doi.org/10.1175/JCLI-D-15-0541.1.
Huang, G., Y. Liu, and R. Huang, 2011: The interannual variability of summer rainfall in the arid and semiarid regions of northern China and its association with the northern hemisphere circumglobal teleconnection. Adv. Atmos. Sci., 28, 257–268, https://doi.org/10.1007/s00376-010-9225-x.
Huang, R., 1992: The East Asia/Pacific pattern teleconnection of summer circulation and climate anomaly in East Asia. J. Meteor. Res., 6, 25–37.
Kobayashi, S., and Coauthors, 2015: The JRA-55 Reanalysis: General specifications and basic characteristics. J. Meteor. Soc. Japan, 93, 5–48, https://doi.org/10.2151/jmsj.2015-001.
Kosaka, Y., and H. Nakamura, 2010: Mechanisms of meridional teleconnection observed between a summer monsoon system and a subtropical anticyclone. Part I: The Pacific–Japan pattern. J. Climate, 23, 5085–5108, https://doi.org/10.1175/2010JCLI3413.1.
Kosaka, Y., S.-P. Xie, and H. Nakamura, 2011: Dynamics of interannual variability in summer precipitation over East Asia. J. Climate, 24, 5435–5453, https://doi.org/10.1175/2011JCLI4099.1.
Kosaka, Y., S.-P. Xie, N.-C. Lau, and G. A. Vecchi, 2013: Origin of seasonal predictability for summer climate over the northwestern Pacific. Proc. Natl. Acad. Sci. USA, 110, 7574–7579, https://doi.org/10.1073/pnas.1215582110.
Li, H., H. Chen, H. Wang, J. Sun, and J. Ma, 2018: Can Barents Sea ice decline in spring enhance summer hot drought events over northeastern China? J. Climate, 31, 4705–4725, https://doi.org/10.1175/JCLI-D-17-0429.1.
Li, X., D. Li, X. Li, and L. Chen, 2018: Prolonged seasonal drought events over northern China and their possible causes. Int. J. Climatol., 38, 4802–4817, https://doi.org/10.1002/joc.5697.
Liu, Y., and Coauthors, 2015: Agriculture intensifies soil moisture decline in northern China. Sci. Rep., 5, 11261, https://doi.org/10.1038/srep11261.
Lu, R.-Y., J.-H. Oh, and B.-J. Kim, 2002: A teleconnection pattern in upper-level meridional wind over the North African and Eurasian continent in summer. Tellus, 54A, 44–55, https://doi.org/10.3402/tellusa.v54i1.12122.
Martens, B., and Coauthors, 2017: GLEAM v3: Satellite-based land evaporation and root-zone soil moisture. Geosci. Model Dev., 10, 1903–1925, https://doi.org/10.5194/gmd-10-1903-2017.
Miralles, D. G., T. R. H. Holmes, R. A. M. De Jeu, J. H. Gash, A. G. C. A. Meesters, and A. J. Dolman, 2011: Global land-surface evaporation estimated from satellite-based observations. Hydrol. Earth Syst. Sci., 15, 453–469, https://doi.org/10.5194/hess-15-453-2011.
Muller, C., and P. O’Gorman, 2011: An energetic perspective on the regional response of precipitation to climate change. Nat. Climate Change, 1, 266–271, https://doi.org/10.1038/nclimate1169.
Nitta, T., 1987: Convective activities in the tropical western Pacific and their impact on the Northern Hemisphere summer circulation. J. Meteor. Soc. Japan, 65, 373–390, https://doi.org/10.2151/jmsj1965.65.3_373.
O’Gorman, P., R. Allan, M. Byrne, and M. Previdi, 2012: Energetic constraints on precipitation under climate change. Surv. Geophys., 33, 585–608, https://doi.org/10.1007/s10712-011-9159-6.
Peng, D., and T. Zhou, 2017: Why was the arid and semiarid northwest China getting wetter in the recent decades? J. Geophys. Res. Atmos., 122, 9060–9075, https://doi.org/10.1002/2016JD026424.
Rodionov, S. N., 2004: A sequential algorithm for testing climate regime shifts. Geophys. Res. Lett., 31, L09204, https://doi.org/10.1029/2004GL019448.
Seager, R., and N. Henderson, 2013: Diagnostic computation of moisture budgets in the ERA-Interim reanalysis with reference to analysis of CMIP-archived atmospheric model data. J. Climate, 26, 7876–7901, https://doi.org/10.1175/JCLI-D-13-00018.1.
Seager, R., N. Naik, and G. Vecchi, 2010: Thermodynamic and dynamic mechanisms for large-scale changes in the hydrological cycle in response to global warming. J. Climate, 23, 4651–4668, https://doi.org/10.1175/2010JCLI3655.1.
Stephan, C., N. Klingaman, P. Vidale, A. Turner, M. Demory, and L. Guo, 2018: A comprehensive analysis of coherent rainfall patterns in China and potential drivers. Part I: Interannual variability. Climate Dyn., 50, 4405–4424, https://doi.org/10.1007/s00382-017-3882-8.
