The Oxford Research Encyclopedia of Climate Science will be available via subscription on April 26. Visit About to learn more, meet the editorial board, or recommend to your librarian.

Dismiss
Show Summary Details

Page of

date: 25 April 2018

Modeling of Regional Climate over the Tibetan Plateau

Summary and Keywords

The modeling of climate over the Tibetan Plateau (TP) started with the introduction of Global Climate Models (GCMs) in the 1950s. Since then, GCMs have been developed to simulate atmospheric dynamics and eventually the climate system. As the highest and widest international plateau, the strong orographic forcing caused by the TP and its impact on general circulation rather than regional climate was initially the focus. Later, with growing awareness of the incapability of GCMs to depict regional or local-scale atmospheric processes over the heterogeneous ground, coupled with the importance of this information for local decision-making, regional climate models (RCMs) were established in the 1970s. Dynamic and thermodynamic influences of the TP on the East and South Asia summer monsoon have since been widely investigated by model. Besides the heterogeneity in topography, impacts of land cover heterogeneity and change on regional climate were widely modeled through sensitivity experiments.

In recent decades, the TP has experienced a greater warming than the global average and those for similar latitudes. GCMs project a global pattern where the wet gets wetter and the dry gets drier. The climate regime over the TP covers the extreme arid regions from the northwest to the semi-humid region in the southeast. The increased warming over the TP compared to the global average raises a number of questions. What are the regional dryness/wetness changes over the TP? What is the mechanism of the responses of regional changes to global warming? To answer these questions, several dynamical downscaling models (DDMs) using RCMs focusing on the TP have recently been conducted and high-resolution data sets generated. All DDM studies demonstrated that this process-based approach, despite its limitations, can improve understandings of the processes that lead to precipitation on the TP. Observation and global land data assimilation systems both present more wetting in the northwestern arid/semi-arid regions than the southeastern humid/semi-humid regions. The DDM was found to better capture the observed elevation dependent warming over the TP. In addition, the long-term high-resolution climate simulation was found to better capture the spatial pattern of precipitation and P-E (precipitation minus evapotranspiration) changes than the best available global reanalysis. This facilitates new and substantial findings regarding the role of dynamical, thermodynamics, and transient eddies in P-E changes reflected in observed changes in major river basins fed by runoff from the TP. The DDM was found to add value regarding snowfall retrieval, precipitation frequency, and orographic precipitation.

Although these advantages in the DDM over the TP are evidenced, there are unavoidable facts to be aware of. Firstly, there are still many discrepancies that exist in the up-to-date models. Any uncertainty in the model’s physics or in the land information from remote sensing and the forcing could result in uncertainties in simulation results. Secondly, the question remains of what is the appropriate resolution for resolving the TP’s heterogeneity. Thirdly, it is a challenge to include human activities in the climate models, although this is deemed necessary for future earth science. All-embracing further efforts are expected to improve regional climate models over the TP.

Introduction

The Tibetan Plateau (TP), being the world’s highest and most extensive plateau, is key to the Third Pole Region. The TP holds the largest amount of glacially stored freshwater outside the Arctic and Antarctica. It is the source of major Asian river systems (Su et al., 2016), and thus vital to nearly one-third of the world’s population in terms of freshwater supplies. Over recent years, the region has gained growing attention because of its significant role in the global climate system and its high sensitivity to global warming. Indeed, its interaction with the Indian monsoon has been identified as one of the plausible “tipping points” of the global climate system (Lenton et al., 2008). Within the Asian monsoon system, it is well recognized that the Himalayas in the TP exert a profound thermal and dynamical influence on the atmospheric circulation, involving a strong interaction between the land and the atmosphere (e.g., Wu et al., 2015).

In principle, the modeling of TP’s climate started with the introduction of Global Climate Models (GCMs) in climate research. Connecting computer technology with meteorology began in 1955 when IBM launched its first machine. Since then, GCMs have been developed to simulate the atmospheric dynamics (Phillips, 1956) and the climate system (Manabe et al., 1965). As an important part of global territory, regional climate over the TP was naturally modeled by GCMs. However, the focus was initially on the strong orographic forcing caused by the TP and its impact on general circulation (e.g., Kasahara & Washington, 1971) rather than regional climate. The need to better understand and predict natural and human induced climate changes, especially the work by the Intergovernmental Panel on Climate Change (IPCC), has promoted the development of GCMs into their current levels of sophistication (Cubasch et al., 2013). As for any other region in the world, the GCM is the most important, effective, and powerful tool for process understanding, simulations of the past, and projection for the future of the TP (e.g., Chen et al., 2015).

The TP possesses a high heterogeneity in terrain and land surface characteristics (Figure 1), which means that even the latest GCMs used by the IPCC do not have the spatial resolution needed to properly treat the complex surface heterogeneity and critical processes such as glacier/snow cover–related albedo feedbacks (Su et al., 2013). In addition, the formulation of adaptation measures in response to climate change requires information at finer spatial scales (Gao et al., 2008, 2011a). More importantly, however, there can be many important processes operating at scales smaller than those resolved by GCMs which cause fine-scale climate variability, especially for precipitation (Chen et al., 2016). Accurate estimates of regional scale climate over the TP require downscaling of the large-scale information on climate changes with help of a Regional Climate Model (RCM) in which GCM outputs are used as initial and lateral boundary conditions (Betts et al., 1996; Entekhabi et al., 1996). Dynamic downscalings (DDMs) have been the main tool for detailed regional climate information, although very fine resolution GCMs (e.g., Guo & Wang, 2016) provide additional hope for the future.

Click to view larger

Figure 1. Domain and topography of the Tibetan Plateau with 83 meteorological stations and typical geographical units.

When regional climate modeling began, researchers focused on understanding the impact of the TP upon the Asian summer monsoon (ASM), including both the Indian and East Asian summer monsoons. The majority of simulations at that time were sensitivity type studies, focusing on the TP’s elevation impact on the Asian monsoons (e.g., Song et al., 2010). The Asian monsoons involve the interactions among many components of the climate system, and fundamental physical processes in the models must therefore be well parameterized. Many of the discussions highlighted different key physical processes and parameters, including land surface processes, atmospheric boundary layer processes, atmospheric heating, precipitation, and ocean data (e.g., Bollasina & Benedict, 2004).

Besides the above-mentioned process studies, the IPCC process (IPCC, 2013) helped maintain high momentum in the community in simulating past and future climate over the TP by using GCMs and RCMs. In particular, the Coupled Model Intercomparison Project (CMIP) and Coordinated Regional Climate Downscaling Experiment (CORDEX), organized by the World Climate Research Programme (WCRP), have played important roles. These two projects provide a community-based infrastructure in support of climate model diagnosis, validation, intercomparison, documentation, and data access. With the release of IPCC AR5, more than 20 GCMs from CMIP5 (Taylor et al., 2012) have been used to study historical and future climate over the TP (Su et al., 2013). Moreover, the potential of the CORDEX simulations has recently been realized and used.

GCMs used by CMIP and RCMs by CORDEX have both been employed for regional climate simulations. In analyzing these simulations, historic simulations not only serve as the baseline to be contrasted with the future projections, but also act as an important step in model verification when compared with observations. Historic simulation by a RCM can be driven by either a reanalysis or a GCM. While a dynamic downscaling using a RCM driven by a reanalysis can be directly compared with historical observations, the climatology from a RCM driven by a GCM can be compared with the RCM simulation which is driven by the reanalysis, and indicates biases in the RCM passed over from the forcing.

Reanalysis typically employs a GCM to periodically ingest observations, with the results being a gridded set of model dependent variables that are consistent with both model dynamics and the information represented in the observations. Such data-assimilation systems are used to define the long-term analysis of atmospheric fields pertaining to past and recent climate, which can be used to provide initial and boundary conditions for RCM simulations. Examples of reanalysis datasets include ERA40 and ERA-Interim by ECMWF (Uppala et al., 2005; Dee et al., 2011) and NCEP/NCAR and NCEP/DOE by NCEP (Kalnay et al., 1996; Kanamitsu et al., 2002), as well as the JRA-25 (Onogi et al., 2007) reanalysis by the Japan Meteorological Agency. While reanalysis can often be thought of as the best estimate for many atmospheric variables (such as winds and temperature), its usage in complex terrains such as the TP must be approached with caution. The degradation, replacement, or changes of instruments (e.g., satellites), as well as changes in methods of observation (e.g., surface, aloft), may inject error (Trenberth et al., 2001). Further, not all reanalysis variables are constrained by observation: some data types, such as precipitation (depending on the reanalysis) and evapotranspiration (for which global observations simply do not exist), are obtained by running (presumably newer) GCMs. A well-known limitation of reanalyses is that they do not conserve moisture (Nigam & Ruiz-Barradas, 2006).

