HTML
--> --> -->Recently, semi-arid climate change in China has varied significantly. For instance, (Huang et al., 2012) found that the largest warming in China during the past 100 years has occurred over semi-arid regions and has been more significant in winter (Ma and Fu, 2003; Ji et al., 2014). Meanwhile, P over the semi-arid regions in China has shown a "wet-west-dry-east" pattern, but has only experienced a slight increase over western semi-arid regions and has not clearly mitigated the drought stress (Gong et al., 2004; Zhang et al., 2011; Li et al., 2015). Moreover, the climate has become drier in northeastern China (Huang et al., 2016a). The population of semi-arid regions in China mainly relies on rainfed agriculture for their livelihood; therefore, climate fluctuations greatly affect the availability of surface water resources (Liu and Xia, 2004). Moreover, drought in semi-arid regions has been one of the most severe manifestations of climate variability in China over the past six decades (Piao et al., 2010). (Ma and Fu, 2003) suggested that the drought in semi-arid regions of northern China mainly results from the decrease in P and increase in evaporation. Conversely, semi-arid climate change strongly influences the economy and society and increases the challenges related to food security and water supply in China (Charney, 1975; Huang et al., 2010; Piao et al., 2010; Li et al., 2015). Climate scientists have achieved many outstanding achievements regarding semi-arid climate change in China. To develop appropriate strategies to cope with significant semi-arid climate change and maintain sustainable development in these regions, it is necessary to review and comment on the latest progress in semi-arid climate change and its dynamic mechanisms. Therefore, by focusing on semi-arid climate change against the background of contemporary climate and global change, this article summarizes the characteristics of semi-arid climate change and its dynamic mechanisms in terms of land-atmosphere interactions, ocean-atmosphere interactions, dust-cloud interactions, and human activities. Finally, the projection of aridity in these regions in the future is reviewed and discussed.
To define the semi-arid regions of drylands and study semi-arid climate change, aridity is the main term used to measure the degree of dryness or water deficiency in a region, as it implies that the permanent water deficiency is closely related to strong insolation, elevated temperature, strong evapotranspiration and low humidity (Mainguet, 1999). Different criteria and climate classification schemes have been proposed for quantifying the degree of aridity at a given location (e.g., K?ppen, 1884; Penman, 1948; Thornthwaite, 1948). For example, (Palmer, 1965) created the original Palmer Drought Severity Index (PDSI) that depends on P, soil moisture, streamflow and potential evapotranspiration (PET), which are mainly used in monitoring and researching drought; however, the PDSI has been criticized for its lack of comparability among diverse climatological regions owing to the empirical parameters used by Palmer in the United States. To make the PDSI usable in China, (An and Xing, 1986) and (Liu et al., 2004) modified the PDSI (denoted by PDSI_CN) with revised weightings and duration factors. (Wells et al., 2004) proposed the self-calibrating PDSI (scPDSI) via calibration by using local conditions to calculate the weightings and duration factors, which improved spatial comparability. In addition, the Standardized Precipitation Index (SPI) and Standardized Precipitation Evapotranspiration Index (SPEI) have been applied widely to monitor moisture conditions (Vicente-Serrano et al., 2010). Meanwhile, the Surface Wetness Index can be used to define drylands and track meteorological drought (Hulme et al., 1992), but it is restricted by the scarcity of station observations and problematic results derived from remote sensing (Kerr, 2007). In addition, the Aridity Index (AI), which is defined by the ratio of annual P to annual PET, is also considered as a measure of terrestrial aridity and is widely used for defining drylands and studying dryland climate change (e.g., Penman, 1948; Hulme, 1996; Feng and Fu, 2013; Huang et al., 2017a).
