Future Projections of summer Tibetan Plateau temperature based on the combined influence of sea surface temperature and soil moisture

Future Projections of summer Tibetan Plateau temperature based on the combined influence of sea surface temperature and soil moisture | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Future Projections of summer Tibetan Plateau temperature based on the combined influence of sea surface temperature and soil moisture Ting Zhang, Ge Liu, Mingkeng Duan, Sulan Nan, Yuhan Feng, Yuwei Zhou, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8673890/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 11 You are reading this latest preprint version Abstract Accurate projections of future warming characteristics in the Tibetan Plateau (TP) are essential for regional ecosystem stability and socioeconomic development and for understanding its role in modulating the Asian Monsoon and broader-scale climate patterns. This study evaluated and selected optimal CMIP6 models by combining an observation-based physical constraint—the synergistic influence of soil moisture anomalies and sea surface temperature anomalies on summer TP surface air temperature (SAT)—with an assessment of SAT spatiotemporal variations. Four multi-model ensemble methods were compared, and the random forest (RF) method was found to most effectively reduce temporal and spatial biases in historical simulations, thus providing a more reliable basis for future projections of summer SAT over the TP. Based on 10 optimal models, the RF-based projections show that SSP1-2.6 scenario, the TP exhibits significant warming trend (0.38°C/10a) in early-term future (2015–2044) but stagnates (0.03°C/10) during the mid- and long-term future (2045–2100). Under the SSP2-4.5 scenario, the TP SAT shows a significant warming trend (0.33°C/10a) during 2015–2070, then slows to 0.14°C/10a. Under the SSP5-8.5 scenario, the TP SAT maintains rapid warming (0.60°C/10a) throughout 2015–2100. The RF-based projections indicate stronger future warming (particularly in western TP) compared to the conventional multi-model ensemble mean (MME) of non-optimized 18 models (MME-18). These results highlight the risk of underestimating TP warming without proper model optimization, warranting particular attention on this vulnerable region. Earth and environmental sciences/Climate sciences Earth and environmental sciences/Environmental sciences Tibetan Plateau CMIP6 projection warming observational constraint Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 1. Introduction As a region highly sensitive to global climate change, the Tibetan Plateau (TP) is experiencing accelerated warming 1 . During the period 1979–2020, the mean annual temperature over the TP increased at 0.34°C per decade, approximately twice the global warming rate 2 . The rapid warming significantly impacts local ecosystem, water resources, agriculture, and infrastructure. Additionally, the TP climate anomalies play a crucial role in modulating the Asian Monsoon and broader-scale climate patterns 2 – 10 . Therefore, studying future TP climate change is critical for both the local and broader climate projection and adaptation. Climate models have been widely applied in simulating historical and projected climate variability 11 – 15 . Based on Coupled Model Intercomparison Project (CMIP) simulations, numerous studies have evaluated and projected climate change over the TP 16–20 . Most findings indicated that the warming trend of the TP will intensify with increasing emissions 19 , 20 . However, significant uncertainties exist in TP simulations and future projections due to amplified model errors caused by complex terrain 21 , 22 . Earlier CMIP simulations exhibited cold biases over the TP 16,23,24 . Although CMIP6 has shown remarkable progress and can more realistically capture the spatial distribution of surface air temperatures (SATs) 25 , many models still exhibit cold biases in the TP 26–29 . Therefore, evaluating CMIP6 model performance is essential for reliable TP climate projections. Current projections of SAT over the TP predominantly rely on multi-model ensemble means (MME), which show considerable uncertainties in estimated warming magnitudes 16 , 17 , 19 , 20 , 30 . Observational constraint methods (e.g., multi-model weighting, attribution-based constraint, emergent constraint) can effectively reduce projection uncertainty 31 . However, few studies have employed such observational constraint methods to improve climate projections for the TP 1,32,,33 . Zhou and Zhang 1 demonstrated that using historical attribution results as constraints leads to the projected TP warming magnitude significantly exceeding the CMIP5 MME estimates, revealing that the latter systematically underestimated the TP’s temperature response to anthropogenic forcing. An ideal model should accurately reproduce observed physical relationships, which can help reduce projection uncertainties 34 ,, 35 . These relationships can help reduce uncertainties in predictions and projections of climate models 36 . For instance, using the observed precipitation-temperature relationship to identify models that correctly simulate land-atmosphere feedback processes, the uncertainties in extreme temperature projections can be effectively reduced 37 , 38 . The tropical temperature-humidity relationship can serve as an observational constraint for evaluating model performance 39 . Using observed relationships as physics-based constraints for model selection can effectively reduce projection uncertainty. Recently, the combined influence of East European-West Siberian soil moisture anomalies and Yellow-Japan Sea SSTAs was found to regulate summer SATs over the eastern TP 40 , offering a potential physics-based constraint. This also raises critical questions: Can CMIP6 models accurately reproduce the combined modulation of the soil moisture anomalies and SSTAs on summer SATs across the entire TP? Can these physics-based observational constraints improve the accuracy of TP SAT projections? These questions warrant investigation to improve the reliability of summer SAT projections over the TP. 2. Data and Methods 2.1 Data This study used the CN05.1 monthly SAT dataset (0.25°×0.25°spatial resolution) obtained from the National Meteorological Information Center to assess the performance of CMIP6 models 41 . This dataset has been extensively applied in climate change studies and model evaluations 42 – 45 . Moreover, the regionally averaged summer SAT over the TP region (73°–105°E, 26°–40°N, above 2,000 m elevation) shows good agreement with the summer TP SAT average from 89 observational stations in the TP, with a correlation coefficient of 0.95 for the period 1961–2014, confirming its reliability. This study also used geopotential heights at 2.5°×2.5° resolution from the National Centers for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR), the ERA-land monthly soil moisture (0–7 cm layer; 0.1°×0.1° resolution) 46 , and the extended reconstructed SST dataset (Version 5; 2.0°×2.0° resolution) obtained from the National Oceanic and Atmospheric Administration (NOAA) 47 . All data were extracted from 1961 to 2014. This study employed model-simulated data from 18 CMIP6 models with topographic data (see the list in Fig. 1 a). The SAT data were topographically adjusted to match the CN05.1 grid topography using atmospheric lapse rates 45 . We used historical simulations (1961–2014) and future projections (2015–2100) under three Shared Socioeconomic Pathway (SSP) scenarios (SSP1-2.6, SSP2-4.5, and SSP5-8.5) 48,49 . All CMIP6 model outputs used in this study represent the first ensemble run (r1i1p1f1) to ensure consistent initial conditions, physical parameterizations, and boundary conditions. The period 1995–2014 was designated as the baseline period for comparative analysis of future climate change. Besides direct evaluation of summer TP SATs, we used the combined effect of the soil moisture anomalies and SSTAs as observational constraints to evaluate the performance of models. As such, we used various model output variables (geopotential heights, SSTs, and surface soil moistures). All model variables were bilinearly interpolated to match observational dataset resolutions. 2.2. Methods The Taylor diagram and Taylor skill score (TSS) were used to quantitatively assess the consistency between simulations and observations 28 , 50 , 51 . The interannual variability skill score (IVS) is used to evaluate the ability of models to reproduce the interannual variability of the summer TP SAT 52 , 53 . We evaluated four multi-model ensemble approaches: the conventional equal-weight MME, Climate model Weighting by Independence and Performance (ClimWIP) 51 , 54 – 57 , Linear Regression (LR) 51 , and Random Forest (RF) 58 . To mitigate overfitting and evaluate generalizability, we train the ensemble models on the period 1961–1994 and validate them on the period 1995–2014, with their robustness and performance assessed using the root mean square error (RMSE) and coefficient of determination (R 2 score). For detailed information on the above evaluation and multi-model ensemble methods, please refer to the supplementary material. 3. Evaluation of CMIP6 models 3.1 Simulation performance for summer SATs over the TP Taylor diagram evaluation (Fig. 1 a) shows that all 18 models reasonably simulate the climatological spatial distribution of summer SATs over the TP. The simulated SAT distributions in most models show high consistency with the observed distribution, with correlation coefficients exceeding 0.80. NorESM2-LM, NorESM2-MM, CESM2-WACCM, ACCESS-CM2, and TaiESM1 demonstrated superior performance in simulating the spatial distribution of SATs (correlation coefficients > 0.95 and TSS > 0.89). In contrast, CanESM5 and CanESM5-1 exhibited relatively poor spatial simulation skills, with lower spatial correlation coefficients and TSS scores below 0.40. The IVS evaluation (Fig. 1 b) shows that most models demonstrated high skills in reproducing the interannual variability (IVS 6.0) (Fig. 1 b), consistent with their poor spatial simulation skills (Fig. 1 a). Based on high spatial similarity with the observed pattern of TP SAT (TSS ≥ 0.8) and strong simulation capability for the interannual variation of TP SAT (IVS ≤ 0.2), we identified 12 optimal models: ACCESS-CM2, ACCESS-ESM1-5, AWI-CM-1-1-MR, BCC-CSM2-MR, CESM2-WACCM, CMCC-CM2-SR5, CMCC-ESM2, MPI-ESM1-2-HR, MRI-ESM2-0, NorESM2-LM, NorESM2-MM, and TaiESM1. 3.2 Evaluation by using physics-based observational constraints The soil moisture anomalies in the East European Plain-West Siberian Plain and SSTAs in the Yellow Sea-Japan Sea jointly modulate multiscale variations of summer SATs over the eastern TP 40 . Specifically, soil moisture anomalies in the East European Plain-West Siberian Plain modulate the overlying upper-tropospheric geopotential heights via local land-air interactions, triggering a downstream Rossby wave train. This wave train generates a high-pressure anomaly extending from the eastern Tibetan Plateau to the Yellow and Japan Seas, which reflects an intensified and eastward-extended South Asian High. The SSTAs reinforce this high-pressure anomaly through local thermal forcing and ultimately cause summer SAT anomalies over the eastern TP. In short, the combined effect of these soil moisture and SST anomalies leads to summer SAT anomalies over the eastern TP 40 . Since this study focuses on the evaluation and projection of summer SATs over the entire TP, we calculated the TP regionally averaged SAT time series for the period 1961–2014 and used it to perform regression analyses on 200 hPa geopotential heights (Fig. 2 a), soil moistures (Fig. 2 b), and SSTAs (Fig. 2 c). All variables in Fig. 2 have been detrended, since we expect CMIP6 models to not only capture the simultaneous variation trends of soil moistures, SSTs, and TP SAT, but also to accurately reproduce the multiscale modulation of soil moisture anomalies and SSTAs on TP SAT during the historical period (1961–2014). Corresponding to higher TP SAT, a positive-negative-positive wave train extends from eastern Europe to the TP-Japan Sea region (Fig. 2 a). Significant negative soil moisture anomalies appear in the East European Plain-West Siberian Plain (Fig. 2 b), and significantly positive SSTAs occur in the Yellow Sea-Japan Sea (Fig. 2 c). These anomalies can excite and maintain the wave train from eastern Europe to the TP-Japan Sea region, leading to positive geopotential height anomalies over the TP (reflecting a stronger and northeastward South Asian high) and thereby causing higher summer TP SAT. These results are consistent with Zhang et al 40 . The 200-hPa geopotential height, soil moisture, and SST anomalies regressed upon the raw (i.e., non-detrended) TP SAT time series show similar spatial patterns to those in Fig. 2 a–c, but with higher significance levels in the key regions (Fig.omitted). The interdecadal variability and trend of the observed TP SAT are successfully captured by a fitting model based on the raw soil moisture index (East European–West Siberian Plain) and the raw SST index (Yellow Sea–Japan Sea) (Fig. 2 d). The correlation coefficient between the low-pass-filtered observed and reconstructed time series reaches 0.95, passing a Monte Carlo test at the 99% confidence level. This indicates that the soil moisture anomalies and SSTAs in these key regions collectively regulate the interdecadal variability and trend of the SAT across the entire TP. This collective modulation mechanism can be used as an