Takaya, K., and H. Nakamura, 2001: A formulation of a phase-independent wave-activity flux for stationary and migratory quasigeostrophic eddies on a zonally varying basic flow. J. Atmos. Sci., 58, 608–627, https://doi.org/10.1175/1520-0469(2001)058<0608:AFOAPI>2.0.CO;2.
Tao, L., T. Li, Y.-H. Ke, and J.-W. Zhao, 2017: Causes of interannual and interdecadal variations of the summertime Pacific–Japan-like pattern over East Asia. J. Climate, 30, 8845–8864, https://doi.org/10.1175/JCLI-D-15-0817.1.
Trenberth, K., and C. Guillemot, 1995: Evaluation of the global atmospheric moisture budget as seen from analyses. J. Climate, 8, 2255–2272, https://doi.org/10.1175/1520-0442(1995)008<2255:EOTGAM>2.0.CO;2.
Wallace, J., and D. Gutzler, 1981: Teleconnections in the geopotential height field during the Northern Hemisphere winter. Mon. Wea. Rev., 109, 784–812, https://doi.org/10.1175/1520-0493(1981)109<0784:TITGHF>2.0.CO;2.
Wang, H., and S. He, 2015: The north China/northeastern Asia severe summer drought in 2014. J. Climate, 28, 6667–6681, https://doi.org/10.1175/JCLI-D-15-0202.1.
Wang, S., X. Yuan, and Y. Li, 2017: Does a strong El Niño imply a higher predictability of extreme drought? Sci. Rep., 7, 40741, https://doi.org/10.1038/srep40741.
Wei, J., Q. Zhang, and S. Tao, 2004: Physical causes of the 1999 and 2000 summer severe drought in North China (in Chinese). Chin. J. Atmos. Sci., 28, 125–137.
Wills, R., and T. Schneider, 2015: Stationary eddies and the zonal asymmetry of net precipitation and ocean freshwater forcing. J. Climate, 28, 5115–5133, https://doi.org/10.1175/JCLI-D-14-00573.1.
Wu, L., and J. Zhang, 2013: Role of land–atmosphere coupling in summer droughts and floods over eastern China for the 1998 and 1999 cases. Chin. Sci. Bull., 58, 3978–3985, https://doi.org/10.1007/s11434-013-5855-6.
Wu, R., and B. Wang, 2002: A contrast of the East Asian summer monsoon–ENSO relationship between 1962–77 and 1978–93. J. Climate, 15, 3266–3279, https://doi.org/10.1175/1520-0442(2002)015<3266:ACOTEA>2.0.CO;2.
Xia, J., L. Zhang, C. Liu, and J. Yu, 2007: Towards better water security in North China. Water Resour. Manage., 21, 233–247, https://doi.org/10.1007/s11269-006-9051-1.
Xie, S.-P., K. Hu, J. Hafner, H. Tokinaga, Y. Du, G. Huang, and T. Sampe, 2009: Indian Ocean capacitor effect on Indo–western Pacific climate during the summer following El Niño. J. Climate, 22, 730–747, https://doi.org/10.1175/2008JCLI2544.1.
Xu, Z., K. Fan, and H. Wang, 2017: Role of sea surface temperature anomalies in the tropical Indo-Pacific region in the northeast Asia severe drought in summer 2014: Month-to-month perspective. Climate Dyn., 49, 1631–1650, https://doi.org/10.1007/s00382-016-3406-y.
Yu, C., and Coauthors, 2018: Assessing the impacts of extreme agricultural droughts in China under climate and socioeconomic changes. Earth’s Future, 6, 689–703, https://doi.org/10.1002/2017EF000768.
Yu, R., B. Wang, and T. Zhou, 2004: Tropospheric cooling and summer monsoon weakening trend over East Asia. Geophys. Res. Lett., 31, L22212, https://doi.org/10.1029/2004GL021270.
Zeng, D., X. Yuan, and J. K. Roundy, 2019: Effect of teleconnected land–atmosphere coupling on northeast China persistent drought in spring–summer of 2017. J. Climate, 32, 7403–7420, https://doi.org/10.1175/JCLI-D-19-0175.1.
Zhang, J., H. Chen, and Q. Zhang, 2019: Extreme drought in the recent two decades in northern China resulting from Eurasian warming. Climate Dyn., 52, 2885–2902, https://doi.org/10.1007/s00382-018-4312-2.
Zhang, L., P. Wu, T. Zhou, and C. Xiao, 2018: ENSO transition from La Niña to El Niño drives prolonged spring–summer drought over North China. J. Climate, 31, 3509–3523, https://doi.org/10.1175/JCLI-D-17-0440.1.
Zhang, L., T. Zhou, P. Wu, and X. Chen, 2019: Potential predictability of North China summer drought. J. Climate, 32, 7247–7264, https://doi.org/10.1175/JCLI-D-18-0682.1.