By evaluating the regional applicability of these schemes in China, (Yang et al., 2017a) found that the scPDSI is more appropriate for China but reduces the value range slightly compared to the PDSI and modified PDSI (PDSI_CN); thus, the classification of aridity should be adjusted accordingly. The SPI and SPEI are more appropriate for humid areas than arid and semi-arid areas because the contributions of temperature change to drought are neglected in the SPI but overestimated in the SPEI when the PET is estimated by the Thornthwaite method instead of the Penman-Monteith algorithm. Meanwhile, (Huang et al., 2016a) found that P alone inaccurately defines drylands because it ignores the effect of temperature and PET, while the K?ppen-Geiger climate classification underestimates the total drylands. Overall, using the AI to identify drylands and aridity produces the most reasonable and reliable results among these three methods. Currently, more researchers have generally divided the climate classification of drylands using AI criteria and considered that the PET estimated by the common Penman-Monteith equation is closer to the real-world conditions under all climates (An et al., 2014; Qian et al., 2017; Yang et al., 2017a). However, reconciling different measures of terrestrial aridity to address various environmental issues in semi-arid climate change is an important issue that should be addressed in future research (Fu et al., 2016).
Using the AI criteria, drylands are defined as areas with an AI (P/PET) less than 0.65, which can be further classified into hyper-arid (AI<0.05), arid (0.05≤ AI<0.2), semi-arid (0.2≤ AI<0.5) and dry sub-humid (0.5≤ AI<0.65) regions (e.g., Feng and Fu, 2013). Meanwhile, drylands using a P-based definition can be divided into hyper-arid (annual P<50 mm), arid (50 mm ≤ annual P<200 mm), semi-arid (200 mm ≤ annual P<400 mm) and dry sub-humid (400 mm ≤ annual P<600 mm) regions. Based on the AI and annual P, (Li et al., 2015) showed the climatological distribution of semi-arid regions of Chinese drylands for the period 1961-90, as shown in Fig. 1. The areas of semi-arid and dry sub-humid regions defined by the AI (Fig. 1a) are smaller than those defined by P (Fig. 1b), and the semi-arid regions of China are mainly distributed in a zonal band from midwest to northeast, with arid and hyper-arid regions to the northwest and dry sub-humid regions to the southeast. The spatial distribution of semi-arid regions based on the AI (Fig. 1a) also matches well with the grasslands of surface vegetation types in China from Moderate Resolution Imaging Spectroradiometer observations (Fig. 1c). In general, the total area of semi-arid regions is 2.14× 106 km2, while the areas of hyper-arid, arid and dry sub-humid regions are 1.06× 106, 1.42× 106 and 1.00× 106 km2, respectively, suggesting that semi-arid regions dominate the coverage of drylands (Fig. 1a). The semi-arid region of northern China is a transitional zone of instability between the inland arid regions in northwestern China and the humid monsoon areas of southeastern China, which is related to strong climate gradients and biological variables, and vegetation growth in these regions is relatively unstable (e.g., Fu and Wen, 2002; Dai et al., 2004). Moreover, the interactions between the East Asian summer monsoon (Huang et al., 2016a), westerlies and mixed agricultural-pastoral activities make these regions highly vulnerable to natural disasters (Shi et al., 1994), climate change and human activities (Fu and Wen, 2002). For example, both the front and rear edges of Transitional Climate Zone (TCZ) exhibit wide year-to-year shifts historically, and the interannual variability of TCZ is mainly affected by P fluctuation while the long-term trend of TCZ is dominated by potential evaporation (Wang et al., 2017). Meanwhile, the most significant expansion and enhanced warming has occurred over the TCZ of semi-arid regions in the last century (Huang et al., 2012; Ji et al., 2014; Guan et al., 2015). The results from (Zhang et al., 2003b) also show that the vegetation-cover change is sensitive to the variation in regional P in the eastern semi-arid regions.