While many GCMs exist, only a few long-term RCM simulations covering the TP have been published (Yu et al., 2013; Ji & Kang, 2013; Gao et al., 2015a, 2015b). The first two were used to evaluate the climate changes over the whole of China. Although the TP is also included in these and several other simulations (e.g., Gao et al., 2013), no specific analysis of the simulations was done for the TP.

This article aims to summarize modeling efforts made so far with regard to the regional climate over the TP. There was little station in the western TP, and the statistical methods were also difficult to implement. Therefore, statistical methods to achieve downscaling for the TP (e.g., Zhu et al., 2013) are not discussed. Before describing GCM and RCM simulations for the past and future regional climates over the TP, the existing sensitivity studies relating to the impact of surface conditions on the regional climate in the TP will be described.

Modeling the Climate Impacts of Land Surface Dynamics

Sensitivity to Surface Properties

Changes in land surface conditions can influence weather pattern and regional climate. GCMs and RCMs serve as useful tools for evaluating the impact of historical or future landscape changes on climate.

For vegetation surfaces, albedo and Leaf Area Index (LAI) are probably the two most important surface factors for the regional climate. Increased LAI (greening) would reduce surface albedo, which would consequently increase incoming solar radiation to the surface and cause surface and near-surface air temperature increase. On the other hand, the greening could enhance evapotranspiration, which contributes to slowing down the temperature rise. Therefore, the net climate impact of greening depends on the battle of these two mechanisms. Land cover changes can result in changes in radiative, thermal, and dynamic characteristics of the surface, which can alter surface energy balance and atmospheric circulations. As an example, Li and Xue (2010) tested two existing land cover maps derived from satellite with very different conditions over the TP in a GCM, one with bare ground and one with vegetation cover. They found that land cover change from vegetated land to bare ground decreases the radiation absorbed by the surface, leading to lower atmospheric temperature as well as weaker vertical ascending motion, low-layer cyclonic, upper-level anticyclonic, and summer monsoon circulation. The changes in the atmospheric circulation caused a precipitation decrease in the southeastern TP in the model.

On the other hand, Shen et al. (2015) claimed that increased growing season vegetation activity over the TP in recent decades may have attenuated surface warming based on observational data and modeling with a RCM (WRF). This negative feedback on growing season vegetation temperature is attributed to enhanced evapotranspiration.

Another RCM (RegCM4) was used to investigate the climate effects of future land use change over China, based on the land cover data recommended by CMIP5 (Hua et al., 2015). Two 15-year simulations for the period 2036–2050, one with the current land-use data and the other with future land-use scenario (2050), were conducted. It is noted that future land use and cover change (LUCC) in China is mainly characterized by the transition from grassland to forest. Results suggest that the magnitudes and ranges of the changes in temperature and precipitation caused by future LUCC show evident seasonality, which are more prominent in summer and autumn. Further, significant responses of climate to future LUCC mainly occurs in Northeast China, North China, the Hetao area, the Eastern Qinghai-Tibetan Plateau, and South China. Further investigation shows that future LUCC can also produce significant impacts on atmospheric circulation. LUCC would result in an abnormal southwesterly wind over extensive areas from the Indian peninsula to the coasts of the South China Sea and South China through the Bay of Bengal. Furthermore, the Indian tropical southwest monsoons and South Sea southwest monsoons would become stronger, and the abnormal water vapor convergence from the South China Sea and the Indian Ocean could result in more precipitation in South China.

Relative Importance of GCM Forcing and Land Surface Processes

The accuracy of RCM simulation depends on the physical processes involved as well as the forcing data sets derived from GCMs. Given the complexity of the TP surface, and to narrow the focus, the relative role of the land surface schemes and forcing data sets in the RCM over the TP were investigated by comparing three Weather Research and Forecasting (WRF) model dynamical downscaling simulations (Gao et al., 2016). Three WRF simulations were configured with two land surface schemes (Noah versus Noah with multi-parameterization (Noah-MP)) and two forcing datasets (ERA-Interim and CCSM4) performed for the period 1980 to 2005. The downscaled temperature and precipitation were evaluated with observations and inter-compared regarding temporal trends, spatial distributions, and climatology. Results show that the temperature and precipitation temporal trends are determined by the forcing data sets. Therefore, choosing a GCM with a small bias and realistic trend or adopting some method to reduce the bias and error in the trends is beneficial and may be necessary before downscaling.

Relative to the forcing data sets, land surface processes play a more critical role in the DDM over the TP. Common cold and wet biases exist in all three simulations regardless of the type of large-scale forcing or land surface model being used. This is especially true in the western TP during the cold season, indicating that the WRF model is still deficient in capturing cold-season processes at high elevations. However, such cold and wet biases in the RCM were greatly constrained by the relative finer topographical features compared to these in forcing GCM. During the wet season, land surface models seem to have a greater impact on simulated air temperature and precipitation than the large-scale forcing. By changing the vertical profiles of temperature in the atmosphere and the horizontal patterns of moisture advection during the monsoon seasons, the land surface schemes significantly regulate the downscaled temperature and precipitation in terms of climatology and spatial patterns.

Although land surface models (LSMs) are important in DDM, modern LSMs showed low applicability in the TP. For instance, significant underestimation of the topsoil moisture in the central TP was simulated using three commonly used modern land surface models—CLM, SIB2, and Noah LSM (Yang et al., 2005). Many efforts were made to mitigate discrepancy, for instance, considering vertical heterogeneity in soil texture and adding organic matter (Yang et al., 2009; Chen et al., 2012). Gao et al. (2015) found the same synoptic utilizing a new-generation LSM—Noah-MP with multiple improvements in model dynamics and a multi-parameterization framework (CTL in Figure 2). They further found that representing vertical heterogeneity in soil texture (LAY) and soil organic matter (SOM) does not reduce the underestimation of topsoil moisture. Instead, they noted the unique vegetation characteristics over the TP: the vegetation is sparse and short, and consequently roots grow in shallow soils. Hence, a new root parameterization (ROOT in Figure 2) for the vegetation over the TP was developed and greatly reduced the overestimation (Gao et al., 2015).

Click to view larger

Figure 2. Variations of the daily mean soil liquid water content (SLW in %) 5cm soil depth for 26 May–16 Sept. 1998 for the observation (OBS) and four simulations (CTL: control run; LAY: layered soil texture; SOM: adding soil organic matter; Root: root parameterization. (From Gao et al., 2015c)

Simulation of the Historical Climate

GCM Simulations

IPCC AR5 recognizes that global models have significant problems for the TP, most likely due to the difficulty of simulating the effects of the dramatic topographic relief, as well as the distorted albedo feedbacks due to extensive snow cover (Su et al., 2013). Considerable spread among the rainfall simulated by the GCMs was found in CMIP3, particularly at the windward side of mountains, e.g., the steep slope of the TP (Kim et al., 2008). Recent studies based on CMIP5 confirm these biases over the Tibet. For the surface air temperature, Figure 9.39 in the IPCC’s AR5 report (IPCC, 2013) shows that the largest cold and wet biases among the global subregions are in the TP, especially during winter. Significant wet bias exists in regions near extensive mountain ranges (Kim et al., 2008). Su et al. (2013) found that most GCMs reasonably capture the climatological patterns and spatial variations of the observed climate. However, most models have cold biases (Figure 3a), with a mean underestimation of 1.18°C–2.58°C for the months December–May, and less than 18°C for June–October. Over the TP, the simulations of all models overestimate precipitation in climatological annual mean biases by 62.0%–183.0% (Figure 3b), and only half of the 24 GCMs can reproduce the observed seasonal pattern, demonstrating the critical need to improve precipitation-related processes in these models.

Click to view larger

Figure 3. Area-averaged (a) annual bias of surface air temperature and (b) relative bias of precipitation relative to the observation for each GCM during 1961–2005 over the entire eastern TP. (From Su et al., 2013)

The large bias associated with precipitation intensity and patterns remains despite the high resolution and inclusion of the indirect effects of sulfate aerosol that have helped to improve the models’ skill in simulating the annual precipitation cycle in GCMs (Duan et al., 2013). Only a few models reproduce the observed seesaw pattern associated with the inter-annual variability of the summer monsoons over the TP. Regarding long-term trends, most models overestimate the amplitude of the tropospheric warming and the declining trend in the surface heat low between 1979 and 2005 (Duan et al., 2013). In addition, the observed cooling trend in the upper troposphere and the declining strength of the Tibetan high pressure system were not reproduced by most models.