observational constraint. To assess whether CMIP6 models can reproduce the multiscale modulation of soil moisture anomalies and SSTAs on TP SAT during the historical period (1961–2014), all variables in the regression analyses of 200 hPa geopotential heights (Fig. 3 ), soil moistures (Fig. 4 ), and SSTAs (Fig. 5 ) have been detrended. In these Figs., NESM3 and AWI-CM-1-1-MR were excluded since they do not provide historical simulated soil moisture data. Most models successfully simulated the significant positive geopotential height anomalies over the TP and accurately reproduced the wave train propagating from Europe to the TP region. However, CanESM5-1 (Fig. 3 g) and CanESM5 (Fig. 3 h) showed no clear wave train. TaiESM1 (Fig. 3 p) even showed no significant positive geopotential height anomaly over the TP. We also examined soil moisture anomalies regressed against the TP SAT time series for each model (Fig. 4 ). Most models capture significant negative soil moisture anomalies around Europe and Siberia, although the locations differ somewhat from observations (Fig. 2 b). Furthermore, the negative soil moisture anomalies generally correspond well with the positive geopotential height anomalies in the overlying troposphere. For instance, in ACCESS-ESM1-5 (Fig. 4 b), the location of significant negative soil moisture anomalies resembles that in the observation (Fig. 2 b). Correspondingly, a positive geopotential height anomaly appears over eastern Europe (Fig. 3 b), with a location highly consistent with the observation (Fig. 2 a). However, in BCC-CSM2-MR (Fig. 4 c) and NorESM2-MM (Fig. 4 o), significant negative soil moisture anomalies shift northwestward compared to the observations (Fig. 2 b). As such, the center of significant positive geopotential height anomalies also shifts northwestward (Fig. 3 c,o), aligning with the locations of soil moisture anomalies in the two models. Despite slight location differences, most models still reproduce the wave trains from Europe to the TP, which contribute to the summer TP SAT anomaly. However, MPI-ESM1-2-LR showed very weak, insignificant negative soil moisture anomalies over Europe (Fig. 4 l). Correspondingly, no significant positive geopotential height anomalies appear over this region (Fig. 3 l). Most models successfully simulate significant positive SSTAs in the Yellow Sea–Japan Sea region associated with higher TP SAT, which is consistent with observations (Fig. 2 c). CanESM5-1, MIROC6, and MPI-ESM1-2-LR showed weak and insignificant SSTAs in this region. Based on physics-based observational constraints (i.e., the combined effect of soil moisture anomalies and SSTAs), we identified 11 optimal models: ACCESS-CM2, ACCESS-ESM1-5, BCC-CSM2-MR, CESM2-WACCM, CMCC-CM2-SR5, CMCC-ESM2, FGOALS-g3, MPI-ESM1-2-HR, MRI-ESM2-0, NorESM2-LM, and NorESM2-MM. We also investigated the 200-hPa geopotential height, soil moisture, and SST anomalies regressed upon the raw (i.e., non-detrended) TP SAT time series (see supplementary Figs. 1–3). The results reveal that these physically better-performing models show similar spatial patterns to those in Figs. 3 – 5 . In contrast, the physically poorer-performing models (e.g., CanESM5-1, CanESM5, and TaiESM1) fail to capture the wave train. Ten of the 11 models align with the selection based on the TSS and IVS criteria, with only FGOALS-g3 having a relatively low TSS. This result implies that the models reproducing the combined effects of soil moisture anomalies and SSTAs on summer TP SAT tend to more accurately simulate the TP SAT (achieving high Taylor and IVS scores). In contrast, those that fail to reproduce the physical link perform poorly. This mutual verification between physics-based constraints and Taylor and IVS scores enhances model credibility. Therefore, based on all these criteria, we ultimately identified 10 optimal models: ACCESS-CM2, ACCESS-ESM1-5, BCC-CSM2-MR, CESM2-WACCM, CMCC-CM2-SR5, CMCC-ESM2, MPI-ESM1-2-HR, MRI-ESM2-0, NorESM2-LM, and NorESM2-MM. Unless otherwise specified, all subsequent multi-model ensemble analyses are derived from the 10 optimally selected models. 3.3 Evaluation of multi-model ensemble methods We use different ensemble methods to establish the base or regression models during the training phase (1961–1994) and assess their performance during the validation phase (1995–2014). Relative to the other methods, the RF approach demonstrates the best agreement with the observation, with the lowest RMSE (0.69) and highest R 2 (0.96) (Table 1 ). The LR method is the second best (RMSE = 0.96, R 2 = 0.92) (Table 1 ). The MME method performs the worst. Nevertheless, the MME of the 10 optimal models (MME-10) outperforms the MME of 18 models (MME-18), with a lower RMSE (1.16) and a higher R 2 (0.91), demonstrating the effectiveness of model selection. Table 1 The assessment results of multi-model ensemble methods. MME-18 MME-10 LR ClimWIP RF R 2 0.85 0.91 0.92 0.92 0.96 RMSE 1.59 1.16 0.96 1.07 0.69 All multi-model ensemble methods successfully capture the spatial distribution of climatological mean summer SATs over the TP (Fig. 6 b, d, f, h, and j), showing a high consistency with the observation (Fig. 6 a). However, the differences between the multi-model ensemble and observed SATs display substantial discrepancies (Fig. 6 c, e, g, i, and k). The MME-18 exhibits the strongest and most extensive cold biases in the western TP (Fig. 6 c.) For the MME-10, the cold biases in the western TP have been to some extent decreased (Fig. 6 e), verifying model selection validity. Among all ensemble methods, the RF yields the best agreement with the observed SATs, with smaller, scatter cold/warm biases rather than large-scale systematic biases (Fig. 6 k). The summer TP SAT has increased significantly since the 1990s (Fig. 2 d). The spatial distribution of the difference in summer SATs between the 1961–1994 and 1995–2014 periods (the latter minus the former) shows that the entire TP experienced warming (Fig. 7 a). Although historical simulations based on different multi-model ensemble methods successfully capture the warming over most of the TP, they incorrectly produce cooling in the western and southern TP and considerably overestimate the warming in the northern TP. This discrepancy is most pronounced for the MME-18 (Fig. 7 b). The MME-10 also erroneously simulated cooling in western and southern TP but with reduced magnitude relative to the MME-18 (Fig. 7 c), indicating improvement through optimal selection. The RF method shows the best agreement with the observed warming over the entire TP, exhibiting only weak, scattered cooling in the western and southern TP and the slightest overestimation of warming over the northern TP (Fig. 7 f). In summary, using the mutual verification between physics-based observational constraints and Taylor and IVS scores to optimally select 10 CMIP6 models, and then applying the RF method for multi-model ensemble, the outputs can effectively reduce historical simulation biases, and therefore are expected to provide more reliable future projections of summer TP SATs. 4. Projected variations in summer TP SATs under different emission scenarios We employed the RF method with 10 optimal models to project and analyze the characteristics of summer SAT variation over the TP under the SSP1-2.6, SSP2-4.5, and SSP5-8.5 (Fig. 8 ). Under SSP1-2.6, the warming in the TP primarily occurs in the early period. From 2015 to 2044, the summer SAT exhibits a significant upward trend, with a warming rate of approximately 0.38°C/10a (exceeding the 99% confidence level) (Fig. 8 a). The warming rate is higher than the historical rate (0.25 ℃/10a for 1961 − 2014). In the mid- to long-term future (2045–2100), the TP SAT shows no clear trend (0.03 ℃/10a, insignificant). Compared to the RF-based projection, the MME-18 projection shows a systematic cold bias (Fig. 8 a), indicating that the latter significantly underestimates the future warming of the TP. Given that the RF-based outputs appear more reliable, the above results suggest that the future warming over the TP may be more severe than the MME-18 projection. Under SSP2-4.5, the TP SAT exhibits a distinct upward trend during 2015–2070 (approximately 0.33°C/10a, exceeding the 99% confidence level) (Fig. 8 b). During the long-term future (2071–2100), the warming persists but decelerates (approximately 0.14°C/10a). The MME-18-based projection exhibits a clear cold bias during 2015–2070 (approximately 0.39°C lower than the RF-based projection). During 2071–2100, the two projections tend to be more consistent. Under SSP5-8.5, the TP SAT shows a more pronounced and sustained upward trend throughout 2015–2100 (0.60°C/10a) (Fig. 8 c). The MME-18 also exhibits a systematic cold bias under this scenario (approximately 0.34°C lower than the RF-based projection) (Fig. 8 c). This discrepancy diminishes to a negligible level during 2071–2100 (approximately 0.02°C difference). Figure 9 displays the spatial distribution of summer SAT anomalies for the early- (2025–2044), mid- (2055–2074), and long-term (2081–2100) future under different scenarios. The RF-based projections show that relative to the baseline period (1995–2014), the SSP1-2.6 scenario exhibits predominant warming across most TP regions in the early future, with temperature increases of 1–2°C (Fig. 9 a). In mid- and long-term future, the increase reaches 2–3°C in these regions (Fig. 9 b,c). In contrast, the MME-18-based projections show relatively weaker warming in central and eastern TP and even cooling in western TP under the SSP1-2.6 scenario (Fig. 9 j-l). A similar west-cooling-east-warming pattern appears under SSP2-4.5 and SSP5-8.5 (Fig. 9 m–q), except during the long-term future under SSP5-8.5 (Fig. 9 r). Compared to the MME-18, the RF-based projections demonstrate stronger warming, particularly in western TP, resulting in more spatially homogeneous warming throughout the entire TP (Fig. 9 d–i). These findings indicate that the RF-based projections primarily reduce the cold bias in western TP. We further investigate the warming magnitude in the western TP (west of 85° E) and find that the RF-based projections show a much more dramatic warming compared to the MME-18 (Fig. 10 ). Under SSP1-2.6, the western TP SAT in the RF-based projections shows persistent positive anomalies relative to the 1995–2014 baseline, with mean warming reaching 1.1°C during 2045–2100 (Fig. 10 a). In contrast, the MME-18-based projections are almost always lower than this baseline (blue line in Fig. 10 a). Under SSP2-4.5, the RF-based projections also show persistent warming (mean 1.66°C during 2071–2100) (Fig. 10 b). The MME-18-based projections remain below the baseline until the late 2040s, showing only modest mean warming of 0.94°C in 2071–2100 (Fig. 10 b). Under SSP2-8.5, the MME-18-based projections do not exceed the observed baseline until the early 2040s (blue line in Fig. 10 c), while the RF-based projections exhibit a more rapid warming trend (Fig. 10 c). Given that western TP has already experienced significant warming, future cooling under global warming seems unlikely. Therefore, the RF-based projections are more reliable. These results suggest that future warming in the western TP may be more severe than the MME-18-based projections. 5. Conclusions and discussion Accurately projecting future TP warming is crucial for local ecological and socioeconomic sustainability. This study evaluated CMIP6 models by combining direct assessment of TP SAT variations with physics-based observational constraints derived from the combined influence of soil moisture anomalies and SSTAs on summer TP SAT. Optimal models and ensemble methods were identified to analyze future changes under different scenarios. The main conclusions are as follows: (1) Using mutual verification combining physics-based observational constraints and statistical skill assessment (Taylor and IVS scores), we identified 10 optimal models. Among four model ensemble methods (MME, ClimWIP, LR, and RF), we found that the RF method effectively reduced the cold bias and accurately simulated the climatological spatial distribution and multi-scale (decadal and trend) temporal variations of summer TP SATs. (2) The RF-based projections with the optimal models show that significant summer TP warming, with its magnitude and evolution strongly dependent on the emission scenario. Under SSP1-2.6, the TP exhibits significant early-term warming (0.38°C/decade during 2015–2044), with no clear trend thereafter. Under SSP2-4.5, the TP SAT shows a significant warming trend (0.33°C/10a) during 2015–2070, and a slower warming trend (0.14°C/10a) afterwards. Under SSP5-8.5, the TP SAT shows a significant warming trend (0.60°C/10a) throughout 2015–2100. The non-optimized MME-18-based projections systematically underestimated the TP warming, especially over western TP. The constrained projections suggest future warming, particularly in the vulnerable western TP, may be more severe than commonly projected. Using optimal fingerprinting, Zhou and Zhang 1 attributed historical warming over the TP to anthropogenic activities, particularly greenhouse gas forcing. Using the attribution results as observational constraints, the predicted TP warming may exceed the CMIP5 multi-model ensemble 1 . Although the constraint conditions in this study differ from Zhou and Zhang 1 , the conclusions are consistent, further supporting the likelihood of more pronounced future TP warming. Sustained and significant warming in the TP would likely increase local soil temperatures, accelerate glacier melt 18 , 59 , alter permafrost 60 , 61 , enhance evaporation, reduce soil moisture, and exacerbate grassland degradation and desertification 62 – 64 . Summer snow cover in the high-altitude western TP can influence summer precipitation anomalies over eastern China 65 , 66 . Under stronger warming, reduced or vanished snow cover could significantly alter climate anomalies in eastern China, which demands particular attention. Additionally, European soil moisture anomalies play a critical regulating role in summer TP SAT. If the regulatory mechanism and its underlying physical constraint remain intact, future extreme warming could be somewhat mitigated. This suggests that multinational and multi-regional cooperation is essential to address climate change. To enhance climate resilience, regional climate impacts and feedback effects should be integrated into global climate governance agendas. Declarations Author Contribution T.Z. and G.L. conceived the study. T.Z. and G.L. performed the analyses and interpreted the data. T.Z., G.L.and M.D. led the writing with input from all co-authors. Acknowledgements This work was jointly sponsored by the National Key Research and Development Program of China (Grant 2023YFF0805300), the Major Science and Technology project of the Xizang Autonomous Region (Grant XZ202402ZD0006), the Youth Innovation Team of China Meteorological Administration “Climate change and its impact in the Tibetan Plateau” (Grant CMA2023QN16), and the Basic Research Fund of CAMS (Grant 2023Z024). We thank the National Tibetan Plateau Data Center for providing the Tibetan Plateau boundary dataset ( http://data.tpdc.ac.cn ). Data Availability The CMIP6 model data is available from https://aims2.llnl.gov/search/cmip6/. References Zhou, T. & Zhang, W. Anthropogenic warming of Tibetan Plateau and constrained future projection. Environ. Res. Lett. 16, 044039 (2021). You, Q. et al. Warming amplification over the Arctic Pole and Third Pole, Trends, mechanisms and consequences. Earth Sci. Rev. 217, 103625 (2021). Duan, A. & Wu, G. Role of the Tibetan Plateau thermal forcing in the summer climate patterns over subtropical Asia. Clim. Dyn. 24, 793–807 (2005). Zhou, X., Zhao, P., Chen, J., Chen, L. & Li, W. Impacts of thermodynamic processes over the Tibetan Plateau on the Northern Hemispheric climate (in Chinese). Sci China Ser D-Earth Sci. 39(11), 1473–1486 (2009). Wu, G., Zhuo, H., Wang, Z. & Liu, Y. Two types of summertime heating over the Asian large-scale orography and excitation of potential-vorticity forcing, I. Over Tibetan Plateau. Sci. China Earth Sci. 59, 1996–2008 (2016). Liu, G., Zhao, P., Nan, S., Chen, J. & Wang, H. Advances in the study of linkage between the Tibetan Plateau thermal anomaly and atmospheric circulations over its upstream and downstream regions (in Chinese). Acta Meteorol Sin. 76(6), 861–869 (2018). Lu, M. et al. Possible effect of the Tibetan Plateau on the “upstream” climate over West Asia, North Africa, South Europe and the North Atlantic. Clim. Dyn. 51, 1485–1498 (2018). Liu, Y. et al. Land-atmosphere-ocean coupling associated with the Tibetan Plateau and its climate impacts. Natl. Sci. Rev. 7(3), 534–552 (2020). Nan, S., Zhao, P., Chen, J. & Liu, G. Links between the thermal condition of the Tibetan Plateau in summer and atmospheric circulation and climate anomalies over the Eurasian continent. Atmos. Res. 247, 105212 (2020). Huang, J. et al. Global climate impacts of land-surface and atmospheric processes over the Tibetan Plateau. Rev. Geophy. 61(3), 000771 (2023). He, S., Yang, J., Bao, Q., Wang, L. & Wang, B. Fidelity of the observational /reanalysis datasets and global climate models in representation of extreme precipitation in East China. J. Clim. 32(1), 195–212 (2019). Yu, E. & Sun, J. Extreme temperature projection over northwestern China based on multiple regional climate models (in Chinese). Trans Atmos Sci. 42(1), 46–57 (2019). Khan, A., Koch, M. & Tahir, A. A. Impacts of climate change on the water availability, seasonality and extremes in the upper Indus Basin (UIB). Sustainability. 12(4), 1283 (2020). Jiang, W. & Chen, H. Assessment and projection of changes in temperature extremes over the mid-high latitudes of Asia based on CMIP6 models (in Chinese). Trans Atmos Sci. 44(4), 592–603 (2021). Xu, R., Liang, X. & Duan, M. Evaluation of CWRF simulation of temperature and precipitation on the Qinghai-Tibet Plateau (in Chinese). Trans Atmos Sci. 44(1), 104–117 (2021). Xu, Y. & Xu, C. Preliminary assessment of simulations of climate changes over China by CMIP5 multi-models. Atmos. Oceanic Sci Lett. 5(6), 489–494 (2012). Hu, Q., Jiang, D. & Fan, G. Climate change projection on the Tibetan Plateau, Results of CMIP5 models (in Chinese). Chinese J. Atmos Sci. 39(2), 260–270 (2015). Yang, M., Wang, X., Pang, G., Wang, G. & Liu, Z. The Tibetan Plateau cryosphere, observations and model simulations for current status and recent changes. Earth Sci. Rev. 190, 353–369 (2019). Zhou, T. et al. The near-term, mid-term and long-term projections of temperature and precipitation changes over the Tibetan Plateau and the sources of uncertainties (in Chinese). J. Meteorol. Res. 40(5), 697–710 (2020). Zhang, J. et al. CMIP6 evaluation and projection of climate change in Tibetan Plateau (in Chinese). Journal of Beijing Normal University (Natural Science). 58(1), 77–88 (2022). Wang, S. & Xiong, Z. The Preliminary Analysis of 5 Coupled Ocean-Atmosphere Global Climate Models Simulation of Regional Climate in Asia (in Chinese). Climatic Environ Res. 9(2), 240 (2004). Ding, Y. et al. Detection, causes and projection of climate change over China, An overview of recent progress. Adv. Atmos. Sci. 24, 954–971 (2007). Jiang, D., Wang, H. & Lang, X. Evaluation of East Asian climatology as simulated by seven coupled models. Adv. Atmos. Sci. 22, 479–495 (2005). Jiang, D., Tian, Z. & Lang, X. Reliability of climate models for China through the IPCC Third to Fifth Assessment Reports. INT J CLIMATOL. 36(3), 1114–1133 (2016). Zhou, T., Zou, L. & Chen, X. Commentary on the Coupled Model Intercomparison Project Phase 6 (CMIP6) (in Chinese). ADV CLIM CHANG RES. 15 (5), 445–456 (2019). Su, F., Duan, X., Chen, D., Hao, Z. & Guo, L. Evaluation of the global climate models in the CMIP5 over the Tibetan Plateau. J. Clim. 26(10), 3187–3208 (2013). Guo, D. L., Sun, J. & Yu, E. Evaluation of CORDEX regional climate models in simulating temperature and precipitation over the Tibetan Plateau. Atmos. Oceanic Sci Lett. 11(3), 219–227 (2018). Zhu, Y. & Yang, S. Evaluation of CMIP6 for historical temperature and precipitation over the Tibetan Plateau and its comparison with CMIP5. Adv. Clim. Chang. Res. 11(3), 239–251 (2020). Chen, W., Jiang, D. & Wang, X. Evaluation and Projection of CMIP6 Models for Climate over the QinghaiXizang (Tibetan) Plateau (in Chinese). Plateau Meteorol. 40(6), 1455–1469 (2021). Yang, Y., Dai, X., Tong, H. & Zhang, B. CMIP5 Model Precipitation Bias-correction Methods and Projected China Precipitation for the Next 30 Years (in Chinese). Climatic Environ Res. 24(6), 769–784 (2019). Zhou, B. & Zhai, P. The constraint methods for projection in the IPCC Sixth Assessment Report on climate change (in Chinese). Acta Meteorol. Sin. 9(6),1063–1070 (2021). Zhao, Y., Zhou, T., Zhang, W. & Li, J. Change in Precipitation over the Tibetan Plateau Projected by Weighted CMIP6 Models. Adv. Atmos. Sci. 39, 1133–1150 (2022). Qiu, H., Zhou, T., Chen, X., Wu, B. & Jiang, J. Understanding the diversity of CMIP6 models in the projection of precipitation over Tibetan Plateau. Geophys. Res. Lett. 51(3), 106553 (2024). Vrac, M. Multivariate bias adjustment of high-dimensional climate simulations, the Rank Resampling for Distributions and Dependences (R2D2) bias correction. Hydrol. Earth Syst. Sci. 22, 3175–3196 (2018). Villalobos-Herrera, R. et al. Towards a compound-event-oriented climate model evaluation, a decomposition of the underlying biases in multivariate fire and heat stress hazards. Nat Hazard Earth Sys. 21(6), 1867–1885 (2021). Orbe, C. GISS Model E2.2, A Climate Model Optimized for the Middle Atmosphere-2. Validation of Large-Scale Transport and Evaluation of Climate Response. J Geophys Res-atmos. 125(24), 033151 (2020). Donat, M., Pitman, A. & Angélil, O. Understanding and reducing future uncertainty in midlatitude daily heat extremes via land surface feedback constraints. Geophys. Res. Lett. 45(19), 10,627 – 10,636 (2018). Vogel, M. M., Zscheischler, J. & Seneviratne, S. I. Varying soil moisture–atmosphere feedbacks explain divergent temperature extremes and precipitation projections in central Europe. Earth Syst. Dyn. 9(3), 1107–1125 (2018). Zhang, Y., Held, I. & Fueglistaler, S. Projections of tropical heat stress constrained by atmospheric dynamics. NAT GEOSCI. 14(3),133–137 (2021). Zhang, T. et al. Synergistic contribution of soil moisture and sea surface temperature to summer Tibetan Plateau temperature. Atmos. Res. 314,107811 (2025). Wu, J. & Gao, X. A gridded daily observation dataset over China region and comparison with the other datasets. Chinese J Geophys-ch. 56(4), 1102–1111 (2013). Guo, D. & Wang, H. Erratum to: comparison of a very-fine-resolution GCM with RCM dynamical downscaling in simulating climate in China. Adv Atmos Sci. 33(6), 794 (2016). Zhou, B., Xu, Y., Wu, J., Dong, S. & Shi, Y. Changes in temperature and precipitation extreme indices over China: analysis of a high-resolution grid dataset. Int J Climatol. 36(3), 1051–1066 (2016). Gao, Y., Xiao, L., Chen, D., Xu, J. & Zhang, H. Comparison between past and future extreme precipitations simulated by global and regional climate models over the Tibetan Plateau. Int J Climatol. 38(3), 1285–1297 (2018). Yang, K., Guo, D., Hua, W., Ma, D. & Xing, Y. Evaluation and projection of CMIP6 HighResMIP in simulating surface air temperature and precipitation over the Tibetan Plateau (in Chinese). Trans Atmos Sci. 46(2), 193–204 (2023). Muñoz-Sabater, J. et al. ERA5-Land, A state-of-the-art global reanalysis dataset for land applications. Earth Syst. Sci. Data. 13 (9), 4349–4383 (2021). Huang, B. et al. Extended Reconstructed Sea Surface Temperature, version 5 (ERSST.v5), upgrades, validations and Intercomparisons. J. Climate. 30(20), 8179–8205 (2017). Eyring, V. et al. Overview of the Coupled Model Intercomparison Project Phase 6 (CMIP6) experimental design and organization. Geosci. Model Dev. 9, 1937–1958 (2016). O'Neill, B.C., Tebaldi, C., Vuuren, D.P. & Eyring, V. The scenario model intercomparison project (ScenarioMIP) for CMIP6. Geosci. Model Dev. 9(9), 3461–82 (2016). Taylor, K.E. Summarizing multiple aspects of model performance in a single diagram. J. Geophys. Res. Atmos. 106(D7), 7183–7192 (2001). Li, T., Jiang, Z. & Treut, H. L. Machine learning to optimize climate projection over China with multi-model ensemble simulations. Environ. Res. Lett. 16(9), 094028 (2021). Scherrer, S. C. Present-day interannual variability of surface climate in CMIP3 models and its relation to future warming. Int. J. Climatol. 31, 1518–1529 (2011). Peng, Y. et al. Observational constraint on the future projection of temperature in winter over the Tibetan Plateau in CMIP6 models. Environ. Res. Lett. 17, 034023 (2022). Chen, W., Jiang, Z. & Li, L. Probabilistic Projections of Climate Change over China under the SRES A1B Scenario Using 28 AOGCMs. J. Climate. 24(17), 4741–4756 (2011). Knutti, R. et al. A climate model projection weighting scheme accounting for performance and interdependence. Geophys Res Lett. 44(14), 1909–1918 (2017). Sanderson, B.M., Knutti, R. & Caldwell, P. Addressing interdependency in a multimodel ensemble by interpolation of model properties. J. Clim. 28(13), 5150–5170 (2015). Sanderson, B. M., Wehner, M. & Knutti, R. Skill and independence weighting for multi-model assessments. Geosci. Model Dev. 10, 2379–2395 (2017). Liaw, A. & Wiener, M. Classification and regression by Random Forest. R News. 2 (3), 18–22 (2002). Oku, Y., Ishikawa, H., Haginoya, S. & Ma, Y. Recent trends in land surface temperature on the Tibetan Plateau. J. Clim. 19, 2995–3003 (2006). Yan, C., Song, X., Zhou, Y., Duan, H. & Li, S. Assessment of aeolian desertification trends from 1975's to 2005's in the watershed of the Longyangxia reservoir in the upper reaches of China's Yellow River. Geomorphology. 112(3), 205–211 (2009). Cheng, G. & Jin, J. Permafrost and groundwater on the Qinghai-Tibet plateau and in northeast China. Hydrogeol. J. 21(1), 5–23 (2013). Li, N., Wang, G., Liu, G., Lin, Y. & Sun, X. The ecological implications of land use change in the source regions of the Yangtze and Yellow Rivers, China. Reg. Environ. Change. 13(5),1099–1108 (2013). Hu, G. et al. Holocene aeolian activity in the headwater region of the Yellow River, northeast Tibet Plateau, China, a first approach by using OSL-dating. Catena. 