Figure1. Climatological distribution of Chinese drylands for the period 1961-90 using the definition of (a) AI [semi-arid regions (0.2≤ AI<0.5)] and (b) annual P [semi-arid regions (200≤ P<400)] (units: mm). (c) Corresponding surface vegetation types from MODIS observations. [Reprinted from (Li et al., 2015). With permission of Springer Nature.]In the past 60 years, many studies have demonstrated that continuous expansion and significant aridity trends have occurred across semi-arid regions in China (Ma and Fu, 2006; Feng and Fu, 2013; Li et al., 2015; Huang et al., 2016a, Huang et al., 2017a). Figure 2 shows the temporal variations in the areal coverage of semi-arid regions compared with those of hyper-arid, arid, and dry sub-humid regions in China. The rate of expansion for semi-arid regions is the fastest [0.111× 106 km2 (10 yr)-1] during 1948-2008, while the areas of arid and dry sub-humid regions have increased by 0.013× 106 km2 (10 yr)-1 and 0.017× 106 km2 (10 yr)-1, respectively, suggesting that the largest expansion over China during the last 61 years has occurred in the semi-arid regions. Relative to 1948-62, the expansion of semi-arid regions, with relative increases of 33%, is the most severe in drylands, where the areal increase of semi-arid regions is more than 10 times than that in arid and dry sub-humid regions (Table 1). Therefore, the rapid expansion of semi-arid regions has dominated the areal increase in total drylands; in contrast, the area of hyper-arid regions has decreased by 6%, but the reduced amplitude is much less than that of semi-arid expansion. Moreover, (Huang et al., 2016a) indicated that the semi-arid expansion over East Asia has been the largest in a single region and accounts for 50% of the global semi-arid expansion. (Feng and Fu, 2013) suggested that the expansion of semi-arid regions due to climate change is responsible for desertification and has increased the population affected by water scarcity and land degradation. Considering that newly formed semi-arid regions have different features compared with the original type of region, (Li et al., 2015) and (Huang et al., 2016a) compared the spatial shifts in semi-arid regions towards drier/wetter types for the last 15 years relative to the 1961-90 climatology, as shown in Fig. 3. The newly formed semi-arid regions that transformed from sub-humid/humid regions mainly occurred over Northeast China, stretching from western Heilongjiang and Jilin provinces and eastern Inner Mongolia, across most of Hebei and Shanxi provinces, and then towards southern Gansu Province (Fig. 3a). The newly formed dry sub-humid regions are adjacent to the east of the formed semi-arid regions in northeastern China, but the expansion in dry sub-humid regions is less than that in semi-arid regions. Meanwhile, the transformations of arid and hyper-arid regions from other types of regions were minimal; thus, the formed coverage of semi-arid regions primarily contributed to the shifts towards drier types of regions in drylands across China. Contrary to the shifts towards drier types, the retreat of arid to semi-arid regions was mainly located in small areas of the central TP and along the edge of the Junggar Basin. According to the contributions of different types of transitions to the areal change in semi-arid regions in China (Table 2), the areal change in newly formed semi-arid areas (from other types to semi-arid regions, 0.73× 106 km2) is six times larger than that in semi-arid decreasing areas (from semi-arid regions to other types, 0.105× 106 km2). Meanwhile, the areas of semi-arid regions with intensified aridity (drying) are much larger than those with decreased aridity (wetting), indicating that the expansion of semi-arid regions is most severe in northern China, and the newly formed semi-arid regions primarily originated from sub-humid and humid regions; thus, more semi-arid regions become drier, and the aridity has intensified dramatically, especially in the newly formed semi-arid regions. (Huang et al., 2016a) also compared the transformation types over East Asia with those of other regions in the world and found that the semi-arid expansion in East Asia is the largest. Many studies have reported that semi-arid expansion associated with an increase in aridity is the result of global warming (Feng and Fu, 2013; Huang et al., 2017b).
Figure2. Temporal variations in areal coverage of drylands in China for the total and individual components, in units of 106 km2. Thick lines are the 15-year smoothing to emphasize the climate change. Solid lines are the fit to the data of the total and individual component of drylands. [Reprinted from (Li et al., 2015). With permission of Springer Nature.]
Figure3. Changes in coverage to (a) drier types and (b) wetter types for 1994-2008 relative to 1961-90. H, SH, SA, A, HA represent humid, sub-humid, semi-arid, arid, and hyper-arid, respectively. [Reprinted from (Li et al., 2015). With permission of Springer Nature.]
