Regional and Teleconnected Impacts of Solar Radiation‐Topography Interaction Over the Tibetan Plateau

Solar radiation‐topography interaction plays an important role in surface energy balance over the Tibetan Plateau (TP). However, the impacts of such interaction over the TP on climate locally and in the Asian regions remain unclear. This study uses the Energy Exascale Earth System Model (E3SM) to evaluate the regional and teleconnected impacts of solar radiation‐topography interaction over the TP. Land‐atmosphere coupled experiments show that topography regulates the surface energy balance, snow processes, and surface climate over the TP across seasons. Accounting for solar radiation‐topography interaction improves E3SM simulation of surface climate. The winter cold bias in air temperature decreases from −4.57 to −3.79 K, and the wet bias in summer precipitation is mitigated in southern TP. The TP's solar radiation‐topography interaction further reduces the South and East Asian summer precipitation biases. Our results demonstrate the topographic roles in regional climate over the TP and highlight its teleconnected climate impacts.


Plain Language Summary
The Tibetan Plateau (TP) is characterized by high elevation and complex topography.Interaction between solar radiation and the undulating topography has important impacts on the regional surface energy balance and hydrologic cycle.Here we use Earth System Model simulations to show the local and remote impacts of the TP's solar radiation-topography interaction on the surface climate of the Asian regions.Such interaction overall increases the air temperature especially in winter over the TP and reduces the summer precipitation in southern TP.Teleconnectedly, the interaction further alters the precipitation patterns in South and East Asia, by altering the atmospheric circulation that influences moisture transport and clouds.Accounting for such interaction generally improves the model performance when benchmarked against observations.These findings underscore the important roles of the TP's solar radiation-topography interaction in modulating the climate of the local and remote Asian regions.HAO ET AL.
• Solar radiation-topography interaction over the Tibetan Plateau (TP) increases annual average regional near-surface air temperature by 0.26 K • Solar radiation-topography interaction over the TP also affects the precipitation patterns in South and East Asia • Including solar radiation-topography interaction improves the simulation of surface climate over the TP and Asian regions Supporting Information: Supporting Information may be found in the online version of this article.
and Geophysical Fluid Dynamics Laboratory ESM (Zorzetto et al., 2023).Besides, a similar scheme was implemented in the Common Land Model (Zhang et al., 2022).Such model enhancement allows us to systematically explore the impacts of radiation-topography interaction regionally and globally.
Radiation-topography interaction has been found to have large impacts on land surface and atmospheric processes over mountainous regions such as the Tibetan Plateau (TP).Driven by meteorological forcing, offline land simulations showed non-negligible effects of topography on surface energy balance and snow dynamics across a wide range of spatial resolutions from 1-km to 2° (Hao et al., 2021(Hao et al., , 2022;;Zorzetto et al., 2023).However, offline land simulations neglect the impacts of land-atmosphere interaction, motivating the use of land-atmosphere coupled experiments to investigate how radiation-topography interaction impacts both atmospheric and land processes over the TP.Most ESMs tend to underestimate air temperature (T air ) and overestimate precipitation (P) over the TP across seasons (Cui et al., 2021;Zhu & Yang, 2020).The inclusion of radiation-topography interaction in land-atmosphere coupled CESM simulations reduces the cold bias over the TP in winter (Lee et al., 2019), and overall decreases P across seasons (Fan et al., 2019).
Besides local impacts, changes in the elevated heating due to radiation-topography interaction over the TP may influence the climate in other regions through teleconnection by excitation of Rossby waves.As the highest plateau in the Earth surface with large snow cover, TP plays an important role in modulating the atmospheric circulation and shaping the weather and climate around the TP (Wu et al., 2007(Wu et al., , 2014;;Yang et al., 2020).For example, the spring surface temperature over the TP shows a lag correlation with summer P in East Asia (Xue et al., 2022).Changing snow cover over the TP can advance/delay the onset and strengthen/weaken the intensity of the East Asian Summer Monsoon, and strongly influence the South Asian Summer Monsoon precipitation (Li et al., 2018;You et al., 2020).The projected surface darkening due to reduced snowpack by global warming will strengthen the elevated heat pump of the TP, and further impact the remote Asian monsoon systems (Tang et al., 2023).Likewise, the radiation-topography interaction-induced changes in land surface thermal conditions over the TP are expected to regulate the transport of water and heat to the Asian downstream regions and further impact the climate of these surrounding regions.
This study aims to investigate the regional and teleconnected impacts of radiation-topography interaction over the TP.Specifically, we used E3SM to carry out present-day 40-year land-atmosphere coupled experiments using three different model configurations to isolate the impact of radiation-topography interaction over the TP.Using these experiments and six benchmark data sets, we evaluated the local impacts of radiation-topography interaction on the TP's surface climate across seasons, followed by analysis of the remote impacts of the TP on T air and P patterns of the East and South Asian regions.