As for the upper-level atmospheric elements, quite large biases appear at the reliefs surrounding the TP and the simulated atmospheric variables at 200 hPa and 500 hPa (wind components at 200 hPa (U200 and V200), air temperature (T500), and geopotential height (Z500) at 500 hPa, specific humidity at 600 hPa (Q600), sea level pressure (SLP)) show remarkable differences among models (Xu et al., 2016). Spatial patterns for wind components, especially for V200, are not well simulated by all the GCMs. Distinct seasonal features in V200 are detected with overestimation in June–August (JJA) over the TP, and underestimation in the southern TP in December–February (DJF). Remarkable cold biases are detected in T500 in most GCMs, especially in DJF, which led to the underestimations in Z500 andU200 in the North, and overestimation in the South. Q600 is underestimated in the TP and pronouncedly underestimated in the southeast of the TP in JJA.

Simulated LAI in CMIPs was evaluated over the TP by comparing with remote sensing (Bao et al., 2015). It was found that most of the models tend to overestimate the satellite LAI magnitude, though LAI patterns generated by the models strongly agree with the references. The wet bias found in most models and overestimations of photosynthesis as well as the bias of satellite data are considered plausible reasons for the overestimation of simulated LAI in most of the models. The model simulations capture the observed increasing trend of LAI over most of the TP during the period 1986–2005; however, the decreasing trend around the headstream of the Yellow River is not detected due to the coarse resolution of the GCMs.

Quality of Reanalysis Data

Reanalysis presents a better surface air temperature mean over the eastern TP than GCMs because of the assimilation process, although the significant cold bias still exists in the reanalysis, especially in the cold season (Gao et al., 2014). This indicates that the assimilation could not conquer the model’s dynamic and physics discrepancies. Although ERA-Interim best resolves the spatial pattern of the GLDAS ensemble mean runoff climatology and temporal trends among the four widely used reanalysis datasets (NCEP-NCAR, NCEP-DOE, ERA-40, and ERA-Interim), which have the closest domain-averaged mean and the highest correlation, the significant wet biases still exist in ERA-Interim. NCEP/NCAR even presents an opposite trend in contrast to ground-based observation and other reanalyses in the TP. Therefore, there is still significant scope for improving reanalyses over the TP.

RCM Simulations

The significant biases in the GCMs and reanalyses over the TP point to problems in the representation of dynamics and physics in the models. Although these biases can be passed over to RCMs, downscaling with RCMs can be expected to improve GCM simulations and add value by increasing spatial resolution. The existing RCM simulations, to a very large extent, confirm that RCMs do a better job than GCMs, which serves as a forcing for RCM (Gao et al., 2003, 2015a, 2015b; Li et al., 2016). This is especially true in the western TP during the cold season. Indeed, RCM simulations have been shown to have closer agreement with the observation than that of the driving GCM in terms of both spatial pattern and precipitation amount (Gao et al., 2003, 2015a). All these factors indicate that RCMs do in principle add value to simulations over the TP, although they still have their deficiencies (e.g., problem in capturing cold-season processes at high elevations).

What follows will focus on how a RCM (WRF) simulates near-surface air temperature, precipitation, and water balance over the TP in comparison to its host GCM and reanalysis (Gao et al., 2014, 2015a, 2015b, 2016; Maussion et al., 2014; Xu et al., 2016; Xiao et al., 2016). Simulation was initialized at 0000 UTC January 1, 1979, and ended at 2300 UTC December 31, 2011. The initial lateral boundary conditions and SST were interpolated from the ERA-interim reanalysis (1979–present; Dee & Uppala, 2009; Dee et al., 2011), which has been proven best among the available reanalysis products in describing temperature and water cycle over the TP (Gao et al., 2014). WRF was set up by 30km horizontal grid spacing with 159×196 grids cells, covering nearly the whole Asian continent following Exp. 6 in Gao et al. (2011b). The Single-Moment 3-class, Grell-Devenyi ensemble scheme, the Yonsei University scheme was used for the PBL parameterization. The LSM used here is the Noah LSM 4-layer soil temperature and moisture model with frozen soil and snow-cover prediction.

Gridded and 83 stations surface air temperature and precipitation observations provided by the National Climate Center, China Meteorological Administration (CMA), are used as references. Comparison is conducted on monthly mean surface air temperature and precipitation. There is no observation for evapotranspiration; the Global Land Data Assimilation Systems (GLDAS) (Rodell et al., 2004) was adapted as a reference. Reanalyses are involved in the comparison of the water balance. P-E for the reanalyses are calculated by not only using the terrestrial precipitation and evapotranspiration, but also the vertically integrated moisture flux convergence using six-hourly specific humidity and wind data at standard pressure levels.

Observed annual surface air temperature averaged across the TP showed abrupt changes around 1998 (Figure 2 in Gao et al., 2014). In the following analysis, 1998 is treated as a pivotal year, and changes in moisture before and after 1998, regarded as the moisture trend over the TP in 1979–2011, are explored. Climatology in the TP is split into dry season (October–April) and wet season (May–September) according to the typical seasonal variation of precipitation in the region (Yao et al., 2013; Gao et al., 2014). Seasonal and annual climatology as well as changes in surface air temperature and precipitation for the dry and wet seasons are analyzed.

Surface Air Temperature

Some previous studies claim an elevation dependency of surface air temperature change over the TP (e.g., Qin et al., 2009) whereas others claim an absence of any elevation dependency in surface air temperature trend (You et al., 2010). Li et al. (2013) addressed an increasing trend over the western and a decreasing trend over the eastern TP in the lapse rate to the south of 35ºN during the period 1962 to 2011. Figure 4a shows the observed inter-annual variability of the lapse rate in 1979–2011 for the eastern and western TP. It was found that the lapse rate estimated from the observations shows a downward trend for both regions, although only the trend for the eastern TP passes a significant t-test at the 95% confidence level. The reason for the disagreement between lapse rate trends revealed in this paper as well as in Li et al. (2013) for the western TP may be due to the different stations and periods used. The difference in the estimating method may also have played a role. As there are many more stations in the eastern TP than in the western TP, the estimated lapse rate for the eastern region is considered more reliable. From Figure 4a we can see that the surface air temperature change is elevation-dependent over the whole TP in the observation.

The simulated surface air temperature in the RCM during 1979–2011 was found to be elevation-dependent over the whole TP in a similar way as the observation. Opposite to the observation, ERA-Interim displays an upward trend in the western TP (Figure 4) due to less warming in the central TP. The stronger large-scale surface warming toward the high latitudes was also found in the RCM simulations. Gao et al. (2015a) discovered that dynamical downscaling better captures the observed downward linear trend in the lapse rate and its temporal trend than the best reanalysis evaluated over the TP based on the 33-year WRF simulations. The magnitude and variability of the lapse rate in the WRF simulation matches the observation much better than ERA-Interim, with higher spatial correlation coefficients in both the dry and wet seasons and smaller RMSE in the wet season. Further, the elevation dependency in the observed temperature in the western and eastern TP is better reproduced by WRF than ERA-Interim.

Click to view larger

Figure 4. Scatter plots of Tair at the 83 stations over the TP against elevation: (a) the station observations, (b) ERA-Interim, c) the WRF simulation. West indicates the area to the west of 95°E; East is the area to the east of 95°E. (From Gao et al., 2015a)

As for the seasonal linear trends of the surface air temperature, the observed monthly mean temperature averaged over all the stations presents a higher warming rate in the dry season (November–April) than the wet season (May– September). This comes from a larger warming rate at stations in the central TP in the dry season than in the wet. ERA-Interim shows smaller warming rates than the observation for both the wet and dry seasons in the central and southern TP. The WRF simulation mostly inherits the ERA-Interim estimate of the temperature trends except for May to July, when the WRF simulates the trends more realistically in the wet season. This improvement originates from closer mean trend, smaller RMSE and higher significant pattern correlation between WRF and observation than ERA-Interim in the wet season, especially in the central TP, where the relative large warming rate is not captured by ERA-Interim. In the cold season, the temperature trend magnitude improvement in the WRF simulation for the central TP is somewhat leveled by the deterioration in the northeastern TP, which results in the same warming rate as ERA-Interim in the dry season. Interestingly, the WRF simulation shows larger spatial correlation with observations than ERA-Interim. The correlation for the WRF simulations passes the statistical two-paired t-test at the 99.9% confidence level, which proves the added skill of the WRF in simulating the regional warming pattern over the TP compared to forcing.

Precipitation

As for precipitation, a significant wet bias was noticed in the dynamical downscaling modeling, although the overestimation has been greatly constrained compared to its forcing. The 33-year WRF simulation driven by ERA-Interim approximately reduces 35% of the wet bias from the driving ERA-Interim in the wet season, although practically no improvement is noticed in the dry season. ERA-Interim suffers from a large number of wet biases over the TP except for the Chaidam Basin in the wet season. In the dry season, wet biases mainly exist in the southern TP in ERA-Interim, which was passed to the WRF simulation showing a similar bias pattern. In the wet season, not only are the magnitudes of wet biases in the driving ERA-Interim reduced in the WRF simulation, but also higher spatial correlation with observations is gained compared to ERA-Interim. This is especially true in the central TP, as several stations even show proximal zero bias.