Figure4. Spatial distributions of the linear trends in the (a) AI [(10 yr)-1], (b) P [mm (10 yr)-1], (c) PET [mm (10 yr)-1], and (d) SAT [°C (10 yr)-1] in semi-arid regions in China from 1948 to 2008. [Reprinted from (Huang et al., 2016a). With permission of Springer Nature.]Recently, some studies have shown that global warming is nonuniform, and the warming of semi-arid regions over the last century has been more significant (Huang et al., 2012; Ji et al., 2014; Guan et al., 2015), especially during the cold season (November to March). The contribution of semi-arid regions in Northern Hemisphere (NH) midlatitudes to continental warming reaches approximately 50% (Huang et al., 2012; Ji et al., 2014); in particular, semi-arid warming in Asia is most remarkable. Similarly, many studies have found that climate warming in the past 100 years is evident in Northeast, North and Northwest China, which are mainly covered by semi-arid regions (IPCC, 2007; Tang et al., 2010). (Li et al., 2015) also showed that warming has enhanced in semi-arid regions in China, especially in northeastern China (Fig. 4b). (Guan et al., 2015) investigated the driving factors behind the enhanced semi-arid warming (ESAW) in East Asia by using a recently developed dynamical adjustment method; the results implied that the ESAW over East Asia is induced by regional anthropogenic forcing, which may be closely related to local human activities (Fig. 5). Moreover, (Zhou et al., 2016) reproduced the ESAW in historical simulations using anthropogenic and natural forcings, but the ESAW disappeared when only natural forcing was considered, suggesting new potential footprints for anthropogenic warming. On the other hand, some studies consider that the increase in water vapor is the main reason for desert amplification (Zhou, 2016), which further demonstrates that the water vapor in the semi-arid regions is likely controlled by water vapor originated from ocean or sea surface temperature (SST). However, in phase 5 of the Coupled Model Intercomparison Project (CMIP5) simulations, the SAT trend is underestimated in semi-arid areas and overestimated in humid areas, indicating that the CMIP5 simulations have large uncertainties in their response of SAT to global warming (Huang et al., 2016b). In addition, the ESAW may accelerate in the context of a warming climate and, thus, has important societal and economic consequences (Zhou et al., 2016).
Figure5. Regionally averaged cold-season mean-temperature trend as a function of annual P for raw (black), dynamically induced (blue), and radiatively forced (red) temperatures (units: °C yr-1) in the cold season from 1902 to 2011 over East Asia. [Reprinted from (Guan et al., 2015).]
Figure6. Schematic diagram of local thermodynamic processes in drylands and humid lands. [Reprinted from (Huang et al., 2017b).In terms of the water cycle, semi-arid regions have become drier in northeastern China during recent years, accompanied by soil drying caused by decreasing P and enhanced warming. At the same time, soil suction has increased with drying soil, and the remaining soil moisture has become less absorbable for plant roots; thus, evapotranspiration and P may be reduced, leading to an increase in SH and temperature. Moreover, the increased temperature can further decrease the soil moisture and P, which could enhance the drying trends and make the regions even drier. The soil water loss, decreased evapotranspiration, and increased temperature form a positive feedback loop, which can continue until the soil is completely dry, resulting in desertification (Seneviratne et al., 2010). Furthermore, soil moisture-carbon-temperature feedbacks also play an important role in local land-atmosphere interactions and the carbon cycle. Excluding the recipients of climate changes induced by increasing carbon dioxide (CO2) (Huang et al., 2016b), semi-arid regions are the largest contributors to the trends and interannual variabilities in the terrestrial carbon sink (Poulter et al., 2014). Semi-arid expansion against the background of global warming reduces the capacity of soil organic carbon storage; thus, soils store less carbon and emit more CO2 into the atmosphere, which aggravates warming and causes a positive feedback cycle, where drying and warming both intensify (Huang et al., 2017a, b). However, these feedbacks are difficult to constrain in a defined area due to their impacts on each other and are still less documented in semi-arid regions in China. To improve our understanding of the role of land-atmosphere interactions over semi-arid regions in China, the temporal- and spatial-scale effects, as well as multiple feedback processes, need to be considered.