Radiation-Topography Interaction in E3SM
E3SM, supported by the United States Department of Energy (DOE), is a state-of-the-art fully coupled ESM developed to address the grand challenge of robust, actionable predictions of the variability and change of the Earth system (Leung et al., 2020).The latest version-2 of E3SM (E3SMv2) features significant developments especially in the atmospheric dynamical core and physics parameterization schemes, river routing, ocean and sea ice components (Golaz et al., 2022).Compared to its predecessor E3SM version-1, E3SMv2 shows a higher computational efficiency and improved performance in simulated clouds and P patterns (Golaz et al., 2022).The E3SM Land Model version-2 (ELMv2), includes a more realistic snow albedo parameterization (Dang et al., 2019) and an improved land biogeochemistry representation (Burrows et al., 2020) for various simulation campaigns.The new optional radiation-related configurations in ELMv2 include the TOP parameterization (Hao et al., 2021), support for multiple types of snow grain shape, and updates to the snow albedo parameterizations to account for different mixing states of snow grain and light-absorbing particles (LAP) (Hao et al., 2023).
ELMv2, by default, uses the PP two-stream approximation parameterization (Sellers, 1985) to calculate surface shortwave radiation balance without accounting for the impacts of topographic relief.The new TOP parameterization in ELMv2 can capture the subgrid topographic effects on solar radiation (Hao et al., 2021).TOP represents the relationship between the topography-related factors and the radiation adjustments caused by subgrid topography via multiple linear regression (Hao et al., 2021).The land-only ELM simulations showed that TOP has better performance in simulating surface energy balance and water cycles in the TP than PP (Hao et al., 2021(Hao et al., , 2022)).
The performance of TOP in land-atmosphere coupled simulations in and around the TP is evaluated in this study.

Experimental Design
We conducted land-atmosphere coupled present-day simulations using E3SMv2 with three different configurations: (a) the default PP scheme, denoted as PP_Globe; (b) the TOP scheme for the TP region and PP for the rest of the globe (Figure S1 in Supporting Information S1), denoted as TOP_TP; and (c) the TOP scheme for the global land, denoted as TOP_Globe.For each simulation, we used the F2010 component set with only active land, atmosphere and river components.In the F2010 configuration, the solar constant, sea surface temperature, sea ice, greenhouse gas concentrations, and aerosol emissions are prescribed at the 2010 level.The E3SM Atmosphere Model version-2 (EAMv2) was set at approximately 1° spatial resolution with 72 vertical layers.
ELMv2 was configured at a 0.5° spatial resolution in the satellite phenology mode driven by the satellite-derived climatological leaf area index data.We ran 40-year global simulations and used the last 20-year simulations for model analysis.Both the EAMv2 and ELMv2 outputs were aggregated to seasonal and annual mean.The EAMv2 outputs were resampled to 0.5° for further analysis.