149, 150–157 (2017). Wang, Y. et al. Changes in mean and extreme temperature and precipitation over the arid region of northwestern China, observation and projection. Adv. Atmos. Sci. 34(3), 289–305 (2017). Liu, G., Wu, R. & Zhang, Y. Persistence of snow cover anomaly over the Tibetan Plateau and implication for forecast of summer precipitation over the Meiyu-Baiu region. Atmos. Oceanic Sci. Lett. 7,115–119 (2014). Liu, G., Wu, R., Zhang, Y. & Nan, S. The summer snow cover anomaly over the Tibetan Plateau and its association with simultaneous precipitation over the mei-yu-baiu region. Adv. Atmos. Sci. 31,755–764 (2014). Additional Declarations No competing interests reported. Supplementary Files SupplementaryMaterialsnew20260122.doc Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 09 Feb, 2026 Reviews received at journal 09 Feb, 2026 Reviews received at journal 05 Feb, 2026 Reviews received at journal 03 Feb, 2026 Reviewers agreed at journal 28 Jan, 2026 Reviewers agreed at journal 27 Jan, 2026 Reviewers agreed at journal 27 Jan, 2026 Reviewers invited by journal 27 Jan, 2026 Editor assigned by journal 26 Jan, 2026 Submission checks completed at journal 26 Jan, 2026 First submitted to journal 22 Jan, 2026 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8673890","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":582082617,"identity":"dfbcd0db-e6ff-4800-b159-e6db0bf94015","order_by":0,"name":"Ting Zhang","email":"","orcid":"","institution":"Chinese Academy of Meteorological Sciences","correspondingAuthor":false,"prefix":"","firstName":"Ting","middleName":"","lastName":"Zhang","suffix":""},{"id":582082618,"identity":"7d80390f-9da5-47e5-9b12-c45256c6dea8","order_by":1,"name":"Ge Liu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAuElEQVRIiWNgGAWjYBACPmYQWUGKFjawljMkaQERjG0kaWHnMX7xcZ5dNH8D+8XHPAx2eUQ4jMfMcua25NwZB3iKjXkYkouJ0mLMu+1AbsMBnjTJGQwHEhuI0vJ3zoHc+aRoMX7M2HAgd8MB9mMSH4jTwlbG2HMsOXfjYR5mgw8GyYS18PMf3vzhR41d7rzj7Q8fJFTYEdYCskgCTDHzGDAwGBChHqT2A4Rmf0Cc+lEwCkbBKBhxAAADdzbeBhQrSAAAAABJRU5ErkJggg==","orcid":"","institution":"Chinese Academy of Meteorological Sciences","correspondingAuthor":true,"prefix":"","firstName":"Ge","middleName":"","lastName":"Liu","suffix":""},{"id":582082619,"identity":"150a5180-98b6-46a0-9bee-f0572a91c6ab","order_by":2,"name":"Mingkeng Duan","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Mingkeng","middleName":"","lastName":"Duan","suffix":""},{"id":582082620,"identity":"a6701b7f-575d-481f-8a10-3c5a3df855fe","order_by":3,"name":"Sulan Nan","email":"","orcid":"","institution":"Chinese Academy of Meteorological Sciences","correspondingAuthor":false,"prefix":"","firstName":"Sulan","middleName":"","lastName":"Nan","suffix":""},{"id":582082621,"identity":"8165c7f1-0d04-4366-8362-cc32b8635c2e","order_by":4,"name":"Yuhan Feng","email":"","orcid":"","institution":"Chinese Academy of Meteorological Sciences","correspondingAuthor":false,"prefix":"","firstName":"Yuhan","middleName":"","lastName":"Feng","suffix":""},{"id":582082622,"identity":"e69a7f6b-b419-4617-9c6c-067c31f599ee","order_by":5,"name":"Yuwei Zhou","email":"","orcid":"","institution":"Chinese Academy of Meteorological Sciences","correspondingAuthor":false,"prefix":"","firstName":"Yuwei","middleName":"","lastName":"Zhou","suffix":""},{"id":582082623,"identity":"2d806994-aa44-487a-9f63-69573d4321e3","order_by":6,"name":"Hancheng Zou","email":"","orcid":"","institution":"Chinese Academy of Meteorological Sciences","correspondingAuthor":false,"prefix":"","firstName":"Hancheng","middleName":"","lastName":"Zou","suffix":""}],"badges":[],"createdAt":"2026-01-23 01:38:22","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8673890/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8673890/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":101396268,"identity":"68b6013f-5fcc-4369-9630-7fc31757bd94","added_by":"auto","created_at":"2026-01-29 09:13:47","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":302784,"visible":true,"origin":"","legend":"\u003cp\u003e(a) Taylor diagrams for the evaluation of 18 CMIP6 models regarding climatological spatial distribution of summer SATs over the TP during the historical period (1961–2014), (b) Interannual variability skill score (IVS) of the 18 models.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8673890/v1/454a28d8678ec36c1a465550.png"},{"id":101398185,"identity":"4a155b72-452a-4389-85e2-12e7e9632216","added_by":"auto","created_at":"2026-01-29 09:39:59","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":837177,"visible":true,"origin":"","legend":"\u003cp\u003eAnomalies of summer (a) 200-hPa geopotential heights (color shadings, units: gpm), (b) soil moisture anomalies (units: %), and (c) SSTAs (units: ℃) regressed against the concurrent TP SAT index during the period 1961–2014. (d) Standardized time series of the observed (red line) and fitted (green line) summer TP SAT index after 9-year low-pass filtering, where the fitting model is constructed using the raw soil moisture index (East European–West Siberian Plain) and the raw SST index (Yellow Sea–Japan Sea). In (c), the green box denotes the Yellow Sea-Japan Sea region. The anomalies significant at the 95% confidence level are stippled. The green contours denote the TP (3000 m above sea level).\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8673890/v1/b4d371fc95ea8e73e1b840ce.png"},{"id":101396264,"identity":"8aea8d52-2c3f-4762-94bd-3f892563faa0","added_by":"auto","created_at":"2026-01-29 09:13:46","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":1431758,"visible":true,"origin":"","legend":"\u003cp\u003eAnomalies of summer 200-hPa geopotential heights (color shadings, units: gpm) regressed against the concurrent TP SAT index, based on the CMIP6 models during 1961–2014. The anomalies significant at the 95% confidence level are stippled. The green contours denote the TP (3000 m above sea level).\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8673890/v1/4194ccda3f76aa843c604fed.png"},{"id":101398195,"identity":"764bccda-dd6b-4e86-bb79-a3ed6fe8384f","added_by":"auto","created_at":"2026-01-29 09:40:05","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":1203716,"visible":true,"origin":"","legend":"\u003cp\u003eAs in Fig.3, but for summer soil moisture anomalies.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8673890/v1/48da6e9f61653d3c72a853ed.png"},{"id":101396259,"identity":"7555d2f8-b3b9-4c5c-b7d1-d1bd666ca455","added_by":"auto","created_at":"2026-01-29 09:13:46","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":961260,"visible":true,"origin":"","legend":"\u003cp\u003eAs in Fig. 3, but for summer SSTAs. The green box denotes the Yellow Sea–Japan Sea region.\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8673890/v1/37291fff9f2938537a925672.png"},{"id":101398805,"identity":"843bc4ed-f6eb-4a9d-95b8-7aae0b2c8e28","added_by":"auto","created_at":"2026-01-29 09:47:44","extension":"jpeg","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1527105,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial distribution of climatological mean summer SATs over the TP 1995–2014 for (a) observation, (b) MME-18, (d) MME-10, (f) ClimWIP, (h) LR, and (i) RF. (c), (e), (g), (i), and (k) display biases of different multi-model ensemble methods, calculated as ensemble-minus-observation differences (units: ℃).\u003c/p\u003e","description":"","filename":"floatimage6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8673890/v1/5762a2284d306d3bd0d0be63.jpeg"},{"id":101751173,"identity":"77d01ed3-fbb6-4f90-a280-a6ab0f69fc38","added_by":"auto","created_at":"2026-02-03 10:17:25","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":789828,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial distribution of the difference in summer SATs over the TP between the 1961–1994 and 1995–2014 periods (the latter minus the former) for (a) observation, (b) MME-18, (d) MME-10, (f) ClimWIP, (h) LR, and (i) RF.\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8673890/v1/3ebc0f9005098f255f12bc79.png"},{"id":101396260,"identity":"bbec73f9-d8f6-43dc-a8b5-719e610f750f","added_by":"auto","created_at":"2026-01-29 09:13:46","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":223687,"visible":true,"origin":"","legend":"\u003cp\u003eTime series of projected summer TP SAT during 2015–2100 under different emission scenarios: (a) SSP1-2.6, (b) SSP2-4.5, (c) SSP5-8.5. The red lines represent the RF-based projection of 10 optimal models, and the blue lines represent the MME-18-based projection. The black dotted lines denote the linear trends.\u003c/p\u003e","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-8673890/v1/3bdf65ca7eef0984cf41a95e.png"},{"id":101396266,"identity":"00b7f314-de3d-4f3f-b87f-6734292ec5d6","added_by":"auto","created_at":"2026-01-29 09:13:46","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":605685,"visible":true,"origin":"","legend":"\u003cp\u003eSpatial distributions of summer SAT anomalies over the TP for the early- (2025–2044), mid-(2055–2074), and long-term (2081–2100) future periods under different emission scenarios. Panels (a)–(i) show the RF-based projections, while (j)–(r) display the MME-18-based projections.\u003c/p\u003e","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-8673890/v1/367e62623c3eb940dab6a669.png"},{"id":101396269,"identity":"7426d589-9dfe-4c15-aea6-58850a5a0251","added_by":"auto","created_at":"2026-01-29 09:13:48","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":252437,"visible":true,"origin":"","legend":"\u003cp\u003eTime series of anomalous summer SAT over the western TP (west of 85° E) for 2015–2100 relative to the observed 1995–2014 baseline, under different emission scenarios: (a) SSP1-2.6, (b) SSP2-4.5, (c) SSP5-8.5. The red and blue lines depict the RF-based (10 optimal models) and MME-18-based projections, respectively.\u003c/p\u003e","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-8673890/v1/9efe0d79323469c0dea2bcc6.png"},{"id":101755039,"identity":"e2bcd9f9-38aa-4f94-9f29-b92d4874e002","added_by":"auto","created_at":"2026-02-03 10:48:36","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":8027392,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8673890/v1/48a5175e-5312-4ed7-bb25-c3037bd92ce4.pdf"},{"id":101396267,"identity":"b17630d9-85c5-485f-b57f-cc50cfcda9fe","added_by":"auto","created_at":"2026-01-29 09:13:46","extension":"doc","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":9062400,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryMaterialsnew20260122.doc","url":"https://assets-eu.researchsquare.com/files/rs-8673890/v1/fe55dfbc253cb66fc1ac545c.doc"}],"financialInterests":"No competing interests reported.","formattedTitle":"Future Projections of summer Tibetan Plateau temperature based on the combined influence of sea surface temperature and soil moisture","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eAs a region highly sensitive to global climate change, the Tibetan Plateau (TP) is experiencing accelerated warming\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. During the period 1979\u0026ndash;2020, the mean annual temperature over the TP increased at 0.34\u0026deg;C per decade, approximately twice the global warming rate\u003csup\u003e\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. The rapid warming significantly impacts local ecosystem, water resources, agriculture, and infrastructure. Additionally, the TP climate anomalies play a crucial role in modulating the Asian Monsoon and broader-scale climate patterns\u003csup\u003e\u003cspan additionalcitationids=\"CR3 CR4 CR5 CR6 CR7 CR8 CR9\" citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e. Therefore, studying future TP climate change is critical for both the local and broader climate projection and adaptation.\u003c/p\u003e \u003cp\u003eClimate models have been widely applied in simulating historical and projected climate variability\u003csup\u003e\u003cspan additionalcitationids=\"CR12 CR13 CR14\" citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. Based on Coupled Model Intercomparison Project (CMIP) simulations, numerous studies have evaluated and projected climate change over the TP\u003csup\u003e16\u0026ndash;20\u003c/sup\u003e. Most findings indicated that the warming trend of the TP will intensify with increasing emissions\u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e,\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. However, significant uncertainties exist in TP simulations and future projections due to amplified model errors caused by complex terrain \u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e,\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e. Earlier CMIP simulations exhibited cold biases over the TP\u003csup\u003e16,23,24\u003c/sup\u003e. Although CMIP6 has shown remarkable progress and can more realistically capture the spatial distribution of surface air temperatures (SATs)\u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e, many models still exhibit cold biases in the TP \u003csup\u003e26\u0026ndash;29\u003c/sup\u003e. Therefore, evaluating CMIP6 model performance is essential for reliable TP climate projections.\u003c/p\u003e \u003cp\u003eCurrent projections of SAT over the TP predominantly rely on multi-model ensemble means (MME), which show considerable uncertainties in estimated warming magnitudes\u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e,\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e,\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e,\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e,\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e. Observational constraint methods (e.g., multi-model weighting, attribution-based constraint, emergent constraint) can effectively reduce projection uncertainty\u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e. However, few studies have employed such observational constraint methods to improve climate projections for the TP \u003csup\u003e1,32,,33\u003c/sup\u003e. Zhou and Zhang\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e demonstrated that using historical attribution results as constraints leads to the projected TP warming magnitude significantly exceeding the CMIP5 MME estimates, revealing that the latter systematically underestimated the TP\u0026rsquo;s temperature response to anthropogenic forcing.