(Wu et al., 2009) indicated that ocean-atmosphere forcing contributes to the formation of climate systems, including deserts and monsoons, in the summertime subtropics. Many studies have shown that oceanic oscillations, as factors in ocean-atmosphere feedbacks, modulate the decadal climate variations in semi-arid regions across China (e.g., Ma and Fu, 2006; Wang et al., 2008; Qian and Zhou, 2014). For example, (Yang et al., 2017b) suggested that the Pacific SST pattern plays a dominant role in the anomalies of annual P and associated atmospheric circulation over eastern China, whereas the Atlantic SSTs contribute to a lesser degree. During the warm phase of the Pacific SST pattern, the semi-arid regions in northeastern China are prone to drought by inducing northwesterly wind anomalies and pushing the monsoons to the south (Ma and Shao, 2006; Ma, 2007), which explains the robust drying trend in northeastern China (Ma and Fu, 2003; Ma and Dan, 2005). (Yang et al., 2017c) indicated that the Pacific Decadal Oscillation (PDO) affects East China P patterns by modulating the large-scale circulation pattern on both the interannual and intraseasonal scales. (Lin et al., 2017) found that the negative phase of the PDO is associated with strong southerly summer monsoons, which are favorable for increasing the occurrence of local extreme rainfall over North China. Furthermore, the interdecadal variability in the PDO can exert a modulating effect on El Ni?o-Southern Oscillation (ENSO) teleconnections over East Asia (Wang et al., 2014; Dong and Dai, 2015). In semi-arid regions over China, ENSO-induced drying trends are magnified when ENSO is in phase with the PDO, while the dry variations weaken or even disappear when ENSO is out of phase with the PDO (Wang et al., 2014). Moreover, the response of the Hadley cell expansion to ocean-atmosphere feedbacks during recent decades has been responsible for enhanced aridity and water shortage changes in semi-arid regions of the NH (Seager et al., 2007).
Meanwhile, disturbances in monsoons and westerly winds are also important factors in semi-arid climate change in China (Qian et al., 2009), as they are closely associated with the land-sea thermal contrast. As suggested by (Wallace et al., 1995) and (He et al., 2014), the land-sea thermal contrast can induce feedbacks between ocean and atmospheric circulations. Both the zonal and meridional land-sea thermal contrast contribute to inducing an abrupt shift in thermally forced atmospheric circulations, such as asymmetric zonal and meridional thermal forcing (ZTF and MTF, respectively), which affects the intensity of westerlies, planetary waves, monsoons, and blocking frequencies and further affects the climate changes over NH continents, including the semi-arid climate change in China (He et al., 2014; Huang et al., 2017a, c). Semi-arid regions are largely influenced by blocking, especially in winter. (He et al., 2014) found that the ESAW is accompanied by a decrease in blocking frequencies. As shown in Fig. 7, when surface thermal forcing in the NH follows the cold-ocean-warm-land pattern due to the CO2 greenhouse effect, the ZTF decreases and reduces the blocking frequencies by modulating planetary waves. In contrast, the westerly wind is weakened, and topographic forcing is enhanced when surface forcing follows the warm-ocean-cold-land pattern, which is favorable for the persistence of blocking and cooling. (Huang et al., 2017c) indicated that the internal climate variability associated with oceanic oscillations and Arctic amplifications, by altering the ZTF and MTF, can enable a decadal modulated oscillation (DMO), which influences the terrestrial temperature in semi-arid regions on decadal to multidecadal time scales. The upward DMO contributes to accelerated warming, similar to that during the last 20 years of the 20th century, while the downward DMO suppresses the long-term warming trend, which resulted in a cooling trend over Eurasia during the recent warming hiatus period.
Figure7. Composite SAT anomaly fields (units: °C) in winter (December-February) for (a) positive and (b) negative phases of land-sea thermal contrast. Contour intervals: 0.5 °C. [Reprinted from (He et al., 2014).]Based on satellite observations, ground-based measurements, and model simulations (Huang et al., 2007; Wang et al., 2010a; Bi et al., 2011; Chen et al., 2013), the large amounts of dust particles in East Asia emitted from arid and semi-arid regions are also considered significant factors affecting semi-arid climate change over China, through their direct effects on solar and thermal radiation (e.g. Li, 2004; Huang et al., 2014) and indirect/semi-direct effects on clouds and P (e.g., Twomey, 1977; Albrecht, 1989; Huang et al., 2006a, b, c, 2009). East Asia is one of the major sources of dust aerosols in the NH, and dust from deserts in China increased most obviously from 1970 to 2000 but has generally decreased since 2000, with oscillations in recent years (e.g., Xia et al., 2016). (Huang et al., 2008) indicated that dust in China often originates from the Taklimakan and Gobi deserts, with more frequent dust storms occurring in the Taklimakan Desert than the Gobi Desert, which can be transported long distances by prevailing westerlies (Fig. 8). Meanwhile, because of the unique TP topography, with dynamical and thermal forcing, dust aerosols from the Taklimakan Desert compared to African dust emissions could be transported to the TP and even eastern China via upper tropospheric westerly jets (Huang et al., 2007, Huang et al., 2008; Ge et al., 2014), which could alter atmospheric stability and circulation (Huang et al., 2007; Zhao et al., 2015; Li et al., 2016b). Moreover, dust outbreaks are sensitive to changes in land-surface conditions, such as moisture and vegetation growth, over East Asia (Kurosaki et al., 2011).