Model Analysis and Evaluation
To clarify the role of radiation-topography interaction, the three model simulations described in Section 2.2 were compared in the TP and downstream over East and South Asia, across the seasons: winter (DJF), spring (MAM), summer (JJA), autumn (SON) as well as annual average.Specifically, the difference between TOP_TP and PP_ Globe was used to investigate the local and remote impacts of radiation-topography interaction over the TP.We also evaluated the difference between TOP_Globe and PP_Globe to diagnose the impacts of non-TP mountainous regions.Specifically, for the local impacts, we compared the spatiotemporal differences in , surface radiation fluxes, turbulent heat fluxes, snow cover fraction (f sno ), snow water equivalent (SWE), T air , and P. For the remote effects, we investigate the impacts on T air and P in two subregions: South Asia (SA; 10-25°N, 70-100°E) and East Asia (EA; 17-49°N, 105-140°E).
We collected six benchmark data sets from 2005 to 2015 for model evaluation (Table S1 in Supporting Information S1): (a) the surface radiation fluxes from the Clouds and the Earth's Radiant Energy System (CERES) Energy Balanced and Filled (EBAF) Edition 4.2 (NASA/LARC/SD/ASDC, 2023), (b) latent (F lat ) and sensible (F sen ) heat fluxes from FLUXCOM (Jung et al., 2019), (c) the spatially-and temporally complete (STC) snow-covered area and grain size (STC-MODSCAG) product (Rittger et al., 2020), (d) the snow property inversion from remote sensing (SPIReS) product (Bair et al., 2021), (e) T air from the University of Delaware (UDel) v5.01 monthly data (Willmott & Matsuura, 1995), and (f) P from the Global Precipitation Climatology Project (GPCP) v2.3 (Adler et al., 2018;Huffman et al., 1997).We used the average of STC-MODSCAG and SPIReS as the reference values of f sno .All the data sets were resampled spatially to 0.5° and temporally to multi-year average seasonal mean to be identical with the model outputs.Based on these benchmark data sets, statistical metrics including correlation coefficient (R), area-weighted mean bias, and area-weighted root-mean-square-error (RMSE) were used to evaluate the model performance.We calculated the relative difference in mean bias and RMSE (δ Bias and δ RMSE ; unit: %) between TOP_TP and PP_Globe as: (2)

Regional Impacts on Surface Energy Balance and Surface Climate
TP's surface energy balance and surface climate show distinct differences from those of the surrounding regions (Figure S2 in Supporting Information S1) due to its high elevation.Radiation-topography interaction regulates the annual average surface energy balance over the TP.In TOP_TP, topography reduces  by 0.01 (mean value), especially in the central and southern TP by more than 0.05, while increasing  in the northern border of the TP (Figure 1a).TOP_TP shows lower cloud cover (Figure S3f in Supporting Information S1) and thus larger downward solar radiation by 0.28 W/m 2 (mean value) than PP_Globe (Figure S3a in Supporting Information S1).
Consequently, TP absorbs more solar radiation with the mean value of about 2.10 W/m 2 (Figure 1b), driven by the reduction of  and increase of downward solar radiation.Given that TP has an annual average downward solar radiation of 231 W/m 2 and annual average  of 0.34 in PP_Globe, the topography-induced  reduction accounts for about 96% of the increase in net solar radiation (  R   ), while the downward solar radiation changes only account for about 4% of the difference between TOP_PP and PP_Globe over the TP.The downward longwave radiation shows a small change of 0.19 W/m 2 in mean value (Figure S3b in Supporting Information S1) responding to the cloud cover change, while the upward longwave radiation increases by 1.27 W/m 2 (mean value) (Figure S3c in Supporting Information S1) associated with surface warming (Figure 1g).The change in radiation fluxes further increases both F lat and F sen (Figures 1c and 1d).F sen shows a larger increase by 0.79 W/m 2 (mean value) than F lat (0.18 W/m 2 ).For the seasonal variation, overall winter shows larger  reduction than summer (Figure S4a in Supporting Information S1) due to higher snow cover and snow albedo.However, the topography-induced changes in    net for all the four seasons are comparable (Figure S4b in Supporting Information S1), because the seasonal variation of solar angle affects the available solar radiation.Although F sen increases for all seasons (Figure S4d in Supporting Information S1), F lat shows smaller changes for all seasons and even decreases in autumn (Figure S4c in Supporting Information S1).Besides the TP's mean changes, the differences in radiative fluxes show larger seasonal variations of spatial patterns (Figures S5-S8 in Supporting Information S1).
The increasing    net overall decreases annual average f sno over the TP (Figure 1e).The spatial pattern of the change in annual average f sno is consistent with that of  (Figures 1a and 1e), attributed to the reduced  that increases

𝐴𝐴
net which further accelerates snowmelt.Different from f sno , SWE increases in the western TP and decreases in the central and southern TP (Figure 1f).Although topography can accelerate snow melt (Figure S3e in Supporting Information S1), the increasing snowfall compensates the loss of snow masses over the western TP (Figure S3f in Supporting Information S1).The snow-atmosphere interaction complicates the snow changes caused by topography compared to the offline land simulations.At the seasonal scale, similar to , f sno shows larger differences in cold seasons and smallest changes in summer due to the smaller snow cover (Figure S4e in Supporting Information S1).Similarly, SWE shows the largest decrease by 4.3 mm in winter (Figure S4f in Supporting Information S1).
The increase in F sen leads to higher annual average near-surface (i.e., 2 m) T air over the whole TP (Figure 1g) by a mean value of 0.26 K.The increase in T air for winter is more pronounced with a mean value of 0.78 K, while there is a slight decrease in the western and Southern TP for summer (Figures S4g and S7 in Supporting Information S1).The changing T air and F lat affect the water and heat exchange between land and atmosphere, and eventually affect the regional P. Overall the central and northern TP shows a decrease in annual average P, while the western TP shows an increase (Figure 1h).The seasonal differences in P are well correlated with the differences in clouds (Figure S3f in Supporting Information S1).Topography reduces the summer P by 0.1 mm/day (mean value), especially in the central and southern TP (Figure 2h), related to the changing winds and cloud cover (Figures S3 and S9 in Supporting Information S1).The topography-induced increase in winter T air and decrease in summer P are expected to reduce the cold and wet bias of E3SM in the TP (Figures 2a, 2c, 2e, and 2g) which is also documented in (Zhu & Yang, 2020).
Including radiation-topography interaction overall improves the E3SM model performance in simulating surface energy balance, snow processes, and surface climate over the TP.For the annual scale, TOP_TP generally shows similar correlation, but smaller mean bias and RMSE with/than PP_Globe (Table 1).For example, δ Bias of  is −7.4% which means that the mean bias of TOP_TP reduces 7.4% compared to that of PP_Globe, while the mean biases of    net , f sno , and T air reduces 23.2%, 6.7%, and 11.1%, respectively.Seasonally (Table S2 in Supporting Information S1), for , both TOP_TP and PP_Globe show high correlation (≥0.59) with CERES across seasons, but large positive mean bias and large RMSE, especially in winter.TOP_TP reduces about 5% and 6% of the mean biases compared to PP_Globe for winter and spring.For    net , TOP_TP improves the correlation with CERES from 0.62 (PP_Globe) to 0.68 in winter and from 0.39 (PP_Globe) to 0.53 in autumn.TOP_TP shows slightly higher positive mean biases in summer, but lower negative mean biases and smaller RMSEs than PP_ Globe in other seasons.For F lat , both TOP_TP and PP_Globe are similarly well correlated with FLUXCOM, but show large overestimations in non-winter seasons, although TOP_TP shows slightly lower negative mean biases in winter than PP_Globe.For F sen , both TOP_TP and PP_Globe show low R values, high negative mean biases and large RMSEs especially in the warm seasons.Compared to PP_Globe, TOP_TP shows higher correlations in winter and autumn, and smaller negative mean biases across seasons than PP_Globe.F lat and F sen generally show opposite mean biases, implying that there are large uncertainties in partitioning the turbulent heat fluxes in E3SM.For f sno , TOP_TP shows slightly better performance for all the three metrics.For T air , TOP_TP shows similar R values with PP_Globe but reduces the cold bias especially in the cold seasons.For example, the negative mean bias of winter T air deceases from −4.57K to −3.79 K.For P, TOP_TP generally has similar mean biases and RMSEs, but higher R values than PP_Globe.For example, R increases from 0.65 to 0.72 in spring, and TOP_TP also slightly reduces the summer wet bias.
Comparing TOP_Globe and PP_Globe produces results that are spatially (Figure S10 in Supporting Information S1) and temporally (Figure S11 in Supporting Information S1) similar to the comparison between TOP_TP and PP_Globe, despite some differences in the magnitude of the statistical metrics (Table S3 in Supporting Information S1).

Teleconnected Impacts on East and South Asian Air Temperature and Precipitation
Including radiation-topography interaction over the TP overall reduces the bias of T air in the land regions of SA and EA.PP_Globe and TOP_TP show similarly high correlations with the UDel data across seasons with R values ≥0.70 and ≥0.89, respectively in SA and EA (Table S4 in Supporting Information S1).PP_Globe shows cold biases in annual T air over the TP's surrounding Asian regions (Figure 2a and Figure S12 in Supporting Information S1).For annual average in SA, PP_Globe has a cold bias of −1.28 K, while in EA, the cold bias is −0.90 K.For summer, PP_Globe has a cold bias of −0.89 K in SA, but a warm bias of +0.54 K in EA.Compared to PP_Globe, TOP_TP increases the annual and summer T air in India, but shows small changes in other SA regions.The winter in SA shows larger reductions of T air bias by 0.24 K than other seasons (Table S4 in Supporting Information S1).In EA, TOP_TP increases annual T air in north China, but reduces annual T air in other EA regions.For summer, TOP_TP reduces the warm biases in northeast Asia, but increase the biases in north China.The summer warm bias in EA reduces from +0.54 K of PP_Globe to +0.49K of TOP_TP.
Radiation-topography interaction over the TP affects the P patterns in EA and SA, possibly through its influence on the atmospheric circulation.The impacts of such interaction on annual and summer P show heterogeneous spatial patterns in the Asian regions (Figures 2d and 2h).In India and East China, TOP_TP overall shows smaller annual average P than PP_Globe (Figure 2d).In summer, TOP_TP reduces P in India, but increases it in the eastern regions of SA (Figure 2h).This is because the topography-induced wind anomaly weakens the climatological westerly wind and associated water transport from the Arabian Sea, while intensifying water transport from the continent to the Bay of Bengal (Figure S9 in Supporting Information S1).The summer P difference between TOP_TP and PP_Globe shows a tripolar structure of "north decrease-middle increase-south decrease" in East China (Figure 2h), which is a dominant pattern of P natural variability of the region (Xue et al., 2023).The tripolar pattern of changes in P is associated with the pattern wind changes at 850-hPa which enhance convergence of water vapor transport to central China while water vapor is diverged northward and southward in northern and southern China (Figure S9 in Supporting Information S1).
Including the TP's radiation-topography interaction overall improves the P simulations in SA and EA.This is already apparent in Figures 2c and 2d for annual P and Figures 2g and 2h for summer P, as the differences between TOP_TP and PP_Globe generally have opposite signs compared to the difference between PP_Globe and the benchmark data.More specifically, PP_Globe shows high correlations to the GPCP especially in cold seasons (Table S4 in Supporting Information S1) for both SA and EA.PP_Globe overall overestimates annual P in SA with the wet mean bias of +0.95 mm/day, while it shows a small positive mean bias of +0.18 mm/day in EA.In summer, PP_Globe has larger overestimations over most SA regions with a mean bias of +2.53 mm/day, and the difference between PP_Globe and GPCP shows heterogeneous spatial distribution in EA (Figure 2g).Specifically, the difference in East China shows a "north wet-middle cold-south wet" tripolar structure in East China.Large positive and negative differences are found in the ocean regions of EA, while the land regions of EA generally show a small deviation from GPCP.By contrast, TOP_TP shows higher R values of 0.66 and 0.83 respectively in summer and autumn than PP_Globe (0.62 and 0.74, respectively) in SA, while TOP_TP has similar R values with PP_Globe across seasons in EA.TOP_TP overall reduces the wet bias in India (Figures 2d  and 2h).Note that the summer spatial patterns between the difference of TOP_TP and PP_Globe and the difference between of PP_Globe and GPCP are opposite in East China (Figures 2g and 2h), which demonstrates that TOP_TP reduces the summer P biases in East China with the tripolar structure.
Although the differences between TOP_Globe and PP_Globe overall show similar patterns to that between TOP_TP and PP_Globe, there are some large differences especially in India (Figure S13 in Supporting Information S1).These demonstrate that the radiation-topography interactions in non-TP regions also affect the climate of the Asian regions.