\u003c/p\u003e \u003cp\u003eAn ideal model should accurately reproduce observed physical relationships, which can help reduce projection uncertainties \u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e,,\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e. These relationships can help reduce uncertainties in predictions and projections of climate models\u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e. For instance, using the observed precipitation-temperature relationship to identify models that correctly simulate land-atmosphere feedback processes, the uncertainties in extreme temperature projections can be effectively reduced \u003csup\u003e\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e,\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e\u003c/sup\u003e. The tropical temperature-humidity relationship can serve as an observational constraint for evaluating model performance\u003csup\u003e\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e. Using observed relationships as physics-based constraints for model selection can effectively reduce projection uncertainty.\u003c/p\u003e \u003cp\u003eRecently, the combined influence of East European-West Siberian soil moisture anomalies and Yellow-Japan Sea SSTAs was found to regulate summer SATs over the eastern TP\u003csup\u003e40\u003c/sup\u003e, offering a potential physics-based constraint. This also raises critical questions: Can CMIP6 models accurately reproduce the combined modulation of the soil moisture anomalies and SSTAs on summer SATs across the entire TP? Can these physics-based observational constraints improve the accuracy of TP SAT projections? These questions warrant investigation to improve the reliability of summer SAT projections over the TP.\u003c/p\u003e"},{"header":"2. Data and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Data\u003c/h2\u003e \u003cp\u003eThis study used the CN05.1 monthly SAT dataset (0.25\u0026deg;\u0026times;0.25\u0026deg;spatial resolution) obtained from the National Meteorological Information Center to assess the performance of CMIP6 models\u003csup\u003e\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e. This dataset has been extensively applied in climate change studies and model evaluations\u003csup\u003e\u003cspan additionalcitationids=\"CR43 CR44\" citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e. Moreover, the regionally averaged summer SAT over the TP region (73\u0026deg;\u0026ndash;105\u0026deg;E, 26\u0026deg;\u0026ndash;40\u0026deg;N, above 2,000 m elevation) shows good agreement with the summer TP SAT average from 89 observational stations in the TP, with a correlation coefficient of 0.95 for the period 1961\u0026ndash;2014, confirming its reliability.\u003c/p\u003e \u003cp\u003eThis study also used geopotential heights at 2.5\u0026deg;\u0026times;2.5\u0026deg; resolution from the National Centers for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR), the ERA-land monthly soil moisture (0\u0026ndash;7 cm layer; 0.1\u0026deg;\u0026times;0.1\u0026deg; resolution)\u003csup\u003e\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e, and the extended reconstructed SST dataset (Version 5; 2.0\u0026deg;\u0026times;2.0\u0026deg; resolution) obtained from the National Oceanic and Atmospheric Administration (NOAA) \u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e. All data were extracted from 1961 to 2014.\u003c/p\u003e \u003cp\u003eThis study employed model-simulated data from 18 CMIP6 models with topographic data (see the list in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). The SAT data were topographically adjusted to match the CN05.1 grid topography using atmospheric lapse rates\u003csup\u003e\u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e\u003c/sup\u003e. We used historical simulations (1961\u0026ndash;2014) and future projections (2015\u0026ndash;2100) under three Shared Socioeconomic Pathway (SSP) scenarios (SSP1-2.6, SSP2-4.5, and SSP5-8.5) \u003csup\u003e48,49\u003c/sup\u003e. All CMIP6 model outputs used in this study represent the first ensemble run (r1i1p1f1) to ensure consistent initial conditions, physical parameterizations, and boundary conditions. The period 1995\u0026ndash;2014 was designated as the baseline period for comparative analysis of future climate change.\u003c/p\u003e \u003cp\u003eBesides direct evaluation of summer TP SATs, we used the combined effect of the soil moisture anomalies and SSTAs as observational constraints to evaluate the performance of models. As such, we used various model output variables (geopotential heights, SSTs, and surface soil moistures). All model variables were bilinearly interpolated to match observational dataset resolutions.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2. Methods\u003c/h2\u003e \u003cp\u003eThe Taylor diagram and Taylor skill score (TSS) were used to quantitatively assess the consistency between simulations and observations\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e,\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e,\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e\u003c/sup\u003e. The interannual variability skill score (IVS) is used to evaluate the ability of models to reproduce the interannual variability of the summer TP SAT \u003csup\u003e\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e,\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eWe evaluated four multi-model ensemble approaches: the conventional equal-weight MME, Climate model Weighting by Independence and Performance (ClimWIP)\u003csup\u003e\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e,\u003cspan additionalcitationids=\"CR55 CR56\" citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e57\u003c/span\u003e\u003c/sup\u003e, Linear Regression (LR)\u003csup\u003e\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e\u003c/sup\u003e, and Random Forest (RF) \u003csup\u003e\u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e58\u003c/span\u003e\u003c/sup\u003e. To mitigate overfitting and evaluate generalizability, we train the ensemble models on the period 1961\u0026ndash;1994 and validate them on the period 1995\u0026ndash;2014, with their robustness and performance assessed using the root mean square error (RMSE) and coefficient of determination (R\u003csup\u003e2\u003c/sup\u003e score). For detailed information on the above evaluation and multi-model ensemble methods, please refer to the supplementary material.\u003c/p\u003e \u003c/div\u003e"},{"header":"3. Evaluation of CMIP6 models","content":"\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003e3.1 Simulation performance for summer SATs over the TP\u003c/h2\u003e \u003cp\u003eTaylor diagram evaluation (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea) shows that all 18 models reasonably simulate the climatological spatial distribution of summer SATs over the TP. The simulated SAT distributions in most models show high consistency with the observed distribution, with correlation coefficients exceeding 0.80. NorESM2-LM, NorESM2-MM, CESM2-WACCM, ACCESS-CM2, and TaiESM1 demonstrated superior performance in simulating the spatial distribution of SATs (correlation coefficients\u0026thinsp;\u0026gt;\u0026thinsp;0.95 and TSS\u0026thinsp;\u0026gt;\u0026thinsp;0.89). In contrast, CanESM5 and CanESM5-1 exhibited relatively poor spatial simulation skills, with lower spatial correlation coefficients and TSS scores below 0.40.\u003c/p\u003e \u003cp\u003eThe IVS evaluation (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb) shows that most models demonstrated high skills in reproducing the interannual variability (IVS\u0026thinsp;\u0026lt;\u0026thinsp;0.20). However, CanESM5 and CanESM5-1 exhibited considerably weaker performance (IVS\u0026thinsp;\u0026gt;\u0026thinsp;6.0) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb), consistent with their poor spatial simulation skills (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eBased on high spatial similarity with the observed pattern of TP SAT (TSS\u0026thinsp;\u0026ge;\u0026thinsp;0.8) and strong simulation capability for the interannual variation of TP SAT (IVS\u0026thinsp;\u0026le;\u0026thinsp;0.2), we identified 12 optimal models: ACCESS-CM2, ACCESS-ESM1-5, AWI-CM-1-1-MR, BCC-CSM2-MR, CESM2-WACCM, CMCC-CM2-SR5, CMCC-ESM2, MPI-ESM1-2-HR, MRI-ESM2-0, NorESM2-LM, NorESM2-MM, and TaiESM1.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec7\" class=\"Section2\"\u003e \u003ch2\u003e3.2 Evaluation by using physics-based observational constraints\u003c/h2\u003e \u003cp\u003eThe soil moisture anomalies in the East European Plain-West Siberian Plain and SSTAs in the Yellow Sea-Japan Sea jointly modulate multiscale variations of summer SATs over the eastern TP\u003csup\u003e40\u003c/sup\u003e. Specifically, soil moisture anomalies in the East European Plain-West Siberian Plain modulate the overlying upper-tropospheric geopotential heights via local land-air interactions, triggering a downstream Rossby wave train. This wave train generates a high-pressure anomaly extending from the eastern Tibetan Plateau to the Yellow and Japan Seas, which reflects an intensified and eastward-extended South Asian High. The SSTAs reinforce this high-pressure anomaly through local thermal forcing and ultimately cause summer SAT anomalies over the eastern TP. In short, the combined effect of these soil moisture and SST anomalies leads to summer SAT anomalies over the eastern TP\u003csup\u003e40\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eSince this study focuses on the evaluation and projection of summer SATs over the entire TP, we calculated the TP regionally averaged SAT time series for the period 1961\u0026ndash;2014 and used it to perform regression analyses on 200 hPa geopotential heights (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea), soil moistures (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb), and SSTAs (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ec). All variables in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e have been detrended, since we expect CMIP6 models to not only capture the simultaneous variation trends of soil moistures, SSTs, and TP SAT, but also to accurately reproduce the multiscale modulation of soil moisture anomalies and SSTAs on TP SAT during the historical period (1961\u0026ndash;2014).\u003c/p\u003e \u003cp\u003eCorresponding to higher TP SAT, a positive-negative-positive wave train extends from eastern Europe to the TP-Japan Sea region (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea). Significant negative soil moisture anomalies appear in the East European Plain-West Siberian Plain (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb), and significantly positive SSTAs occur in the Yellow Sea-Japan Sea (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ec). These anomalies can excite and maintain the wave train from eastern Europe to the TP-Japan Sea region, leading to positive geopotential height anomalies over the TP (reflecting a stronger and northeastward South Asian high) and thereby causing higher summer TP SAT. These results are consistent with Zhang et al\u003csup\u003e40\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eThe 200-hPa geopotential height, soil moisture, and SST anomalies regressed upon the raw (i.e., non-detrended) TP SAT time series show similar spatial patterns to those in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea\u0026ndash;c, but with higher significance levels in the key regions (Fig.omitted). The interdecadal variability and trend of the observed TP SAT are successfully captured by a fitting model based on the raw soil moisture index (East European\u0026ndash;West Siberian Plain) and the raw SST index (Yellow Sea\u0026ndash;Japan Sea) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ed). The correlation coefficient between the low-pass-filtered observed and reconstructed time series reaches 0.95, passing a Monte Carlo test at the 99% confidence level. This indicates that the soil moisture anomalies and SSTAs in these key regions collectively regulate the interdecadal variability and trend of the SAT across the entire TP.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThis collective modulation mechanism can be used as an observational constraint. To assess whether CMIP6 models can reproduce the multiscale modulation of soil moisture anomalies and SSTAs on TP SAT during the historical period (1961\u0026ndash;2014), all variables in the regression analyses of 200 hPa geopotential heights (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e), soil moistures (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e), and SSTAs (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e) have been detrended. In these Figs., NESM3 and AWI-CM-1-1-MR were excluded since they do not provide historical simulated soil moisture data.\u003c/p\u003e \u003cp\u003eMost models successfully simulated the significant positive geopotential height anomalies over the TP and accurately reproduced the wave train propagating from Europe to the TP region. However, CanESM5-1 (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eg) and CanESM5 (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eh) showed no clear wave train. TaiESM1 (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ep) even showed no significant positive geopotential height anomaly over the TP.