Figure8. Dust event vertical structure from CALIPSO observations over the Taklamakan Desert (80°E), Gobi Desert (102°E), eastern China (124°E), Yellow Sea (146°E), and the Pacific Ocean (168°E) superimposed on a regional map. Coloring on the right represents the dust event frequency of occurrence in percent. [Reprinted from (Huang et al., 2008).]In addition, dust aerosols in East Asia have an important impact on the microphysical properties and formation of clouds, by acting as cloud condensation nuclei and ice nuclei (first/second indirect effect and invigoration effect) and altering the relative humidity and stability of the atmosphere (semidirect effect) (Li et al., 2010, Li et al., 2011). Moreover, the effects of dust also interact with cloud properties to alter the atmospheric dynamics, surface energy budget, and hydrological cycle, which further influences semi-arid climate change. Huang et al. (2006a, b, 2010) compared cloud properties under dusty and dust-free conditions over semi-arid regions in Asia and found that dust aerosols decreased the ice-cloud effective particle size and optical depth of cirrus clouds (Kawamoto et al., 2004; Su et al., 2008; Wang et al., 2010b). Dust aerosols over Asian semi-arid regions have a near warming effect at the top of the atmosphere but a cooling effect at the surface, while a warming effect occurs within the atmosphere (Huang et al., 2014). Furthermore, dust-cloud interactions can enhance or suppress semi-arid P due to the spatial distributions of dust aerosols, TP topography, cloud types, local humidity, and atmospheric circulation. (Han et al., 2013) indicated using a climate-chemistry-aerosol model that dust radiative forcing decreased the P over the Yangtze River region and large areas of northeastern China. There is a positive feedback loop in the dust aerosol-cloud-P interactions over semi-arid regions (Huang et al., 2014, Huang et al., 2017a); for instance, a decrease in rainfall and subsequent deficit in soil moisture contribute to increasing dust storms, which further reduces the low-cloud cover and water vapor and increases the high-cloud cover, resulting in suppressed P and increased dust storms. Therefore, more dust storms occur in arid and semi-arid regions under feedback conditions, which further contributes to the desertification observed in recent decades and accelerated aridity over semi-arid regions in Asia (Huang et al., 2014).
With rapid economic development, more GHGs and pollutants, such as aerosols, have been emitted, which are key forcing agents associated with human activities (Huang et al., 2017a). Desertification, which is the loss of biodiversity and other forms of environmental deterioration, is partially the result of human activities (Huang et al., 2016a, b). Many studies have indicated that human activities, such as increased CO2 emissions, anthropogenic aerosols, land use and urbanization, play an important role in semi-arid climate change over China (e.g., Dai, 2013; Guan et al., 2015; Lin et al., 2015; Fu et al., 2016; Huang et al., 2017a; Xu and Yang, 2017; Zhang et al., 2017). For instance, under high CO2 emissions, projected drying trends exist over semi-arid regions during the 21st century (Feng and Fu, 2013; Huang et al., 2016b). (Zhao et al., 2017) found that the increase in anthropogenic aerosols in the atmosphere enhanced the terrestrial aridity and resulted in semi-arid expansion over East Asia. Lin et al. (2015, 2016) examined the changes in AI forced by CO2 and aerosols over China using Community Earth System Model simulations and found that GHGs and aerosols can significantly affect regional AI, and the AI response to CO2 decreased (drying) in almost all semi-arid regions in China, while the AI response to aerosols (black carbon and sulfate) increased (wetting) to the west and northeast but decreased (drying) in central areas across semi-arid regions of China (Fig. 9). (Li et al., 2016a) investigated the effects of anthropogenic aerosols on temperature variability in semi-arid regions over China using CMIP5 models and found that anthropogenic aerosols reduced the temperature in the region. Moreover, (Zhao et al., 2015) found that dust aerosols can cause surface cooling and reduced P over semi-arid regions in the NH, based on aerosol-climate coupled simulations. (Guan et al., 2016) found that population density/change was correlated with anthropogenic dust aerosols in the semi-arid regions of East China, which further aggravated drought, indicating that human activities affect semi-arid climate change (Huang et al., 2014, 2015).
