**1. Introduction**

The summer of 2013 was unprecedentedly hot in Eastern China, causing substantial societal and economic impacts [1]. Such a phenomenon has drawn widespread concerns, and the physical mechanism behind such heatwave is gradually being discovered [2–4]. The changes of large-scale atmospheric circulations may be the main cause of temperature anomalies, and small-scale physical processes of local energy balance such as soil moisture–atmosphere coupling could also make a contribution to them [5]. Many studies have shown that soil moisture anomalies play an important role in soil moisture–temperature coupling [6,7], as it could control the energy budget by the partitioning of latent heat flux and sensible heat flux, further impacting the air temperature [8,9]. When soil moisture decreases, less water can be used for evapotranspiration, resulting in a decrease of latent heat flux [8,10]. Based on the energy balance, the decline of latent heat flux causes an increase of the sensible heat flux, thus enhancing the air temperature. These conditions indicate a negative feedback between soil moisture and air temperature: Soil moisture deficit results in the rise of air temperature [11–13].

Nowadays, various studies have shown that dry soil moisture conditions can have a substantial influence on the severity of heat waves and drought through the coupling between soil moisture and atmosphere [10,11]. As shown, soil moisture–temperature coupling helps to explain heat waves in summer climate [14,15]. Modeling experiments have focused on identifying the strong coupling regions, and how these regions are influenced by the changing climate [14,16]. The Global Land–Atmosphere Coupling (GLACE) project indicated that the strongest coupling regions (hot spots) of soil moisture–temperature are between the transitional regions of wet and dry climates [16,17]. In addition, some numerical experiments have been devoted to studying the soil moisture–temperature coupling at regional scales [18–20] finding that soil moisture anomalies impact air temperature during summer mainly in areas like Northern China [21–23].

At present, many studies have tried to use di fferent metrics for assessing land–atmosphere coupling strengths. Koster et al. [16] proposed to use correlation between evapotranspiration and temperature, and found results agreeing with other metrics, for example, those based on the correlation between evapotranspiration and radiation. Gallego-Elvira et al. [24] used the dependence of the surface heating rate of di fferent cover types on the previous precipitation during drought to identify di fferent evaporative regions. Dirmeyer [25] devoted an index of surface flux sensitivity to soil moisture variability and applied it to global atmospheric reanalysis datasets. However, most studies assessing the coupling of soil moisture to climate were based primarily on summer (June–July–August, JJA) [9,14], but there is a lack of understanding on its seasonal changes; therefore, it is important to devote more attention to studying the coupling within the other seasons as well. Moreover, the studies of the soil moisture–temperature coupling were mostly based on modeling experiments [10,26], which show a large di fference in the regions and strengths of land–atmosphere coupling [8]. However, the limited ground measurements of soil moisture cannot meet the research needs of land–atmosphere coupling on regional scales. Given the limitation of ground measurements, satellite data could be used in soil moisture–temperature coupling studies from an observational point of view, and the recent development of soil moisture and evapotranspiration products from remote sensing technology provides the possibility for such studies.

A wide variety of datasets makes it possible for us to study soil moisture–temperature coupling from an observational perspective [24,27,28]. This study utilized one of these satellite-based datasets with a coupling diagnostic to show the spatial distribution and interannual variation of strong coupling regimes between soil moisture and temperature over China in di fferent seasons. This diagnostic focused on two di fferent timescales to fill the gap between extremes and climatological studies of soil moisture–temperature coupling. Di fferent season coupling hot spots of China are illustrated in the following sections. Subsequently, we explored the role of soil moisture during the 2013 heatwave in Southeast China and the 2009 event in North China.

Furthermore, this study depicts not only soil moisture–temperature coupling in long-term variations, but also soil moisture–temperature coupling during the heat wave events and the related heating processes. It may help to point us toward better understanding the underlying processes of soil moisture–temperature coupling, and thus improve the prediction skills of heat waves [14,29].

#### **2. Materials and Methods**

## *2.1. Study Area*

China has a complex climate due to its topography [30]. Mainland China is generally divided into three di fferent climatic zones: Arid region with annual precipitation below 200 mm, humid region with annual precipitation more than 800 mm, and transitional region with annual precipitation from 200 to 800 mm [31]. The arid region mainly includes Xinjiang province and western Inner Mongolia Plateau. The humid region is mainly South China. The transitional region mainly includes North China Plain, Northeast Plain, and part of Tibetan Plateau, where there is more rain in summer but less in winter. Known as the "The Third Pole", the Tibetan Plateau is the highest and most unique

geographical unit on earth [32]. It is one of the most sensitive regions to global climate change and its hydrological processes are quite different from that of other regions of China [33,34]; thus, it is regarded as a separate typical region in this study. The humid, transitional, and dry regions, as well as the Tibetan Plateau, are shown in Figure 1.

**Figure 1.** The typical climate regions of China (arid, humid, transitional, the Tibetan Plateau).

## *2.2. Data Sources*

The evapotranspiration (ET) and potential evapotranspiration (PET) data from the GLEAM v3.0a (Global Land Evaporation Amsterdam Model) product were used in this study, which span the period from 1980 to 2015 with a spatial resolution of 0.25◦. The GLEAM model is a simplified land model, which is fully dedicated to estimating the terrestrial evaporation and root zone soil moisture based on satellite data [35]. It comprises a set of algorithms using multiyear satellite observations to estimate the components of terrestrial ET. The PET is calculated within the Priestley–Taylor equation via the observations of net radiation and near-surface air temperature [36]. The 2-m air temperature and the top layer (0–7 cm) volumetric soil moisture from the ERA-Interim reanalysis data were used [37].