Based on the observation, there are positive precipitation trends for both the dry and wet seasons. The largest increasing trend occurs in May, followed by August, which accounts for a large portion of the increasing trend in the wet season (0.024mm d−1 (10a)−1). At the same time, slightly negative trends are found in July and September. Observed precipitation from January to April has essentially no trend while showing a slightly positive trend in the dry season due to increases in October and November, despite a small decreasing trend in December. The observation presents no trend at most stations, scattered with few stations with positive trends from October to April of the following year. In the wet season, there is a positive trend in the central and northern TP which leads to the average positive trend. However, negative or no trends exist at several stations in the eastern TP. Annually, precipitation presents plentiful positive trends in the northwestern TP and none or negative trends in the southeastern TP. Pattern similarity of precipitation trends between ERA-Interim, the WRF simulation and the observation is quite weak, in line with the weak and scattered precipitation trend in the observation.

ERA-Interim presents greater positive trends than the observation for both seasons, especially from May to August. Spatially, larger trends appear at stations on the east margin of the TP in the dry season and most stations in the wet season. Annually, ERA-Interim presents the same trend pattern, which has a much greater magnitude than the observation. The WRF simulated trends are closer to the observation than those in ERA-Interim in the wet season. This further confirms the added value of the WRF simulation. Compared to its forcing, the WRF simulation improves the simulation of the annual cycles and temporal trends of precipitation in the wet season. Nevertheless, there is hardly any improvement in the dry season. The land surface model impacts the precipitation climatology and spatial distribution more significantly than the large-scale forcing through including the surface heating over the TP. Large-scale forcing has more influence on the trends than on the spatial characteristics in the WRF simulation.

Another 11-year (2001–2011) DDM, named High Asia Reanalysis (HAR) and driven by the NCEP analysis FNL using WRF with horizontal resolution of 30km and 10km, also proved the added values regarding snowfall retrieval, precipitation frequency, and orographic precipitation in terms of spatial patterns and seasonality (Maussion et al., 2014), compared to both rain-gauge observations and satellite-based precipitation estimates from the Tropical Rainfall Measurement Mission (TRMM). This focused on precipitation amounts, type, seasonality, and inter-annual variability. Special attention was given to the links between the observed patterns and regional atmospheric circulation. It found that the HAR reproduces previously reported spatial patterns and seasonality of precipitation and that the higher solution data add value regarding snowfall retrieval, precipitation frequency, and orographic precipitation. An improvement when increasing horizontal resolution from 30 to 10 km was demonstrated, quantitatively (by comparing to observations) as well as qualitatively (in the reproduction of documented orographic precipitation features). As an example of an application of the HAR, a new classification of glaciers on the TP according to their accumulation regimes is proposed (Figure 5) which illustrates the strong spatial variability of precipitation seasonality. For Maussion et al.’s study, the annual cycle of precipitation on the TP is characterized by a winter precipitation regime in the west, a spring precipitation regime in northern and southern TP, and a summer precipitation regime elsewhere. Glaciers in different precipitation regimes will respond differently to changes in climate and shifts in precipitation seasonality. The authors proposed a new map of glacier accumulation regimes as one possible application for the HAR. Their analysis illustrates the high spatial variability of precipitation seasonality and provides the first approach for glacier energy and mass balance considerations. The map resulting from this cluster analysis emphasizes that glaciers on the TP cannot be considered as one entity with uniform mass balance sensitivity (see Figure 5).

Click to view larger

Figure 5. Classification of glacier accumulation regimes according to precipitation seasonality. A k-means clustering algorithm is run on three input variables (percentage of precipitation falling in DJF, MAM (March–May), and JJA) and five output clusters. We focus on glacierized grid points only (bottom left). The histogram plot showing the relative occurrence of each class in the map with the color legend is described below (bottom left). The clusters are named after their cluster centers characteristics. As SON is a linear combination of the three other variables, it was not included in the clustering procedure. (From Maussion et al., 2014)

Water Balance

Net precipitation (P-E) was examined in water balances. As shown by Global Land Surface Data Assimilation System (GLDAS), ERA-Interim, and WRF simulation, P-E exhibits general increases in most of the northwestern TP and decreases in the southeastern TP from 1979 to 2011. However, the pattern of P-E changes in the WRF simulation (Figure 6h) resembles GLDAS (Figure 6g) more than ERA-Interim (Figure 6i). Pattern correlation of P-E changes between GLDAS and WRF is 0.17, which passes the statistical significance t-test at the 98% confidence level. In contrast, the correlation (0.04) between GLDAS and ERA-Interim does not pass statistical significance at even the 70% confidence level, although ERA-Interim is already reproducing the observed P-E climatology and trend in the TP better than other popular reanalyses. Most notably, the WRF simulation captures the larger increase in the Qiangtang Plateau over the northwest comparable to GLDAS, but ERA-Interim shows increases that are more evenly distributed over the TP and a slight positive gradient from northwestern to southeastern TP (Figure 6i). Furthermore, the decrease in the southeastern TP has smaller values and spreads across the lower reaches of the Upper Brahmaputra River Basin in WRF (Figure 6h) compared to the larger, more concentrated decreases in ERA-Interim (Figure 6i).

Click to view larger

Figure 6. Changes in (a–c) P, (d–f) E, and (g–i) P-E (unit: mm d−1) of 1998–2011 compared to 1979–1997 for OBS/GLDAS, WRF simulation, and ERA-Interim respectively over the TP. (From Gao et al., 2015b)

Interestingly, the pattern correlation coefficients between P-E in GLDAS with HGT and MODIS annual mean LAI averaged from 2001 to 2012 are 0.20 and −0.27 respectively. The positive correlation of the GLDAS P-E change with topography indicates larger P-E increase at higher elevation, while the negative correlation of the GLDAS P-E change with LAI suggests that P-E increases more over the dry land cover type with smaller LAI in GLDAS. These relationships of P-E changes in GLDAS with elevation and LAI are well captured by the WRF simulations showing high pattern correlations of P-E with topography and LAI (Gao et al., 2015b). However, none of these relationships are captured in ERA-Interim. This suggests a potential shift towards topography and vegetation distribution as the stronger drivers of P-E changes at the regional scale, and further supports the importance of resolving topography and land cover using models with higher resolution to simulate regional changes in P-E. Table 1 outlines the pattern correlations between these observations.

Table 1. Pattern correlations between annual P-E change in GLDAS and ERA-Interim/WRF simulation in 1998–2011 compared to 1979–1997 and pattern correlations between ERA-Interim/WRF simulation with topography (HGT) and leaf area index (LAI) climatology in 2000–2011. The degree of freedom is 189 (From Gao et al., 2015b).

GLDAS

ERA-Interim

WRF

GLDAS

1

0.04

0.17*

HGT

0.20*

0.02

0.16*

LAI

−0.27*

0.01

−0.20*

*Correlation coefficients with an asterisk are statistically significant at the 98% confidence level based on two-tailed t-test.

The high-resolution WRF climate simulation not only improves the pattern of P-E changes compared to GLDAS over the best available reanalysis, it also provides new and substantial findings regarding the contributions of the thermodynamic and the transient eddy components of moisture flux convergence. Most notably, the WRF simulation better represents the observed positive/negative changes in the vast northwestern/southeastern TP than the coarse resolution reanalysis forcing. The improved P-E change pattern is attributed to improved P changes and reduced P biases at high elevations in the high-resolution simulation. This demonstrates that for the TP with complex terrain, high-resolution climate simulations can provide important insights into regional water cycle changes.

The changes of P-E in post-phase relative to pre-phase are decomposed into dynamic contributor due to changes in mean circulation (MCD), thermodynamic contributor (TH) due to changes in mean specific humidity, and transient eddy moisture convergence (TE):

$Display mathematics$
(1)

It is noted that the MCD change plays a key role in the changes of P-E in ERA-Interim (Figure 7 right) over the TP. The TH contributes positively over the southern and central TP, while the TE tends to reinforce (offset) the dynamic component over the southern and parts of the northern TP (central TP). In dynamical downscaling, besides the predominant contribution of the MCD inherited from the large-scale forcing, TH and TE changes importantly contribute to P-E changes in the WRF simulation but not the reanalysis (Figure 7 left). Contrasting divergence changes at low levels between the northwestern and southeastern TP induced by land surface heterogeneity trigger stronger (weaker) upward motion in the vast northwestern (southeastern) TP, leading to P-E increases/decreases by MCD changes. The larger convergence in the northwestern TP attracts more moisture to Qiangtang Plateau and Qilian Mountain from the surroundings and strengthens P-E changes through latent heat released and TH change contribution.