Figure9. Changes (units: % °C-1) in P, PET and AI due to black carbon (BC) and sulfate (SO4) aerosols (left-hand and middle panels) and CO2 (right-hand panels), scaled by global-mean SAT changes. Gray regions do not have statistically significant changes based on a 95% confidence interval from a two-sided t-test. [Reprinted from (Lin et al., 2016).]In addition, natural and human activities have modified land-use and cover change (LUCC), which interacts with the environment, has significant effects on the ecosystems of semi-arid regions and, consequently, directly or indirectly exerts significant influence on climate change (Deng et al., 2013). Negative LUCC may dry the local climate, leading to degradation, especially in semi-arid regions (Fu and Wei, 1993; Fu et al., 2016). The land cover degradation in northern China has accelerated in the last five decades, and the remaining grassland area in Inner Mongolia has significantly degraded (Wang et al., 2004), which could lead to a decrease in P and an increase in surface temperature in semi-arid regions by altering various physical characteristics of the land surface, such as albedo, surface roughness, soil water and thermal variables (Zhang et al., 2005; Zhang et al., 2009; Yu and Xie, 2013). Using a regional climate model, (Chen et al., 2017) indicated that net radiation and evaporation reduced within degraded areas after simulating vegetation degradation over the semi-arid regions of China. (Li et al., 2017) indicated that the observed warming trend over semi-arid regions in northern China during 1946-2005 could be largely attributable to anthropogenic forcings, while the internal variability is still larger than the forced variation in semi-arid P. To better understand and quantify the contributions of human activities to semi-arid climate change, additional studies are needed.
Figure10. Area changes in semi-arid and other subtype regions of drylands in East Asia from the corrected CMIP5 and RCP8.5 relative to the baseline (1961-90) for the future period of 2096-2100. The gray shading denotes the baseline drylands present in 1961-90. Changes include any transition from adjacent and nonadjacent subtypes. [Reprinted from (Huang et al., 2016b)]During the process of climate change, land-atmosphere interactions and human activities can influence the intensity of the regional temperature response in semi-arid regions, in which the effect of anthropogenic forcing is greater than that of natural forcing. Ocean-atmosphere interactions may modulate decadal climate change in semi-arid regions by altering the intensities of westerlies, monsoons, planetary waves, and blocking frequencies. Moreover, dust-cloud-P interactions have a significant impact on P by affecting the local energy and hydrological cycles in semi-arid regions across China. However, current findings indicate that CMIP5 simulations do not perform well in simulating semi-arid climate change. After correcting the CMIP5 simulations based on historical data, semi-arid regions in China are projected to continuously expand in the near future under a high emissions scenario, which will increase the risk of land degradation and water scarcity.
Although significant progress has been made in recent studies on semi-arid climate change and its dynamics over China, many unsolved issues remain to be further investigated in future studies. For instance:
(1) There is still a lack of high-quality datasets for studying and forecasting semi-arid climate change over China. Future studies should combine various datasets based on ground-based measurements, satellite observations, and climate reconstructions to study semi-arid climate change on multiple time scales and further compare the regional characteristics and correlations within different semi-arid regions across China and other regions of the world.
(2) Previous studies have mainly focused on the regional means of climatic factors in semi-arid regions across China, but few studies have focused on typical underlying surfaces in different semi-arid regions to examine climate change and dynamics in these regions. Meanwhile, the semi-arid regions across China are some of the major semi-arid regions in East Asia, and further studies need to analyze typical semi-arid climate change in China.
(3) The mechanisms of semi-arid climate change are still not well understood, and there is still a lack of robust research on the mechanisms and numerical simulations of semi-arid climate change in the context of global change. Moreover, the attributions of human-induced perturbations and natural forcings to semi-arid climate change should be further quantified and evaluated.
(4) Climate models have not performed well in the simulation and prediction of semi-arid climate change across China; thus, the ability to simulate aridity and climate change in semi-arid regions should be further improved in future climate models.