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eWe also examined soil moisture anomalies regressed against the TP SAT time series for each model (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). Most models capture significant negative soil moisture anomalies around Europe and Siberia, although the locations differ somewhat from observations (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb). Furthermore, the negative soil moisture anomalies generally correspond well with the positive geopotential height anomalies in the overlying troposphere. For instance, in ACCESS-ESM1-5 (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb), the location of significant negative soil moisture anomalies resembles that in the observation (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb). Correspondingly, a positive geopotential height anomaly appears over eastern Europe (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb), with a location highly consistent with the observation (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea). However, in BCC-CSM2-MR (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ec) and NorESM2-MM (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eo), significant negative soil moisture anomalies shift northwestward compared to the observations (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb). As such, the center of significant positive geopotential height anomalies also shifts northwestward (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec,o), aligning with the locations of soil moisture anomalies in the two models. Despite slight location differences, most models still reproduce the wave trains from Europe to the TP, which contribute to the summer TP SAT anomaly. However, MPI-ESM1-2-LR showed very weak, insignificant negative soil moisture anomalies over Europe (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003el). Correspondingly, no significant positive geopotential height anomalies appear over this region (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003el).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eMost models successfully simulate significant positive SSTAs in the Yellow Sea\u0026ndash;Japan Sea region associated with higher TP SAT, which is consistent with observations (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ec). CanESM5-1, MIROC6, and MPI-ESM1-2-LR showed weak and insignificant SSTAs in this region.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eBased on physics-based observational constraints (i.e., the combined effect of soil moisture anomalies and SSTAs), we identified 11 optimal models: ACCESS-CM2, ACCESS-ESM1-5, BCC-CSM2-MR, CESM2-WACCM, CMCC-CM2-SR5, CMCC-ESM2, FGOALS-g3, MPI-ESM1-2-HR, MRI-ESM2-0, NorESM2-LM, and NorESM2-MM. We also investigated the 200-hPa geopotential height, soil moisture, and SST anomalies regressed upon the raw (i.e., non-detrended) TP SAT time series (see supplementary Figs.\u0026nbsp;1\u0026ndash;3). The results reveal that these physically better-performing models show similar spatial patterns to those in Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e\u0026ndash;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e. In contrast, the physically poorer-performing models (e.g., CanESM5-1, CanESM5, and TaiESM1) fail to capture the wave train.\u003c/p\u003e \u003cp\u003eTen of the 11 models align with the selection based on the TSS and IVS criteria, with only FGOALS-g3 having a relatively low TSS. This result implies that the models reproducing the combined effects of soil moisture anomalies and SSTAs on summer TP SAT tend to more accurately simulate the TP SAT (achieving high Taylor and IVS scores). In contrast, those that fail to reproduce the physical link perform poorly. This mutual verification between physics-based constraints and Taylor and IVS scores enhances model credibility. Therefore, based on all these criteria, we ultimately identified 10 optimal models: ACCESS-CM2, ACCESS-ESM1-5, BCC-CSM2-MR, CESM2-WACCM, CMCC-CM2-SR5, CMCC-ESM2, MPI-ESM1-2-HR, MRI-ESM2-0, NorESM2-LM, and NorESM2-MM. Unless otherwise specified, all subsequent multi-model ensemble analyses are derived from the 10 optimally selected models.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003e3.3 Evaluation of multi-model ensemble methods\u003c/h2\u003e \u003cp\u003eWe use different ensemble methods to establish the base or regression models during the training phase (1961\u0026ndash;1994) and assess their performance during the validation phase (1995\u0026ndash;2014). Relative to the other methods, the RF approach demonstrates the best agreement with the observation, with the lowest RMSE (0.69) and highest R\u003csup\u003e2\u003c/sup\u003e (0.96) (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The LR method is the second best (RMSE\u0026thinsp;=\u0026thinsp;0.96, R\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.92) (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). The MME method performs the worst. Nevertheless, the MME of the 10 optimal models (MME-10) outperforms the MME of 18 models (MME-18), with a lower RMSE (1.16) and a higher R\u003csup\u003e2\u003c/sup\u003e (0.91), demonstrating the effectiveness of model selection.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eThe assessment results of multi-model ensemble methods.\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eMME-18\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMME-10\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eLR\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eClimWIP\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eRF\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eR\u003csup\u003e2\u003c/sup\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.85\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.91\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.92\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e0.92\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.96\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRMSE\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e1.59\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e1.16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.96\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e \u003cp\u003e1.07\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e \u003cp\u003e0.69\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eAll multi-model ensemble methods successfully capture the spatial distribution of climatological mean summer SATs over the TP (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003eb, d, f, h, and j), showing a high consistency with the observation (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ea). However, the differences between the multi-model ensemble and observed SATs display substantial discrepancies (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ec, e, g, i, and k). The MME-18 exhibits the strongest and most extensive cold biases in the western TP (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ec.) For the MME-10, the cold biases in the western TP have been to some extent decreased (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ee), verifying model selection validity. Among all ensemble methods, the RF yields the best agreement with the observed SATs, with smaller, scatter cold/warm biases rather than large-scale systematic biases (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003ek).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe summer TP SAT has increased significantly since the 1990s (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ed). The spatial distribution of the difference in summer SATs between the 1961\u0026ndash;1994 and 1995\u0026ndash;2014 periods (the latter minus the former) shows that the entire TP experienced warming (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003ea). Although historical simulations based on different multi-model ensemble methods successfully capture the warming over most of the TP, they incorrectly produce cooling in the western and southern TP and considerably overestimate the warming in the northern TP. This discrepancy is most pronounced for the MME-18 (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003eb). The MME-10 also erroneously simulated cooling in western and southern TP but with reduced magnitude relative to the MME-18 (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003ec), indicating improvement through optimal selection. The RF method shows the best agreement with the observed warming over the entire TP, exhibiting only weak, scattered cooling in the western and southern TP and the slightest overestimation of warming over the northern TP (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003ef).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eIn summary, using the mutual verification between physics-based observational constraints and Taylor and IVS scores to optimally select 10 CMIP6 models, and then applying the RF method for multi-model ensemble, the outputs can effectively reduce historical simulation biases, and therefore are expected to provide more reliable future projections of summer TP SATs.\u003c/p\u003e \u003c/div\u003e"},{"header":"4. Projected variations in summer TP SATs under different emission scenarios","content":"\u003cp\u003eWe employed the RF method with 10 optimal models to project and analyze the characteristics of summer SAT variation over the TP under the SSP1-2.6, SSP2-4.5, and SSP5-8.5 (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eUnder SSP1-2.6, the warming in the TP primarily occurs in the early period. From 2015 to 2044, the summer SAT exhibits a significant upward trend, with a warming rate of approximately 0.38\u0026deg;C/10a (exceeding the 99% confidence level) (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ea). The warming rate is higher than the historical rate (0.25 ℃/10a for 1961\u0026thinsp;\u0026minus;\u0026thinsp;2014). In the mid- to long-term future (2045\u0026ndash;2100), the TP SAT shows no clear trend (0.03 ℃/10a, insignificant). Compared to the RF-based projection, the MME-18 projection shows a systematic cold bias (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ea), indicating that the latter significantly underestimates the future warming of the TP. Given that the RF-based outputs appear more reliable, the above results suggest that the future warming over the TP may be more severe than the MME-18 projection.\u003c/p\u003e \u003cp\u003eUnder SSP2-4.5, the TP SAT exhibits a distinct upward trend during 2015\u0026ndash;2070 (approximately 0.33\u0026deg;C/10a, exceeding the 99% confidence level) (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003eb). During the long-term future (2071\u0026ndash;2100), the warming persists but decelerates (approximately 0.14\u0026deg;C/10a). The MME-18-based projection exhibits a clear cold bias during 2015\u0026ndash;2070 (approximately 0.39\u0026deg;C lower than the RF-based projection). During 2071\u0026ndash;2100, the two projections tend to be more consistent.\u003c/p\u003e \u003cp\u003eUnder SSP5-8.5, the TP SAT shows a more pronounced and sustained upward trend throughout 2015\u0026ndash;2100 (0.60\u0026deg;C/10a) (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ec). The MME-18 also exhibits a systematic cold bias under this scenario (approximately 0.34\u0026deg;C lower than the RF-based projection) (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003ec). This discrepancy diminishes to a negligible level during 2071\u0026ndash;2100 (approximately 0.02\u0026deg;C difference).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003e displays the spatial distribution of summer SAT anomalies for the early- (2025\u0026ndash;2044), mid- (2055\u0026ndash;2074), and long-term (2081\u0026ndash;2100) future under different scenarios. The RF-based projections show that relative to the baseline period (1995\u0026ndash;2014), the SSP1-2.6 scenario exhibits predominant warming across most TP regions in the early future, with temperature increases of 1\u0026ndash;2\u0026deg;C (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003ea). In mid- and long-term future, the increase reaches 2\u0026ndash;3\u0026deg;C in these regions (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003eb,c).\u003c/p\u003e \u003cp\u003eIn contrast, the MME-18-based projections show relatively weaker warming in central and eastern TP and even cooling in western TP under the SSP1-2.6 scenario (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003ej-l). A similar west-cooling-east-warming pattern appears under SSP2-4.5 and SSP5-8.5 (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003em\u0026ndash;q), except during the long-term future under SSP5-8.5 (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003er).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eCompared to the MME-18, the RF-based projections demonstrate stronger warming, particularly in western TP, resulting in more spatially homogeneous warming throughout the entire TP (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003ed\u0026ndash;i). These findings indicate that the RF-based projections primarily reduce the cold bias in western TP. We further investigate the warming magnitude in the western TP (west of 85\u0026deg; E) and find that the RF-based projections show a much more dramatic warming compared to the MME-18 (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003e). Under SSP1-2.6, the western TP SAT in the RF-based projections shows persistent positive anomalies relative to the 1995\u0026ndash;2014 baseline, with mean warming reaching 1.1\u0026deg;C during 2045\u0026ndash;2100 (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003ea). In contrast, the MME-18-based projections are almost always lower than this baseline (blue line in Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003ea). Under SSP2-4.5, the RF-based projections also show persistent warming (mean 1.66\u0026deg;C during 2071\u0026ndash;2100) (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eb). The MME-18-based projections remain below the baseline until the late 2040s, showing only modest mean warming of 0.94\u0026deg;C in 2071\u0026ndash;2100 (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003eb). Under SSP2-8.5, the MME-18-based projections do not exceed the observed baseline until the early 2040s (blue line in Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003ec), while the RF-based projections exhibit a more rapid warming trend (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003ec). Given that western TP has already experienced significant warming, future cooling under global warming seems unlikely. Therefore, the RF-based projections are more reliable. These results suggest that future warming in the western TP may be more severe than the MME-18-based projections.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"5. Conclusions and discussion","content":"\u003cp\u003eAccurately projecting future TP warming is crucial for local ecological and socioeconomic sustainability. This study evaluated CMIP6 models by combining direct assessment of TP SAT variations with physics-based observational constraints derived from the combined influence of soil moisture anomalies and SSTAs on summer TP SAT. Optimal models and ensemble methods were identified to analyze future changes under different scenarios. The main conclusions are as follows:\u003c/p\u003e \u003cp\u003e(1) Using mutual verification combining physics-based observational constraints and statistical skill assessment (Taylor and IVS scores), we identified 10 optimal models. Among four model ensemble methods (MME, ClimWIP, LR, and RF), we found that the RF method effectively reduced the cold bias and accurately simulated the climatological spatial distribution and multi-scale (decadal and trend) temporal variations of summer TP SATs.\u003c/p\u003e \u003cp\u003e(2) The RF-based projections with the optimal models show that significant summer TP warming, with its magnitude and evolution strongly dependent on the emission scenario. Under SSP1-2.6, the TP exhibits significant early-term warming (0.38\u0026deg;C/decade during 2015\u0026ndash;2044), with no clear trend thereafter. Under SSP2-4.5, the TP SAT shows a significant warming trend (0.33\u0026deg;C/10a) during 2015\u0026ndash;2070, and a slower warming trend (0.14\u0026deg;C/10a) afterwards. Under SSP5-8.5, the TP SAT shows a significant warming trend (0.60\u0026deg;C/10a) throughout 2015\u0026ndash;2100. The non-optimized MME-18-based projections systematically underestimated the TP warming, especially over western TP. The constrained projections suggest future warming, particularly in the vulnerable western TP, may be more severe than commonly projected.\u003c/p\u003e \u003cp\u003eUsing optimal fingerprinting, Zhou and Zhang\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e attributed historical warming over the TP to anthropogenic activities, particularly greenhouse gas forcing. Using the attribution results as observational constraints, the predicted TP warming may exceed the CMIP5 multi-model ensemble\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e. Although the constraint conditions in this study differ from Zhou and Zhang\u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u003c/sup\u003e, the conclusions are consistent, further supporting the likelihood of more pronounced future TP warming.\u003c/p\u003e \u003cp\u003eSustained and significant warming in the TP would likely increase local soil temperatures, accelerate glacier melt\u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e,\u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e59\u003c/span\u003e\u003c/sup\u003e, alter permafrost\u003csup\u003e\u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e60\u003c/span\u003e,\u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e61\u003c/span\u003e\u003c/sup\u003e, enhance evaporation, reduce soil moisture, and exacerbate grassland degradation and desertification\u003csup\u003e\u003cspan additionalcitationids=\"CR63\" citationid=\"CR62\" class=\"CitationRef\"\u003e62\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e64\u003c/span\u003e\u003c/sup\u003e. Summer snow cover in the high-altitude western TP can influence summer precipitation anomalies over eastern China\u003csup\u003e\u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e65\u003c/span\u003e,\u003cspan citationid=\"CR66\" class=\"CitationRef\"\u003e66\u003c/span\u003e\u003c/sup\u003e. Under stronger warming, reduced or vanished snow cover could significantly alter climate anomalies in eastern China, which demands particular attention.\u003c/p\u003e \u003cp\u003eAdditionally, European soil moisture anomalies play a critical regulating role in summer TP SAT. If the regulatory mechanism and its underlying physical constraint remain intact, future extreme warming could be somewhat mitigated. This suggests that multinational and multi-regional cooperation is essential to address climate change. To enhance climate resilience, regional climate impacts and feedback effects should be integrated into global climate governance agendas.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eT.Z. and G.L. conceived the study. T.Z. and G.L. performed the analyses and interpreted the data. T.Z., G.L.and M.D. led the writing with input from all co-authors.\u003c/p\u003e\u003ch2\u003eAcknowledgements\u003c/h2\u003e \u003cp\u003eThis work was jointly sponsored by the National Key Research and Development Program of China (Grant 2023YFF0805300), the Major Science and Technology project of the Xizang Autonomous Region (Grant XZ202402ZD0006), the Youth Innovation Team of China Meteorological Administration \u0026ldquo;Climate change and its impact in the Tibetan Plateau\u0026rdquo; (Grant CMA2023QN16), and the Basic Research Fund of CAMS (Grant 2023Z024). We thank the National Tibetan Plateau Data Center for providing the Tibetan Plateau boundary dataset (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://data.tpdc.ac.cn\u003c/span\u003e\u003cspan address=\"http://data.tpdc.ac.cn\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe CMIP6 model data is available from https://aims2.llnl.gov/search/cmip6/.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eZhou, T. \u0026amp; Zhang, W. Anthropogenic warming of Tibetan Plateau and constrained future projection. Environ. Res. Lett. 16, 044039 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYou, Q. et al. Warming amplification over the Arctic Pole and Third Pole, Trends, mechanisms and consequences. Earth Sci. Rev. 217, 103625 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDuan, A. \u0026amp; Wu, G. Role of the Tibetan Plateau thermal forcing in the summer climate patterns over subtropical Asia. Clim. Dyn. 24, 793\u0026ndash;807 (2005).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou, X., Zhao, P., Chen, J., Chen, L. \u0026amp; Li, W. Impacts of thermodynamic processes over the Tibetan Plateau on the Northern Hemispheric climate (in Chinese). Sci China Ser D-Earth Sci. 39(11), 1473\u0026ndash;1486 (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu, G., Zhuo, H., Wang, Z. \u0026amp; Liu, Y. Two types of summertime heating over the Asian large-scale orography and excitation of potential-vorticity forcing, I. Over Tibetan Plateau. Sci. China Earth Sci. 59, 1996\u0026ndash;2008 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, G., Zhao, P., Nan, S., Chen, J. \u0026amp; Wang, H. Advances in the study of linkage between the Tibetan Plateau thermal anomaly and atmospheric circulations over its upstream and downstream regions (in Chinese). Acta Meteorol Sin. 76(6), 861\u0026ndash;869 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLu, M. et al. Possible effect of the Tibetan Plateau on the \u0026ldquo;upstream\u0026rdquo; climate over West Asia, North Africa, South Europe and the North Atlantic. Clim. Dyn. 51, 1485\u0026ndash;1498 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, Y. et al. Land-atmosphere-ocean coupling associated with the Tibetan Plateau and its climate impacts. Natl. Sci. Rev. 7(3), 534\u0026ndash;552 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNan, S., Zhao, P., Chen, J. \u0026amp; Liu, G. Links between the thermal condition of the Tibetan Plateau in summer and atmospheric circulation and climate anomalies over the Eurasian continent. Atmos. Res. 247, 105212 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang, J. et al. Global climate impacts of land-surface and atmospheric processes over the Tibetan Plateau. Rev. Geophy. 61(3), 000771 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHe, S., Yang, J., Bao, Q., Wang, L. \u0026amp; Wang, B. Fidelity of the observational /reanalysis datasets and global climate models in representation of extreme precipitation in East China. J. Clim. 32(1), 195\u0026ndash;212 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYu, E. \u0026amp; Sun, J. Extreme temperature projection over northwestern China based on multiple regional climate models (in Chinese). Trans Atmos Sci. 42(1), 46\u0026ndash;57 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKhan, A., Koch, M. \u0026amp; Tahir, A. A. Impacts of climate change on the water availability, seasonality and extremes in the upper Indus Basin (UIB). Sustainability. 12(4), 1283 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiang, W. \u0026amp; Chen, H. Assessment and projection of changes in temperature extremes over the mid-high latitudes of Asia based on CMIP6 models (in Chinese). Trans Atmos Sci. 44(4), 592\u0026ndash;603 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu, R., Liang, X. \u0026amp; Duan, M. Evaluation of CWRF simulation of temperature and precipitation on the Qinghai-Tibet Plateau (in Chinese). Trans Atmos Sci. 44(1), 104\u0026ndash;117 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu, Y. \u0026amp; Xu, C. Preliminary assessment of simulations of climate changes over China by CMIP5 multi-models. Atmos. Oceanic Sci Lett. 5(6), 489\u0026ndash;494 (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHu, Q., Jiang, D. \u0026amp; Fan, G. Climate change projection on the Tibetan Plateau, Results of CMIP5 models (in Chinese). Chinese J. Atmos Sci. 39(2), 260\u0026ndash;270 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang, M., Wang, X., Pang, G., Wang, G. \u0026amp; Liu, Z. The Tibetan Plateau cryosphere, observations and model simulations for current status and recent changes. Earth Sci. Rev. 190, 353\u0026ndash;369 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou, T. et al. The near-term, mid-term and long-term projections of temperature and precipitation changes over the Tibetan Plateau and the sources of uncertainties (in Chinese). J. Meteorol. Res. 40(5), 697\u0026ndash;710 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, J. et al. CMIP6 evaluation and projection of climate change in Tibetan Plateau (in Chinese). Journal of Beijing Normal University (Natural Science). 58(1), 77\u0026ndash;88 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, S. \u0026amp; Xiong, Z. The Preliminary Analysis of 5 Coupled Ocean-Atmosphere Global Climate Models Simulation of Regional Climate in Asia (in Chinese). Climatic Environ Res. 9(2), 240 (2004).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDing, Y. et al. Detection, causes and projection of climate change over China, An overview of recent progress. Adv. Atmos. Sci. 24, 954\u0026ndash;971 (2007).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiang, D., Wang, H. \u0026amp; Lang, X. Evaluation of East Asian climatology as simulated by seven coupled models. Adv. Atmos. Sci. 22, 479\u0026ndash;495 (2005).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiang, D., Tian, Z. \u0026amp; Lang, X. Reliability of climate models for China through the IPCC Third to Fifth Assessment Reports. INT J CLIMATOL. 36(3), 1114\u0026ndash;1133 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou, T., Zou, L. \u0026amp; Chen, X. Commentary on the Coupled Model Intercomparison Project Phase 6 (CMIP6) (in Chinese). ADV CLIM CHANG RES. 15 (5), 445\u0026ndash;456 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSu, F., Duan, X., Chen, D., Hao, Z. \u0026amp; Guo, L. Evaluation of the global climate models in the CMIP5 over the Tibetan Plateau. J. Clim. 26(10), 3187\u0026ndash;3208 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuo, D. L., Sun, J. \u0026amp; Yu, E. Evaluation of CORDEX regional climate models in simulating temperature and precipitation over the Tibetan Plateau. Atmos. Oceanic Sci Lett. 11(3), 219\u0026ndash;227 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhu, Y. \u0026amp; Yang, S. Evaluation of CMIP6 for historical temperature and precipitation over the Tibetan Plateau and its comparison with CMIP5. Adv. Clim. Chang. Res. 11(3), 239\u0026ndash;251 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen, W., Jiang, D. \u0026amp; Wang, X. Evaluation and Projection of CMIP6 Models for Climate over the QinghaiXizang (Tibetan) Plateau (in Chinese). Plateau Meteorol. 