Click to view larger

Figure 7. Distributions of annual (a, b) P-E changes and contributions from (c, d) mean circulation dynamics (δ‎MCD), (e, f) thermodynamics (δ‎TH), and (g, h) transient eddy (δ‎TE) from the WRF simulation (left), and ERA-Int (right) in 1998–2011 compared to 1979–1997 (Unit: mm d−1). (From Gao et al., 2015b)

Projections of Future Climate Change

Surface Air Temperature

CMIP3 and CMIP5 both project warming over the TP, but with different magnitudes. CMIP3 projects that the annual mean surface air temperature will increase by 3.8ºC with similar magnitudes in winter and summer by the end of the 21st century (Table 11.1 in AR4 report). Chen et al. (2011) analyzed 28 GCM outputs and found the surface air temperature during winter in the TP will increase by 3ºC with 100% possibility under SRES A1B. Further, the possibility of increasing 4ºC is above 80%. Compared to CMIP3, CMIP5 projects an increase of 2.6ºC in the annual mean surface air temperature (Table 14.1 in AR5 report). Differences in IPCC AR4 and AR5 temperature and precipitation projections over the TP may reflect the impact of the cryosphere change on local climate, which was included more in the projections of AR5 than AR4. Hu et al. (2015) examined an ensemble of 30 GCMs of CMIP5 and found the annual mean surface air temperature will increase 1.1ºC, 2.1ºC, and 2.7ºC in the early (2016–2035), middle (2046–2065), and latter (2081–2100) parts of the 21st century respectively under the Representative Concentration Pathway (RCP) 4.5. Su et al. (2013) claimed that all models produce a warming trend in the 21st century under the RCP8.5 scenario. In contrast, the RCP2.6 scenario predicts a lower average warming rate for the near term and a small cooling trend in the long term with decreasing radiative forcing. Temperature increases slightly more in winter than summer.

Projections based on RCM generally follow projections of its GCM forcing but with better spatial distributions (Yu et al., 2013; Gao et al., 2013; Ji & Kang, 2013; Tang et al., 2016). The ensemble averages of temperature are more consistent with observations than the simulations from individual RCMs (Tang et al., 2016).

Precipitation

As with temperature, we start by looking at GCM projections. For the annual mean precipitation, GCMs agreement is low on changes both for winter and summer precipitation. Liu et al. (2009) claim a slight increase no more than 5% in annual mean precipitation. However, Chen et al. (2011) project a significant increase in winter and summer precipitation by the end of the 21st century, with possibility over 60% and 80%.

IPCC AR4 and AR5 project similar magnitude increase over the TP, with 10% and 9% respectively. All MMD-A1B (Multi-Model Dataset-A1B) models in the AR4 project increased precipitation during DJF (median 19%). Most but not all models also simulate increased precipitation in the other seasons (Table 11.1 in AR4 report). The ability of CMIP5 models to simulate precipitation over the TP varies. Hu et al. (2015) examined the ensemble of 30 GCMs from CMIP5 and found the annual mean precipitation will increases by 4.4%, 7.9%, and 11.7% in the early (2016–2035), middle (2046–2065), and latter (2081–2100) parts of the 21st century respectively under RCP4.5. Su et al. (2013) projects about 3.2% higher than the 1961–2005 annual mean in the near future (2006–2035). In the long term future, precipitation is projected to increase 6.0% under RCP2.6 and 12.0% under the RCP8.5. Seasonally, the largest increase occurs in the summer and the lowest in winter.

RCMs simulations driven by CMIP3 are consistent with GCMs findings (e.g., Gao et al., 2003, 2011b). Some RCM simulations driven by CMIP5 project increase slightly more in winter than summer with a mild annual variation (Table 14.1 in AR5 report). However, some project less increases. It is worth to note that the projection is highly forcing (GCM) dependent. Due to the huge computing sources needed for RCM simulations, only a relatively small number of RCMs have been used for the TP. As a result, the uncertainty associated with RCM simulations may not sample as well as those of the GCMs.

Extreme Events

Future warming is expected to change the characteristics of extreme events. Indeed, with a faster warming rate over the TP than the global average, extreme events are projected to significantly increase. Ensembles of 24 GCMs in CMIP5 project an increase in extreme events related to maximum surface air temperature and a decrease in extreme events related to minimum temperature (Zou & Zhou, 2013). Under the RCP2.6, RCP4.5, and RCP8.5, a GCM (MPI_ESM_LR) in CMIP5 also projects the same increase in mean temperature related extreme indices. As expected, RCP8.5 increases the most and RCP2.6 the least. A consistency among RCPs is the significant increase in the maximum temperature at night and no changes in diurnal temperature range (DTR) because of the same increase magnitudes in the day time maximum temperatures and minimum nightly temperatures.

All CMIP3 and CMIP5 GCMs project increases in precipitation related extreme indices over the TP. The same as temperature projection, RCP8.5 increases the most and RCP2.6 the least. Jiang et al. (2012) project the intensity will increase 10%–26% by the end of the 21st century. The maximum of five consecutive day precipitation will increase 25%–45% during the same period.

Results based on available RCM studies agree that warming will continue over the TP, but the warming predicted by RCMs is less than that of the forcing GCMs. Dynamical downscaling projects increases in the maximum and minimum temperatures and decrease in the DTR in China (Zhang et al., 2006). Greater increases occur during DJF compared with JJA. The increasing temperature trend is more pronounced over the Gangdise Mountains and the Himalayas than in the central TP.

Dynamical downscaling driven by the CMIP3 GCMs projects that extreme precipitation over most parts of China will increase (Gao et al., 2002; Yu et al., 2013; Ji & Kang, 2013; Bao et al., 2015; Xu et al., 2012; Lang & Sui, 2013; Hu et al., 2013), especially in western and northern China, and that precipitation over some southern regions will decrease. The projected increase of future extreme precipitation makes great contributions to the total precipitation increase. Over the TP, the projection of precipitation shows the main increases during DJF. For JJA, it predicts decreases or slight changes in the southern TP. The comparison between RCP8.5 and RCP4.5 scenarios shows similar spatial distributions of temperature and precipitation, whereas the respective values of RCP8.5 are enhanced compared with those under RCP4.5 (Ji & Kang, 2013).

The impacts of the land surface model and the boundary conditions on the downscaled extreme temperature indices are investigated using different LSMs (Noah LSM versus Noah-MP) and boundary conditions (ERA-Interim versus CCSM) with the same model configurations except for the land surface model and the boundary conditions (Xiao et al., 2016). The results show that the basic features of the spatial patterns of all extreme temperature indices over the TP can be finely reproduced by the RCM. Due to the impacts of the boundary conditions, a climatic tendency rate of extreme temperature indices possesses negative biases. Despite being forced by the different boundary conditions, the two experiments coupled with the same land surface model reproduce similar spatial distributions of extreme temperature indices under current model configurations; of the two land surface models tested, Noah LSM produced significantly larger cold bias. However, the boundary conditions have more influence on climatic tendency rates of extreme temperature indices compared to the LSM, and thus the climatic tendency rates of WRF simulation driven by CCSM4 are close to observation.

Gao et al. (2017) evaluated nine extreme precipitation indices from three dynamic downscalings driven by ERA-Interim reanalysis against a gridded observational dataset for the historical period 1980–2005. From this, future projections were inter-compared with forcing for the period 2005–2100 under two scenarios (RCP4.5 and RCP8.5). The authors found that the coarse resolution forcings greatly overestimate the extreme precipitation indices. This is more pronounced for the frequencies of the extreme rains. The large overestimations in precipitation frequency and amount in GCMs are greatly constrained in the dynamical downscaling. More significantly, downscalings better present the observed interannual variabilities in extreme precipitation indices than their forcings. As for future projections, GCMs project a general wetting over the entire TP. Downscaling also projects wetting but with more light rain and less heavy rain than in GCM projections, which result in less sensitivity to warming in dynamical downscaling than the course-resolution forcing.

Concluding Remarks and Future Perspectives

Climate modeling works have thus far advanced our knowledge about the regional climate in the TP. For instance, all DDM studies have demonstrated that this process-based approach, despite some unavoidable shortcomings, can improve understandings of the processes leading to precipitation on the TP. However, much more is needed for applications to put the fragmented information and knowledge into a system perspective. Obviously, there are several scientific and practical challenges that remain to be resolved.