Discussion and Conclusions
Radiation-topography interaction plays an important role in regulating regional climate.By using land-atmosphere coupled experiments based on E3SM, we demonstrate that radiation-topography interaction can influence the TP's surface energy balance by reducing  and f sno (Figure 1), which is consistent with the offline ELM simulations (Hao et al., 2021).Such interaction further warms regional near-surface atmosphere, modifies the clouds and affect the local P (Figure 1).Accounting for such interaction in E3SM shows reduced cold biases over the whole TP (Figure 2) especially in winter, which is in line with Lee et al. (2019) and Fan et al. (2019).The interaction is expected to affect the glacier evolution over the TP by accelerating glacier melt and retreat (Tang et al., 2023).Radiation-topography interaction over the TP could further affect the East and South Asian climate.Due to the important role of TP as an "elevated heat pump," the topography-induced albedo change can affect wind and moisture transport over SA and EA, and thus redistribute P (Figure 2) over the Asian regions (Tang et al., 2023).Specifically, the TP's albedo change can affect the intensities and movement of the South Asian High and West Pacific Subtropical High (Tang et al., 2023) through the Rossby wave trains (Wang et al., 2008).Our simulations show that these changes are manifested in the iconic tripolar pattern change in P in East China which is a dominant pattern of P variability in the region.The surface temperature anomalies over the TP have significant impacts on the East Asian summer monsoon precipitation (Diallo et al., 2022).Besides, the snow cover change in winter over the TP is linked to the variation of summer P in the downstream regions of China (Li et al., 2018;You et al., 2020).All of these can contribute to the change of regional P patterns and timing.However, the nonlinear responses of Asian climate to the topography-induced albedo change and associated dominant pathways need further investigations.It is noted that non-TP mountainous regions can also contribute to the change of P patterns in the Asian regions (Figure S13 in Supporting Information S1).
There are still large systemic biases in simulating surface climate over the TP and surrounding regions, despite improved E3SM model performance after accounting for radiation-topography interaction.Such issues have been found in all the ESMs (including E3SM) participating in the "Impact of Initialized Land Surface Temperature and Snowpack on Subseasonal-to-Seasonal Prediction" project (Xue et al., 2021).The large cold and wet biases imply that there are additional important physical processes over the TP but are not well represented or even missing in E3SM and other ESMs.For example, the coupling of convection to the large-scale environment needs to be improved to reduce the P biases in E3SM (Zheng et al., 2019).The T air and P biases in E3SM contribute to the uncertainties in snowpack simulations (Brunke et al., 2021).Besides, the LAP deposition over the TP shows large impacts on the TP's snow cover (Sarangi et al., 2020) and Asian monsoon climate (Qian et al., 2011).However, there is still limited knowledge on snow grain shape and mixing state between LAP and snow grain over the TP, which has been demonstrated to have large impacts on TP's energy balance and water cycle (Hao et al., 2023;He et al., 2018).Better considering the snow-aerosol-radiation interaction is necessary to reduce the uncertainties in simulating climate over the TP and surrounding regions.
Our findings underscore the important regional and teleconnected impacts of radiation-topography interaction over the TP.Improved understanding of the topographic roles stresses the significance of parameterizing such important physical processes in CMIP6 models for future climate projections.Neglecting such interaction will bias the simulations and projections of surface energy balance, snow processes and surface climate over complex terrain and surrounding regions.While this study focuses on the solar radiation-topography interaction, the impacts of the longwave radiation-topography interaction (Feldman et al., 2022) can be investigated in the future.
Further explorations can also be extended to global mountainous regions using the ensemble of multiple ESMs under both satellite phenology and biogeochemistry modes.

Figure 1 .
Figure 1.Regional impacts of radiation-topography interaction on annual (a-d) surface energy balance, (e-f) snow, (g) air temperature (T air ) and (h) precipitation (P) over the Tibetan Plateau.For each panel, the impacts are represented by the difference (∆) between TOP_TP and PP_Globe.

Figure 2 .
Figure 2. Teleconnected impacts of radiation-topography interaction on annual and summer air temperature (T air ) and precipitation (P).Panels (a, c, e, f) are the differences between PP_Globe and Benchmark data sets, and Panels (b, d, f, h) are the differences between TOP_TP and PP_Globe.For each panel, the thick black solid line is the Tibetan Plateau boundary, and the two black dashed lines are the boundaries of South and East Asia regions defined in Section 2.3.
This research was supported by the U.S. DOE, Office of Science, Office of Biological and Environmental Research, Earth System Model Development program area, as part of the Climate Process Team projects.Y.G. was supported by the National Oceanic and Atmospheric Administration under award number NA19OAR4310243, and acknowledges the support by (while serving at) the National Science Foundation.The research was conducted at the Pacific Northwest National Laboratory (PNNL), operated for the DOE by the Battelle Memorial Institute under contract DE-AC05-76RL01830.This research used the computational resources of the National Energy Research Scientific Computing Center, and DOE's Biological and Environmental Research Earth System Modeling program's Compy computing cluster located at PNNL.
Note.The sources of the benchmark data sets are indicated in the second column.The corresponding metrics across seasons are listed in TableS2in Supporting Information S1.Statistical Metrics of Energy Exascale Earth System Model Simulated Annual Surface Energy Balance, Snow Variables, Air Temperature (T air ) and Precipitation (P) Against the Benchmark Data Sets Over the Tibetan Plateau for Both PP_Globe and TOP_TP