40(6), 1455\u0026ndash;1469 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang, Y., Dai, X., Tong, H. \u0026amp; Zhang, B. CMIP5 Model Precipitation Bias-correction Methods and Projected China Precipitation for the Next 30 Years (in Chinese). Climatic Environ Res. 24(6), 769\u0026ndash;784 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou, B. \u0026amp; Zhai, P. The constraint methods for projection in the IPCC Sixth Assessment Report on climate change (in Chinese). Acta Meteorol. Sin. 9(6),1063\u0026ndash;1070 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao, Y., Zhou, T., Zhang, W. \u0026amp; Li, J. Change in Precipitation over the Tibetan Plateau Projected by Weighted CMIP6 Models. Adv. Atmos. Sci. 39, 1133\u0026ndash;1150 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eQiu, H., Zhou, T., Chen, X., Wu, B. \u0026amp; Jiang, J. Understanding the diversity of CMIP6 models in the projection of precipitation over Tibetan Plateau. Geophys. Res. Lett. 51(3), 106553 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVrac, M. Multivariate bias adjustment of high-dimensional climate simulations, the Rank Resampling for Distributions and Dependences (R2D2) bias correction. Hydrol. Earth Syst. Sci. 22, 3175\u0026ndash;3196 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVillalobos-Herrera, R. et al. Towards a compound-event-oriented climate model evaluation, a decomposition of the underlying biases in multivariate fire and heat stress hazards. Nat Hazard Earth Sys. 21(6), 1867\u0026ndash;1885 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOrbe, C. GISS Model E2.2, A Climate Model Optimized for the Middle Atmosphere-2. Validation of Large-Scale Transport and Evaluation of Climate Response. J Geophys Res-atmos. 125(24), 033151 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDonat, M., Pitman, A. \u0026amp; Ang\u0026eacute;lil, O. Understanding and reducing future uncertainty in midlatitude daily heat extremes via land surface feedback constraints. Geophys. Res. Lett. 45(19), 10,627\u0026thinsp;\u0026ndash;\u0026thinsp;10,636 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVogel, M. M., Zscheischler, J. \u0026amp; Seneviratne, S. I. Varying soil moisture\u0026ndash;atmosphere feedbacks explain divergent temperature extremes and precipitation projections in central Europe. Earth Syst. Dyn. 9(3), 1107\u0026ndash;1125 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, Y., Held, I. \u0026amp; Fueglistaler, S. Projections of tropical heat stress constrained by atmospheric dynamics. NAT GEOSCI. 14(3),133\u0026ndash;137 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, T. et al. Synergistic contribution of soil moisture and sea surface temperature to summer Tibetan Plateau temperature. Atmos. Res. 314,107811 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWu, J. \u0026amp; Gao, X. A gridded daily observation dataset over China region and comparison with the other datasets. Chinese J Geophys-ch. 56(4), 1102\u0026ndash;1111 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuo, D. \u0026amp; Wang, H. Erratum to: comparison of a very-fine-resolution GCM with RCM dynamical downscaling in simulating climate in China. Adv Atmos Sci. 33(6), 794 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou, B., Xu, Y., Wu, J., Dong, S. \u0026amp; Shi, Y. Changes in temperature and precipitation extreme indices over China: analysis of a high-resolution grid dataset. Int J Climatol. 36(3), 1051\u0026ndash;1066 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGao, Y., Xiao, L., Chen, D., Xu, J. \u0026amp; Zhang, H. Comparison between past and future extreme precipitations simulated by global and regional climate models over the Tibetan Plateau. Int J Climatol. 38(3), 1285\u0026ndash;1297 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang, K., Guo, D., Hua, W., Ma, D. \u0026amp; Xing, Y. Evaluation and projection of CMIP6 HighResMIP in simulating surface air temperature and precipitation over the Tibetan Plateau (in Chinese). Trans Atmos Sci. 46(2), 193\u0026ndash;204 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMu\u0026ntilde;oz-Sabater, J. et al. ERA5-Land, A state-of-the-art global reanalysis dataset for land applications. Earth Syst. Sci. Data. 13 (9), 4349\u0026ndash;4383 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang, B. et al. Extended Reconstructed Sea Surface Temperature, version 5 (ERSST.v5), upgrades, validations and Intercomparisons. J. Climate. 30(20), 8179\u0026ndash;8205 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eEyring, V. et al. Overview of the Coupled Model Intercomparison Project Phase 6 (CMIP6) experimental design and organization. Geosci. Model Dev. 9, 1937\u0026ndash;1958 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eO'Neill, B.C., Tebaldi, C., Vuuren, D.P. \u0026amp; Eyring, V. The scenario model intercomparison project (ScenarioMIP) for CMIP6. Geosci. Model Dev. 9(9), 3461\u0026ndash;82 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTaylor, K.E. Summarizing multiple aspects of model performance in a single diagram. J. Geophys. Res. Atmos. 106(D7), 7183\u0026ndash;7192 (2001).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, T., Jiang, Z. \u0026amp; Treut, H. L. Machine learning to optimize climate projection over China with multi-model ensemble simulations. Environ. Res. Lett. 16(9), 094028 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eScherrer, S. C. Present-day interannual variability of surface climate in CMIP3 models and its relation to future warming. Int. J. Climatol. 31, 1518\u0026ndash;1529 (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePeng, Y. et al. Observational constraint on the future projection of temperature in winter over the Tibetan Plateau in CMIP6 models. Environ. Res. Lett. 17, 034023 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen, W., Jiang, Z. \u0026amp; Li, L. Probabilistic Projections of Climate Change over China under the SRES A1B Scenario Using 28 AOGCMs. J. Climate. 24(17), 4741\u0026ndash;4756 (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKnutti, R. et al. A climate model projection weighting scheme accounting for performance and interdependence. Geophys Res Lett. 44(14), 1909\u0026ndash;1918 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSanderson, B.M., Knutti, R. \u0026amp; Caldwell, P. Addressing interdependency in a multimodel ensemble by interpolation of model properties. J. Clim. 28(13), 5150\u0026ndash;5170 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSanderson, B. M., Wehner, M. \u0026amp; Knutti, R. Skill and independence weighting for multi-model assessments. Geosci. Model Dev. 10, 2379\u0026ndash;2395 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiaw, A. \u0026amp; Wiener, M. Classification and regression by Random Forest. R News. 2 (3), 18\u0026ndash;22 (2002).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOku, Y., Ishikawa, H., Haginoya, S. \u0026amp; Ma, Y. Recent trends in land surface temperature on the Tibetan Plateau. J. Clim. 19, 2995\u0026ndash;3003 (2006).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYan, C., Song, X., Zhou, Y., Duan, H. \u0026amp; Li, S. Assessment of aeolian desertification trends from 1975's to 2005's in the watershed of the Longyangxia reservoir in the upper reaches of China's Yellow River. Geomorphology. 112(3), 205\u0026ndash;211 (2009).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCheng, G. \u0026amp; Jin, J. Permafrost and groundwater on the Qinghai-Tibet plateau and in northeast China. Hydrogeol. J. 21(1), 5\u0026ndash;23 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, N., Wang, G., Liu, G., Lin, Y. \u0026amp; Sun, X. The ecological implications of land use change in the source regions of the Yangtze and Yellow Rivers, China. Reg. Environ. Change. 13(5),1099\u0026ndash;1108 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHu, G. et al. Holocene aeolian activity in the headwater region of the Yellow River, northeast Tibet Plateau, China, a first approach by using OSL-dating. Catena. 149, 150\u0026ndash;157 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, Y. et al. Changes in mean and extreme temperature and precipitation over the arid region of northwestern China, observation and projection. Adv. Atmos. Sci. 34(3), 289\u0026ndash;305 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, G., Wu, R. \u0026amp; Zhang, Y. Persistence of snow cover anomaly over the Tibetan Plateau and implication for forecast of summer precipitation over the Meiyu-Baiu region. Atmos. Oceanic Sci. Lett. 7,115\u0026ndash;119 (2014).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, G., Wu, R., Zhang, Y. \u0026amp; Nan, S. The summer snow cover anomaly over the Tibetan Plateau and its association with simultaneous precipitation over the mei-yu-baiu region. Adv. Atmos. Sci. 31,755\u0026ndash;764 (2014).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"npj-climate-and-atmospheric-science","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"npjclimatsci","sideBox":"Learn more about [npj Climate and Atmospheric Science](http://www.nature.com/npjclimatsci/)","snPcode":"41612","submissionUrl":"https://submission.springernature.com/new-submission/41612/3","title":"npj Climate and Atmospheric Science","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"NPJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Tibetan Plateau, CMIP6, projection, warming, observational constraint","lastPublishedDoi":"10.21203/rs.3.rs-8673890/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8673890/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eAccurate projections of future warming characteristics in the Tibetan Plateau (TP) are essential for regional ecosystem stability and socioeconomic development and for understanding its role in modulating the Asian Monsoon and broader-scale climate patterns. This study evaluated and selected optimal CMIP6 models by combining an observation-based physical constraint\u0026mdash;the synergistic influence of soil moisture anomalies and sea surface temperature anomalies on summer TP surface air temperature (SAT)\u0026mdash;with an assessment of SAT spatiotemporal variations. Four multi-model ensemble methods were compared, and the random forest (RF) method was found to most effectively reduce temporal and spatial biases in historical simulations, thus providing a more reliable basis for future projections of summer SAT over the TP. Based on 10 optimal models, the RF-based projections show that SSP1-2.6 scenario, the TP exhibits significant warming trend (0.38\u0026deg;C/10a) in early-term future (2015\u0026ndash;2044) but stagnates (0.03\u0026deg;C/10) during the mid- and long-term future (2045\u0026ndash;2100). Under the SSP2-4.5 scenario, the TP SAT shows a significant warming trend (0.33\u0026deg;C/10a) during 2015\u0026ndash;2070, then slows to 0.14\u0026deg;C/10a. Under the SSP5-8.5 scenario, the TP SAT maintains rapid warming (0.60\u0026deg;C/10a) throughout 2015\u0026ndash;2100. The RF-based projections indicate stronger future warming (particularly in western TP) compared to the conventional multi-model ensemble mean (MME) of non-optimized 18 models (MME-18). These results highlight the risk of underestimating TP warming without proper model optimization, warranting particular attention on this vulnerable region.\u003c/p\u003e","manuscriptTitle":"Future Projections of summer Tibetan Plateau temperature based on the combined influence of sea surface temperature and soil moisture","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-01-29 09:13:41","doi":"10.21203/rs.3.rs-8673890/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-02-09T06:08:13+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-02-09T05:49:31+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-02-06T01:45:46+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-02-04T00:58:01+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"6370880163252618942769730650365107926","date":"2026-01-28T23:58:42+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"245571574270412154438743284403845300579","date":"2026-01-27T13:01:30+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"233388780108496077042962475278376871177","date":"2026-01-27T07:10:22+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-01-27T06:56:25+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-01-26T14:20:30+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-01-26T13:50:07+00:00","index":"","fulltext":""},{"type":"submitted","content":"npj Climate and Atmospheric Science","date":"2026-01-23T01:33:57+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"npj-climate-and-atmospheric-science","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"npjclimatsci","sideBox":"Learn more about [npj Climate and Atmospheric Science](http://www.nature.com/npjclimatsci/)","snPcode":"41612","submissionUrl":"https://submission.springernature.com/new-submission/41612/3","title":"npj Climate and Atmospheric Science","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"NPJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"4a437adb-c063-44cf-938f-ba34f15571be","owner":[],"postedDate":"January 29th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":61928643,"name":"Earth and environmental sciences/Climate sciences"},{"id":61928644,"name":"Earth and environmental sciences/Environmental sciences"}],"tags":[],"updatedAt":"2026-04-28T06:53:50+00:00","versionOfRecord":[],"versionCreatedAt":"2026-01-29 09:13:41","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8673890","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8673890","identity":"rs-8673890","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}