Firstly, all the existing dynamical downscaling is conducted using a specific RCM driven by a specific GCM, which means that the RCM skill is strongly limited by the skill of its driving GCM (Racherla et al., 2012). A way to deal with this problem is to use a multi-model ensemble approach (Gao et al., 2011). The CORDEX framework has been developed along this direction to improve the regional downscaling techniques and achieve robust and valuable conclusions in future regional climate projections (Giorgi et al., 2009). Li et al. (2016) and Tang et al. (2016) employed four ensemble methods using CORDEX outputs to generate the multi-model projection of regional climate change over East Asia. The historical results show that the regional temperature ensembles from all four methods differ from each other, whereas the ensembles outperform a single RCM result in aspects of the spatial distribution as well as the seasonal variation over East Asia. The potential of the CORDEX products for the TP has yet to be fully utilized.

Secondly, although most of the above-mentioned studies agree on the added value of RCMs in relation to their host GCMs, few showed the opposite. One such example is a recent analysis on the CORDEX South Asia regional climate models (CORDEX-RCMs) (Mishra, 2015). This study found that the RCMs are not able to reproduce the observed warming in the Himalayan water towers. CORDEX-RCMs overestimate observed warming by threefold in the Ganges and Brahmaputra basins, which raises questions about their reliability in projecting future warming trends for the region. The CORDEX-RCMs overestimate the area that experienced significant warming and fail to reproduce precipitation trends in both magnitude as well as direction. It was also established that in observational data sets, uncertainty in precipitation and air temperature increases with elevation, which may be associated with sparse observations. However, the CORDEX-RCMs showed larger uncertainty at the lower elevations in both precipitation and temperature. More importantly, the hosts GCMs show better performance in simulating winter climate than the CORDEX-RCMs. This issue certainly deserves more detailed studies.

Thirdly, dynamical downscaling is a local diagnostic process that represents a post processing of GCM simulations. This means that the local forcing could not be fed back into the large scales represented by GCMs to improve the rendering of climates for other regions. A natural way to improve this would be the use state-of-the-art GCMs with very fine resolution. Recently, Guo and Wang (2016) compared simulation results from a very-fine-resolution GCM and a RCM dynamical downscaling over China. They found that the very-fine-resolution GCM reproduces the climatology and trends of both air temperature and precipitation, as well as inter-monthly variations of air temperature in terms of spatial pattern and amount, closer to observations than the coarse-resolution version of the GCM. However, this is not the case for the inter-monthly variations of precipitation. The RCM dynamical downscaling method performs better than the very-fine-resolution GCM in terms of the climatology and inter-monthly variation of precipitation. This is most likely because RCMs have better microphysics and cumulus parameterizations compared with GCMs. This comparison suggests that very-fine-resolution GCMs possess great potential regarding their application in regional climate simulation in the future.

In terms of future development, the importance of high-quality observational data for the TP must not be underestimated, as any modeling efforts need to be verified by observations. However, due to the harsh environmental conditions, observation sites are sparsely scattered over the TP in locations to which humans have relatively easy access. Moreover, most observation sites are in the eastern TP valley (Gao et al., 2015a); for a large portion of the vast northwestern TP, there are no sites. Therefore, there is an unavoidable uncertainty in climate change analysis with in situ observations due to the poorly gauged observation network over the TP (Zhang et al., 2016). Several global and regional gridded datasets indeed cover the whole TP (e.g., Mitchell & Jones, 2005; Adler et al., 2003; Xu et al., 2008; Wu & Gao, 2013), but these heavily rely on observations which are scattered in the eastern TP, and therefore the reliability of these gridded products remains limited for the west. To reduce the uncertainty of available datasets and enhance our knowledge in the regional climate over the TP, more reliable products and innovative approaches are necessary. While satellite data have been playing an increasingly role (e.g., Chen et al., 2016), more advanced and integrated observation systems (e.g., Xu et al., 2008), fine-scale modeling (Curio et al., 2015), and regional reanalysis such as those for North America (Mesinger et al., 2006) hold promising future prospects.

Acknowledgments

Gao Y. is supported by the Ministry of Science and Technology of China (2013CB956004), the National Natural Science Foundation of China (91537105, 41322033), and the “100-Talent” program granted by the Chinese Academy of Sciences. Chen D. is supported by Swedish VR, STINT, BECC, and MERGE, as well as SNIC through S-CMIP.

Gao, Y., Xu, J., Chen, D., Gao, Y., Xu, J., & Chen, D. (2015). Evaluation of WRF mesoscale climate simulations over the Tibetan Plateau during 1979–2011. Journal of Climate, 28(7), 2823–2841.Find this resource:

IPCC. (2013). Climate change 2013: The physical science basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, U.K.: Cambridge University Press.Find this resource:

Maussion, F., Scherer, D., Mölg, T., Collier, E., Curio, J., & Finkelnburg, R. (2014). Precipitation seasonality and variability over the Tibetan Plateau as resolved by the High Asia Reanalysis. Journal of Climate, 27(5), 1910–1927.Find this resource:

Warner, T. T. (2011). Numerical weather and climate prediction. Cambridge, U.K.: Cambridge University Press.Find this resource:

References

Adler, R. F., Huffman, G. J., Chang, A., Ferraro, R., Xie, P.-P., Janowiak, J., . . . Nelkin, E. (2003). The Version-2 Global Precipitation Climatology Project (GPCP) Monthly Precipitation Analysis (1979–present). Journal of Hydrometeorology, 4(6), 1147–1167.Find this resource:

Bao, J., Feng, J., & Wang, Y. (2015). Dynamical downscaling simulation and future projection of precipitation over China. Journal of Geophysical Research: Atmospheres, 120(16), 8227–8243.Find this resource:

Bao, Y., Gao, Y., Lü, S., Wang, Q., Zhang, S., Xu, J., . . . Chang, Y. (2014). Evaluation of CMIP5 earth system models in reproducing leaf area index and vegetation cover over the Tibetan Plateau. Journal of Meteorological Research, 28(6), 1041–1060.Find this resource:

Betts, A. K., Ball, J. H., Beljaars, A. C. M., Miller, M. J., & Viterbo, P. A. (1996). The land surface-atmosphere interaction: A review based on observational and global modeling perspectives. Journal of Geophysical Research: Atmospheres, 101(D3), 7209–7225.Find this resource:

Bollasina, M., & Benedict, S., (2004). The role of the Himalayas and the Tibetan Plateau within the Asian monsoon system. Bulletin of the American Meteorological Society, 85(7), 1001–1004.Find this resource:

Chen, D., Tian, Y., Yao, T., & Ou, T. (2016). Satellite measurements reveal strong anisotropy in spatial coherence of climate variations over the Tibet Plateau. Scientific Reports, 6, 30304.Find this resource:

Chen, D., Xu, B., Yao, T., Guo, Z., Cui, P., Chen, F., . . . Zhang, T. (2015). Assessment of past, present and future environmental changes on the Tibetan Plateau. Chinese Science Bulletin, 60(32), 3025–3035.Find this resource:

Chen, W., Jiang, Z., Li, L., Chen, W., Jiang, Z., & Li, L. (2011). Probabilistic projections of climate change over China under the SRES A1B scenario using 28 AOGCMs. Journal of Climate, 24(17), 4741–4756.Find this resource:

Chen, Y., Yang, K., Tang, W., Qin, J., & Zhao, L. (2012). Parameterizing soil organic carbon’s impacts on soil porosity and thermal parameters for Eastern Tibet grasslands. Science China Earth Science, 55(6), 1001-1011.Find this resource:

Cubasch, U., Wuebbles, D., Chen, D., Facchini, M. C., Frame, D., Mahowald, N., & Winther, J.-G. (2013). Introduction. In T. F. Stocker, D. Qin, G.-K. Plattner, M. Tignor, S. K. Allen, J. Boschung, . . .P. M. Midgley (Eds.), Climate change 2013: The physical science basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, U.K.: Cambridge University Press.Find this resource:

Curio, J., Maussion, F., & Scherer, D. (2015). A 12-year high-resolution climatology of atmospheric water transport over the Tibetan Plateau. Earth System Dynamics, 6(1), 109–124.Find this resource:

Dee, D., Uppala, S. M., Simmons, A. J., Berrisford, P., Poli, P., Kobayashi, S., . . . Vitart, F. (2011). The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quarterly Journal of the Royal Meteorological Society, 137, 553–597.Find this resource:

Dee, D. P., & Uppala, S. (2009). Variational bias correction of satellite radiance data in the ERA-Interim reanalysis. Quarterly Journal of the Royal Meteorological Society, 135, 1830–1841.Find this resource:

Duan, A., Hu, J., Xiao, Z., Duan, A., Hu, J., & Xiao, Z. (2013). The Tibetan Plateau summer monsoon in the CMIP5 simulations. Journal of Climate, 26(19), 7747–7766.Find this resource:

Entekhabi, D., Rodriguez-Iturbe, I., & Castelli, F. (1996). Mutual interaction of soil moisture state and atmospheric processes. Journal of Hydrology, 184(1–2), 3–17.Find this resource:

Gao, X., Li, D., Zhao, Z., & Giorgi, F. (2003). Climate change due to greenhouse effects in Qinghai-Xizang Plateau and along the Qianghai-Tibet Railway. Plateau Meteorology, 22(5), 458–463.Find this resource:

Gao, X., Shi, Y., & Giorgi, F. (2011). A high resolution simulation of climate change over China. Science China Earth Sciences, 54(3), 462–472.Find this resource:

Gao, X., Wang, M., & Giorgi, F. (2013). Climate change over China in the 21st century as simulated by BCC_CSM1.1-RegCM4.0. Atmospheric and Oceanic Science Letters, 6(5), 381–386.Find this resource:

Gao, X., Zhao, Z., & Giorgi, F. (2002). Changes of extreme events in regional climate simulations over East Asia. Advances in Atmospheric Sciences, 19(5), 927–942.Find this resource:

Gao, Y., Chen, F., Barlage, M., Liu, W., Cheng, G., Li, X., . . .Ma, M. (2008). Enhancement of land surface information and its impact on atmospheric modeling in the Heihe River Basin, northwest China. Journal of Geophysical Research, 113(D20), D20S90.Find this resource:

Gao, Y., Cuo, L., Zhang, Y., Gao, Y., Cuo, L., & Zhang, Y. (2014). Changes in moisture flux over the Tibetan Plateau during 1979–2011 and possible mechanisms. Journal of Climate, 27(5), 1876–1893.Find this resource:

Gao, Y., Leung, L. R., Salathé, E. P., Dominguez, F., Nijssen, B., & Lettenmaier, D. P. (2012). Moisture flux convergence in regional and global climate models: Implications for droughts in the southwestern United States under climate change. Geophysical Research Letters, 39(9).Find this resource:

Gao, Y., Leung, L. R., Zhang, Y., & Cuo, L. (2015b). Changes in moisture flux over the Tibetan Plateau during 1979–2011: Insights from a high-resolution simulation. Journal of Climate, 28(10), 4185–4197.Find this resource:

Gao, Y., Li, K., Chen, F., Jiang, Y., & Lu, C. (2015c). Assessing and improving Noah-MP land model simulations for the central Tibetan Plateau. Journal of Geophysical Research: Atmospheres, 120, 9258–9278.Find this resource:

Gao, Y., Li, X., Leung, R. L., Chen, D., & Xu, J. (2015d). Aridity changes in the Tibet Plateau in a warming climate. Environmental Research Letters.Find this resource:

Gao, Y., Vano, J. A., Zhu, C., & Lettenmaier, D. P. (2011a). Evaluating climate change over the Colorado River basin using regional climate models. Journal of Geophysical Research, 116(D13), D13104.Find this resource:

Gao, Y., Xiao, L., Chen, D., Chen, F., Xu, J., & Xu, Y. (2016). Quantification of the relative role of land-surface processes and large-scale forcing in dynamic downscaling over the Tibetan Plateau. Climate Dynamics, 48(5–6), 1–17.Find this resource:

Gao, Y., Xiao, L., Chen, D., Xu, J., & Zhang, H. (2017). Comparison between past and future extreme precipitations simulated by global and regional climate models over the Tibetan Plateau. International Journal of Climatology, 16.Find this resource:

Gao, Y., Xu, J., Chen, D., Gao, Y., Xu, J., & Chen, D. (2015a). Evaluation of WRF mesoscale climate simulations over the Tibetan Plateau during 1979–2011. Journal of Climate, 28(7), 2823–2841.Find this resource:

Gao, Y., Xue, Y., Peng, W., Kang, H.-S., & Waliser, D. (2011b). Assessment of dynamic downscaling of the extreme rainfall over East Asia using a regional climate model. Advances in Atmospheric Sciences, 28(5), 1077–1098.Find this resource:

Giorgi, F., Jones, C., & Asrar, G. (2009). Addressing climate information needs at the regional level: The CORDEX framework. Organization (WMO) Bulletin, 58(3), 175.Find this resource:

Guo, D., & Wang, H. (2016). Comparison of a very-fine-resolution GCM with RCM dynamical downscaling in simulating climate in China. Advances in Atmospheric Sciences, 33(5), 559–570.Find this resource:

Held, I. M., & Soden, B. J. (2006). Robust responses of the hydrological cycle to global warming. Journal of Climate, 19, 5686–5699.Find this resource:

Hu, B. Y., Tang, J. P., & Wang, S. Y. (2013). Evaluation and projection of extreme events over China under IPCC A1B scenario by MM5V3 model. Chinese Journal of Geophysics-Chinese Edition, 56(7), 2195–2206.Find this resource:

Hu, Q., Jiang, D., & Fan, G. (2015). Climate change projection on the Tibetan Plateau: Results of CMIP5 models. Chinese Journal of Atmospheric sciences, 39(2), 260–270.Find this resource:

Hua, W., Chen, H., & Li, X. (2015). Effects of future land use change on the regional climate in China. Science China Earth Sciences, 58(10), 1840–1848.Find this resource:

IPCC. (2013). Climate change 2013: The physical science basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, U.K.: Cambridge University Press.Find this resource:

Ji, Z., & Kang, S. (2013). Double nested dynamical downscaling experiments over the Tibetan Plateau and their projection of climate change under RCPs scenarios. Journal of the Atmospheric Sciences, 70, 1278–1290.Find this resource:

Ji, Z., & Kang, S. (2013). Projection of snow cover changes over China under RCP scenarios. Climate Dynamics, 41, 589–600.Find this resource:

Jiang, Z., Song, J., Li, L., Chen, W., Wang, Z., & Wang, J. (2012). Extreme climate events in China: IPCC-AR4 model evaluation and projection. Climatic Change, 110(1–2), 385–401.Find this resource:

Kalnay, E., Kanamitsu, M., Kistler, R., Collins, W., Deaven, D., Gandin, L., . . . Joseph, D. (1996). The NCEP/NCAR 40-year reanalysis project. Bulletin of the American Meteorological Society, 77(3), 437–471.Find this resource:

Kanamitsu, M. W., Ebisuzaki, J., Woollen, S., Yang, S. K., Hnilo, J. J., Fiorino, M., & Potter, G. L. (2002). NCEP–DOE AMIP-II reanalysis (R-2). Bulletin of the American Meteorological Society, 83, 1631–1643.Find this resource:

Kasahara, A., & Washington, W. M., (1971). General circulation experiments with a six-layer NCAR Model, including orography, cloudiness and surface temperature calculations. Journal of the Atmospheric Sciences, 28(5), 657–701.Find this resource:

Kim, J., Jung, H.-S., Mechoso, C. R., & Kang, H.-S. (2008). Validation of a multidecadal RCM hindcast over East Asia. Global and Planetary Change, 61(3), 225–241.Find this resource:

Lang, X., & Sui, Y. (2013). Changes in mean and extreme climates over China with a 2°C global warming. Chinese Science Bulletin, 58(12), 1453–1461.Find this resource:

Lenton, T. M., Held, H., Kriegler, E., Hall, J. W., Lucht, W., Rahmstorf, S., & Schellnhuber, H. J. (2008). Tipping elements in the Earth’s climate system. Proceedings of the National Academy of Sciences of the United States of America, 105(6), 1786–1793.Find this resource:

Li, Q., Wang, S., Lee, D.-K., Tang, J., Niu, X., Hui, P., . . . Sasaki, H. (2016). Building Asian climate change scenario by multi-regional climate models ensemble. Part II: Mean precipitation. International Journal of Climatology, 36(13), 4253–4264.Find this resource:

Li, Q., & Xue, Y. (2010). Simulated impacts of land cover change on summer climate in the Tibetan Plateau. Environmental Research Letters, 5(1), 15102.Find this resource:

Liu, X., Cheng, Z., & Zhang, R. (2009). The A1B scenario projection for climate change over the Tibetan Plateau in the next 30–50 years. Plateau Meteorology, 28(3), 475–484.Find this resource:

Li, X., Wang, L., Chen, D., Yang, K., Xue, B., & Sun, L. T. (2013). Near-surface air temperature lapse rates in the mainland China during 1962–2011. Journal of Geophysical Research: Atmospheres, 118(14), 7505–7515.Find this resource:

Manabe, S., Smagorinsky, J., & Strickler, R. F. (1965). Simulated climatology of a general circulation model with a hydrologic cycle. Monthly Weather Review, 93(December), 769–798.Find this resource:

Maussion, F., Scherer, D., Mölg, T., Collier, E., Curio, J., & Finkelnburg, R. (2014). Precipitation seasonality and variability over the Tibetan Plateau as resolved by the High Asia Reanalysis. Journal of Climate, 27(5), 1910–1927.Find this resource:

Mesinger, F., DiMego, G., Kalnay, E., Mitchell, K., Shafran, P. C., Ebisuzaki, W., . . . Shi, W. (2006). North American regional reanalysis. Bulletin of the American Meteorological Society, 87(3), 343–360.Find this resource:

Mishra, V. (2015). Climatic uncertainty in Himalayan water towers. Journal of Geophysical Research: Atmospheres, 120(7), 2689–2705.Find this resource:

Mitchell, T. D., & Jones, P. D. (2005). An improved method of constructing a database of monthly climate observations and associated high-resolution grids. International Journal of Climatology, 25(6), 693–712.Find this resource:

Nigam, S., and Ruiz-Barradas, A., (2006). Seasonal hydroclimate variability over North America in global and regional reanalyses and AMIP simulations: Varied representation. Journal of Climate, 19(5), 815–837.Find this resource:

Onogi, K., Tsutsui, J., Koide, H., Sakamoto, M., Kobayashi, S., Hatsushika, H., . . . Taira, R. (2007). The JRA-25 Reanalysis. Journal of the Meteorological Society of Japan, 85(3), 369–432.Find this resource:

Ou, T., Chen, D., Linderholm, H. W., & Jeong, J.-H. (2013). Evaluation of global climate models in simulating extreme precipitation in China. Tellus A, 65, 1–16.Find this resource:

Phillips, N. A. (1956). The general circulation of the atmosphere: A numerical experiment. Quarterly Journal of the Royal Meteorological Society, 82(352), 123–164.Find this resource:

Qin, J., Yang, K., Liang, S., & Guo, X. (2009). The altitudinal dependence of recent rapid warming over the Tibetan Plateau. Climatic Change, 97(1–2), 321–327.Find this resource:

Racherla, P. N., Shindell, D. T., & Faluvegi, G. S. (2012). The added value to global model projections of climate change by dynamical downscaling: A case study over the continental U.S. using the GISS-ModelE2 and WRF models. Journal of Geophysical Research: Atmospheres, 117(D20).Find this resource:

Rodell, M., Houser, P. R., Jambor, U., Gottschalck, J., Mitchell, K., Meng, C., . . . Toll, D. (2004). The Global Land Data Assimilation System. Bulletin of the American Meteorological Society, 85(3), 381–394.Find this resource:

Shen, M., Piao, S., Jeong, S.-J., Zhou, L., Zeng, Z., Ciais, P., . . .Yao, T. (2015). Evaporative cooling over the Tibetan Plateau induced by vegetation growth. Proceedings of the National Academy of Sciences of the United States of America, 112(30), 9299–9304.Find this resource:

Song, J.-H., Kang, H.-S., Byun, Y.-H., & Hong, S.-Y. (2010). Effects of the Tibetan Plateau on the Asian summer monsoon: A numerical case study using a regional climate model. International Journal of Climatology, 30(5).Find this resource:

Su, F., Duan, X., Chen, D., Hao, Z., & Cuo, L. (2013). Evaluation of the Global Climate Models in the CMIP5 over the Tibetan Plateau. Journal of Climate, 26(10), 3187–3208.Find this resource:

Su, F., Zhang, L., Ou, T., Chen, D., Yao, T., Tong, K., & Qi, Y. (2016). Hydrological response to future climate changes for the major upstream river basins in the Tibetan Plateau. Global and Planetary Change, 136, 82–95.Find this resource:

Tang, J., Li, Q., Wang, S., Lee, D.-K., Hui, P., Niu, X., . . . Sasaki, H. (2016). Building Asian climate change scenario by multi-regional climate models ensemble. Part I: Surface air temperature. International Journal of Climatology, 36(13), 4241–4252.Find this resource:

Taylor, K. E., Stouffer, R. J., Meehl, G. A., Taylor, K. E., Stouffer, R. J., & Meehl, G. A. (2012). An overview of CMIP5 and the experiment design. Bulletin of the American Meteorological Society, 93(4), 485–498.Find this resource:

Trenberth, K. E., Stepaniak, D. P., Hurrell, J. W., & Fiorino, M. (2001). Quality of reanalyses in the tropics. Journal of Climate, 14(7), 1499–1510.Find this resource:

Uppala, S. M., Kallberg, P. W., Simmons, A. J., Andrae, U., Da Costa Bechtold, V., Fiorino, M., K., . . .Woollen, J. (2005). The ERA-40 re-analysis. Quarterly Journal of the Royal Meteorological Society, 131(612), 2961–3012.Find this resource:

Wu, G., Duan, A., Liu, Y., Mao, J., Ren, R., Bao, Q., . . .Hu, W. (2015). Tibetan Plateau climate dynamics: Recent research progress and outlook. National Science Review, 2(1), 100–116.Find this resource:

Wu, J., & Gao, X. (2013). A gridded daily observation dataset over China region and comparison with the other datasets. Chinese Journal of Geophysics-Chinese Edition, 56(4), 1102–1111.Find this resource:

Xiao, L., Gao, Y., Chen, F., Xu, J., Li, K., Li, X., & Jiang, Y. (2016). Dynamic downscaling simulation of extreme temperature indices over the Qinghai-Xizang Plateau. Plateau Meteorology, 35(3), 574–589.Find this resource:

Xu, J., Gao, Y., Chen, D., Xiao, L., & Ou, T. (2016). Evaluation of global climate models for downscaling applications centred over the Tibetan Plateau. International Journal of Climatology, 37, 657–671.Find this resource:

Xu J., Shi, Y., & Gao, X. (2012). Changes in extreme events as simulated by a high-resolution regional climate model for the next 20–30 years over China. Atmospheric and Oceanic Science Letters, 5(6), 483–488.Find this resource:

Xu, X., Zhang, R., Shi, X., Zhang, S., Bian, L., Cheng, X., & Ding, G. (2008). A new integrated observational system over the Tibetan Plateau. Bulletin of the American Meteorological Society, 89(10), 1492–1496.Find this resource:

Yang, K., Chen, Y., & Qin, J. (2009). Some practical notes on the land surface modeling in the Tibetan Plateau. Hydrology and Earth System Sciences, 13(5), 687–701.Find this resource:

Yang, K., Koike, T., Ye, B., & Bastidas, L. (2005). Inverse analysis of the role of soil vertical heterogeneity in controlling surface soil state and energy partition. Journal of Geophysical Research: Atmospheres, 110(8), D13105.Find this resource:

Yao, T., Masson-Delmotte, V., Gao, J., Yu, W., Yang, X., Risi, C., . . . Hou, S. (2013). A review of climatic controls on δ‎18O in precipitation over the Tibetan Plateau: Observations and simulations. Review of Geophysics, 51(4), 525–548.Find this resource:

Yao, T., Thompson, L., Yang, W., Yu, W., Gao, Y., Guo, X., . . . Joswiak, D. (2012). Different glacier status with atmospheric circulations in Tibetan Plateau and surroundings. Nature Climate Change, 2(9), 663–667.Find this resource:

You, Q., Kang, S., Pepin, N., Fluegel, W., Yan, Y., Behrawan, H., & Huang, J. (2010). Relationship between temperature trend magnitude, elevation and mean temperature in the Tibetan Plateau from homogenized surface stations and reanalysis data. Global and Planet. Change, 71(1–2), 124–133.Find this resource:

Yu, E., Wang, H., Sun, J., & Gao, Y. (2013). Climatic response to changes in vegetation in the Northwest Hetao Plain as simulated by the WRF model. International Journal of Climatology, 33(6), 1470–1481.Find this resource:

Zhang, C., Tang, Q., & Chen, D. (2016). Recent changes in the moisture source of precipitation over the Tibetan Plateau. Journal of Climate, 30, 1807–1819.Find this resource:

Zhang, Y., Xu, Y., Dong, W., Cao, L., & Sparrow, M. (2006). A future climate scenario of regional changes in extreme climate events over China using the PRECIS climate model. Geophysical Research Letters, 33(24), L24702.Find this resource:

Zhu, X., Wang, W., Fraedrich, K., Zhu, X., Wang, W., & Fraedrich, K. (2013). Future climate in the Tibetan Plateau from a statistical regional climate model. Journal of Climate, 26(24), 10125–10138.Find this resource:

Zou, L., & Zhou, T. (2013). Near future (2016–40) summer precipitation changes over China as projected by a regional climate model (RCM) under the RCP8.5 emissions scenario: Comparison between RCM downscaling and the driving GCM. Advances in Atmospheric Sciences, 30(3), 806–818.Find this resource: