As an important approach to improving prediction accuracy,the post-process error correction of climate model products plays an indispensable role in global operational climate systems.To enhance the prediction precision of numerical climate prediction models,this study applies the Convolutional Neural Network(CNN)approach to conduct post-process correction on key operational prediction products of the third-generation climate operational prediction system of the China Meteorological Administration(CMA),i.e.,CMA-CPSv3.The targeted products include monthly 2 m air temperature,precipitation over China,and the El Niño-Southern Oscillation(ENSO)index during the period 2001-2023.Using reanalysis data from the National Centers for Environmental Prediction(NCEP)as the observational benchmark,a dedicated correction model has been developed through deep learning training of a multi-layer CNN architecture.After model construction,changes in the model performance before and after correction are evaluated during an independent test period.Results indicate that the CNN model significantly improves the prediction accuracy of climate model products.For temperature and precipitation predictions in China,the correlation coefficient of 1-7 months lead predictions is increased by 0.1-0.5.Among these improvements,the Root Mean Square Error(RMSE)of temperature is decreased by 0.5-1.0℃,representing a reduction rate of 20%—30%.For precipitation,the correlation coefficient is increased by 0.1-0.2(an increase of 10%—20%),and the RMSE is decreased by 0.1-1.0 mm/d(a reduction rate of 3%—30%),with the RMSE reduction rate reaching 30%—50%in Eastern and Southeastern China.For the ENSO index,the correlation skill for forecasts with a lead time of 1-7 months is enhanced by 5%—7%,and the RMSE at a lead time of 7 months is reduced by 50%,suggesting that the model effectively addresses the issue of excessive oscillation amplitude of the ENSO index in the original CMA-CPSv3 model.Furthermore,this study explicitly identifies a limitation of the CNN model,i.e.,excessive intensity smoothing,when applied to the correction of extreme climate events,and proposes multi-dimensional directions for future optimization.It thus provides a technical solution that integrates scientific rigor and practical applicability for operational post-processing of CMA's climate models.

Interactions between atmospheric chemical compounds and climate have a great impact on the earth system and atmospheric chemistry. However, the online two-way chemistry-climate coupled model, an indispensable tool for quantifying chemistry-climate interactions and projecting future air quality with climate change, remains sparse due to the considerable challenge in model complexity and computational resources. We present the development and evaluation of BCC-GEOS-Chem v2.0, which couples the GEOS-Chem chemical transport model (v14.0.1) with the Beijing Climate Centre Earth System Model (BCC-ESM). Based on the modular framework of BCC-GEOS-Chem v1.0, BCC-GEOS-Chem v2.0 further couples the Harmonized Emissions Component (HEMCO) to manage anthropogenic emission inventories and natural emissions, updates the chemical mechanism, includes the feedback of aerosols and greenhouse gases, and develops the capability for high-resolution simulation. The standard chemical mechanism in the BCC-GEOS-Chem v2.0 features a comprehensive Ox-NOx-VOC-halogen-aerosol chemical scheme for the troposphere and the stratosphere. We further evaluate the performance of the BCC-GEOS-Chem v2.0 simulation in representing atmospheric chemistry and compare with the model outputs from the BCC-GEOS-Chem v1.0 and BCC-AGCM-Chem over the simulated time period (2012-2014) at a spatial resolution of T42L26 (approximately 2.8 degrees & times;2.8 degrees and 26 vertical layers with a top at 2.914 hPa). BCC-GEOS-Chem v2.0 accurately depicts the primary seasonal and spatial distributions of tropospheric ozone observed by multiple instruments, showing small global mean biases of -2.1-1.8 ppbv for mid-tropospheric (700-400 hPa) ozone concentrations relative to satellite observations, along with a high spatial correlation coefficient (r) of 0.77-0.92 for individual seasons. It also demonstrates improved performance in simulating tropospheric carbon monoxide (CO), nitrogen dioxide (NO2), formaldehyde (CH2O) and surface PM2.5 compared to both BCC-GEOS-Chem v1.0 and the BCC-AGCM-Chem. The diagnostics of tropospheric ozone budgets (a global tropospheric ozone burden of 355 Tg) and OH concentrations (0.97 & times;106 molec.cm-3) are generally consistent with observation-constrained estimates and multi-model assessment. With the inclusions of aerosol-radiation and aerosol-cloud interactions, BCC-GEOS-Chem v2.0 reproduces the expected impacts of aerosols on radiative and cloud properties, e.g., decreasing shortwave downward solar radiation and outgoing longwave radiation, increasing cloud liquid water, and suppressing precipitation. The high-resolution simulation at T159L72 (approximately 0.75 degrees & times;0.75 degrees and 72 vertical layers with a top at 0.01 hPa) further improves the model capability in resolving the fine-scale plume transport dynamics and the pollution hotspot of NO2 and PM2.5, as well as the low ozone concentration in high-NOx environment in wintertime China. The development of the BCC-GEOS-Chem v2.0 model provides a powerful tool to study climate-chemistry interactions and for future projection of global atmospheric chemistry and regional air quality.

Abstract. This study provides a comprehensive description of the China Meteorological Administration Climate Prediction System version 4 (CMA-CPSv4), which is developed based on the fully coupled global climate-aerosol Beijing Climate Center Earth System Model (BCC-ESM1). It is updated from its previous version, CMA-CPSv3, which was based on the high-resolution Beijing Climate Center Climate System Model version 2 (BCC-CSM2-HR). In contrast to CMA-CPSv3, CMA-CPSv4 is capable of simulating the dynamic evolution of aerosols and their feedback on the climate system. This study aims to evaluate the reproducibility of atmospheric aerosols in CMA-CPSv4 under the forcing of observed atmospheric circulation. The 20-year simulations for the period 2001–2020 are conducted. The results show that CMA-CPSv4 reasonably captures the global spatial distribution and temporal variations in mass concentrations for five categories of dust, sea salt, sulfates, organic carbon, and black carbon, as well as aerosol optical depth (AOD). In East Asia, simulated fine-mode particulate matter PM2.5 concentrations are in good agreement with the CMIP6 multi-model ensemble mean (MME), although dust concentrations over the Taklamakan–Mongolia–North China regions are slightly underestimated, and sulfate concentrations are overestimated over the oceans. In addition, several severe dust pollution events in northern China are successfully reproduced, demonstrating the capability of CMA-CPSv4 to simulate aerosol concentrations and extreme events. The reasonable simulation of aerosol distribution is fundamental for studying aerosol-climate interactions and the impact of aerosols on numerical weather and climate prediction in our future work.

The newly developed third-generation Beijing Climate Center Climate System Model Medium Resolution (BCC-CSM3-MR) exhibits pronounced and seasonally varying sea surface temperature (SST) biases across the North Pacific, characterized by a “sandwich” pattern with warm biases east of Japan and cold biases in both the Okhotsk Sea and the subtropical North Pacific. These SST biases are closely associated with systematic errors in surface ocean currents, surface heat fluxes, and wind forcing, including a northward-shifted and weakened Kuroshio Extension, misrepresented westerlies and easterlies, and erroneous seasonal cooling and heating. Regional diagnostics reveal that cold SST biases intensify from late spring through early autumn and are subsequently transported downward, driven by underestimated stratification, excessive vertical mixing, and enhanced vertical shear of both zonal and meridional currents. Mixed-layer heat budget analyses further indicate that surface heat flux errors dominate SST bias formation in the Okhotsk Sea and the subtropical North Pacific, whereas the residual term, including the vertical processes, plays a leading role in the northern North Pacific during late autumn to spring. Sensitivity experiments using a forced ocean model provide a heuristic demonstration that biases in atmospheric forcing can contribute to surface cold signals consistent with those simulated by BCC-CSM3-MR. In particular, shortwave radiation errors produce the largest cooling, while precipitation and sea-level pressure biases have relatively minor impacts. Overall, the SST biases in BCC-CSM3-MR likely arise from a combined influence of atmospheric forcing errors and excessive oceanic vertical processes, highlighting key pathways for improving future model performance in the North Pacific.

Including sophisticated aerosol schemes in the models of the sixth Coupled Model Inter-comparison Project (CMIP6) has not improved historical climate simulations. In particular, the models underestimate the surface air temperature anomaly (SATa) when anthropogenic sulfur emissions increased in 1960-1990, making the reliability of the CMIP6 projections questionable. This cooling bias is largely attributable to the unreasonable simulated atmospheric sulfate burden changes. Sulfate burden anomaly are closely linked to both sulfate and SO2 deposition processes. Intensified sulfate deposition directly reduces atmospheric sulfate loading, while enhanced SO2 deposition limits precursor availability for sulfate formation by oxidation. These deposition processes regulate sulfate concentrations directly and indirectly. The systematically underestimated sulfate turnover time in CMIP6 models suggests that refining SO2 deposition process rather than sulfate deposition would be a more scientific approach for model improvement. This is supported by two post-CMIP6 models that show better SATa reproduction after improving the SO2 deposition parameterizations. Strong correlations between sulfate burden anomaly and SATa persist before, during, and after the 1960-1990 period. Such temporal consistency confirms the dominant role of sulfate-related physical processes across all examined time intervals.

Achieving the 2 °C climate target requires the coordination of strategies for greenhouse gases (GHGs) and air pollutants mitigation, yet their complex interactions remain insufficiently explored. BCC-ESM1 Earth system model is employed to compare global climate responses under the novel SSP2-com scenario, in which both GHGs and aerosols undergo reduction, with that under the SSP2-4.5 scenario. Moreover, the relative contributions of carbon dioxide (CO2), sulfur dioxide (SO2), and black carbon (BC) to future temperature increases are analyzed. Results reveal that compared with the SSP2-4.5 scenario, the SSP2-com scenario can stabilize the end-21st-century temperature rise well below 2 °C, primarily driven by the reduction of anthropogenic CO2 emissions. A mid-term warming rebound between 2061 and 2080 is observed due to reduced aerosol cooling. SO2 reductions result in a weakening aerosol-induced radiative forcing, driving regional warming asymmetries—particularly in northern high latitudes (up to +1.5 °C in winter). Compared to CO2-only mitigation, experiments involving SO2 reductions also exhibit stronger global precipitation increases, suggesting an acceleration of the hydrological cycle under lower aerosol loading. Energy budget analysis further indicates that SO2 mitigation results in an increase in net shortwave radiation at the top of the atmosphere by approximately 0.23 W/m2 during the mid-term (2061–2080), and consequently leads to an accumulated surface energy gain of about 0.15 W/m2. These findings highlight a key trade-off: aerosol mitigation may induce mid-term warming, but remains essential for achieving air quality and climate goals. This work underscores the necessity of balancing mid-term climate–air quality trade-offs with long-term decarbonization, offering actionable insights for policymakers to design integrated pathways align with the Paris Agreement.

Ensemble prediction has been an important tool for weather forecasting, sub-seasonal to seasonal prediction, seasonal prediction, interannual prediction and even simulation of climate change, which has garnered widespread attention in the field of meteorology. This paper introduces the ensemble prediction scheme of China Meteorological Administration Climate Prediction System version 3 (CMA-CPSv3). In this scheme, we adopt the approach of combining stochastic perturbations of physical process tendencies in the atmosphere and the air-sea flux with time-lagged initial value perturbation. Based upon the version 2 of High-Resolution Beijing Climate Centre Climate System Model (BCC-CSM2-HR), we have developed a multi-layer random perturbation ensemble prediction system with relatively good ensemble sample dispersion, stability, and reliability. Results of evaluation for hindcasts over the past 20 years show that this ensemble prediction system significantly improves the prediction of precipitation and 2 m air temperature over China, as well as the El Niu00F1o-Southern Oscillation (ENSO), Indian Ocean Dipole (IOD) and Asian Monsoon. In particular, the random perturbation of air-sea flux shows a positive effect on improving the prediction skills of ENSO and Southeast Asian Monsoon (SEAM) and Western North Pacific Summer Monsoon (WNPSM) indices. This study offers useful insights for further characterizing uncertainty in other component models of the climate system.

Achieving the 2 u2103 climate target requires the coordination of strategies for greenhouse gases (GHGs) and air pollutants mitigation, yet their complex interactions remain insufficiently explored. BCC-ESM1 Earth system model is employed to compare global climate responses under the novel SSP2-com scenario, in which both GHGs and aerosols undergo reduction, with that under the SSP2-4.5 scenario. Moreover, the relative contributions of carbon dioxide (CO2), sulfur dioxide (SO2), and black carbon (BC) to future temperature increases are analyzed. Results reveal that compared with the SSP2-4.5 scenario, the SSP2-com scenario can stabilize the end-21st-century temperature rise well below 2 u2103, primarily driven by the reduction of anthropogenic CO2 emissions. A mid-term warming rebound between 2061 and 2080 is observed due to reduced aerosol cooling. SO2 reductions result in a weakening aerosol-induced radiative forcing, driving regional warming asymmetriesu2015particularly in northern high latitudes (up to +1.5 u2103 in winter). Compared to CO2-only mitigation, experiments involving SO2 reductions also exhibit stronger global precipitation increases, suggesting an acceleration of the hydrological cycle under lower aerosol loading. Energy budget analysis further indicates that SO2 mitigation results in an increase in net shortwave radiation at the top of the atmosphere by approximately 0.23 W/m2 during the mid-term (2061u20152080), and consequently leads to an accumulated surface energy gain of about 0.15 W/m2. These findings highlight a key trade-off: aerosol mitigation may induce mid-term warming, but remains essential for achieving air quality and climate goals. This work underscores the necessity of balancing mid-term climateu2015air quality trade-offs with long-term decarbonization, offering actionable insights for policymakers to design integrated pathways align with the Paris Agreement.

Belg is the primary rainy season in southern Ethiopia and the secondary rainy season in the remaining regions. The major water supply for Belg is from the Southwest Indian Ocean, leading to the strong rainfall over southwestern and southern Ethiopia. This study investigates the spatiotemporal variability of Belg over the common period 1985–2023 across Ethiopia. The first and second typical anomalous modes of Belg account for 41.9

Accurate simulation of land surface processes is pivotal for advancing climate model fidelity and projecting hydrological and ecological responses to climate change. In this study, we incorporate a dynamic root water uptake scheme (DROOT) into the Beijing Climate Center Climate System Model (BCC-CSM), enabling dynamic root distribution and plant water stress responses. This approach provides a more physiologically realistic representation of root-mediated water uptake than conventional static root parameterizations. Model performance was assessed through simulations spanning 1990–2014, focusing on key variables: soil moisture (SM), latent heat flux (LE), gross primary productivity (GPP), precipitation (PR), 2-m air temperature (T2M), and downward shortwave radiation (SW). Our results demonstrate that DROOT substantially enhances SM simulations, particularly in regions where the original model exhibited significant biases, such as the Amazon and mid-latitude zones. Tropical regions also show marked improvements in LE and GPP simulations. Although DROOT’s influence on PR and SW remains marginal, it effectively mitigates warm biases south of 50°N. Furthermore, the scheme refines vegetation’s role in the land–atmosphere water cycle by intensifying SM-LE coupling in semi-arid regions while attenuating the direct PR-SM relationship. This study highlights the critical role of accurately representing land surface ecohydrological processes in climate modeling.

This study evaluates the ocean climatology simulated by the Beijing Climate Center Climate System Models (BCC-CSMs) participating in phases 5 and 6 of the Coupled Model Intercomparison Project (CMIP5 and CMIP6). CMIP6 BCC models generally outperform CMIP5 ones in reproducing ocean states. The CMIP6 high-resolution model, BCC-CSM2-HR, with an enhanced ocean component, exhibits the best simulation performance overall. Specifically, only BCC-CSM2-HR can accurately reproduce the southern equatorial current in the Pacific Ocean, implying the benefits of an enhanced ocean component. Persistent biases are also identified in BCC models across CMIP5 to CMIP6, including substantial biases in sea surface salinity in the Arctic Ocean, warm biases in the intermediate and deep ocean, and notable salinity biases in the northern Indian Ocean. These biases are also commonly presented in other CMIP5 and CMIP6 models. Furthermore, this study evaluates how BCC models simulate modes of climate variability, such as ENSO (El Niño–Southern Oscillation), PDO (Pacific Decadal Oscillation), and NPGO (North Pacific Gyre Oscillation). Future plans are also outlined, including the online integration of an ocean surface wave model and the refinement of model resolution, for development efforts aimed at bolstering the accuracy and reliability of BCC model simulations of ocean climatology.

Achieving the Paris Agreement’s 2 °C target demands regionally tailored climate policies and proven negative emission strategies. Here, we use a novel SSP2-com scenario that integrates updated emissions trajectories, China’s carbon neutrality pledge, and mid-to-late 21st century carbon dioxide removal (CDR) deployment to assess Earth system responses under a 2 °C-aligned pathway. Employing a state-of-the-art Earth system model, we demonstrate that SSP2-com limits end-of-century warming to 1.87 °C (below 2 °C), 0.51 °C and 1.47 °C lower than SSP2-4.5 and SSP3-7.0, respectively. Despite a reduction in global carbon sinks, SSP2-com’s stringent mitigation avoids 0.08 °C and 0.29 °C of warming relative to SSP2-4.5 and SSP3-7.0, respectively, attributed to China’s accelerated decarbonization. Our findings suggest that combining ambitious regional action (e.g. China’s 2060 net-zero goal) with global CDR deployment provides a critical pathway to stabilize below 2 °C, yet underscores unresolved risks from uneven cooling efficacy and ecosystem feedbacks. This work advances scenario design by reconciling national climate pledges with global targets, offering actionable insights for policymakers to prioritize equity-driven mitigation.

This study investigates the impacts of modifying the deep convection scheme on the ability to simulate the Madden–Julian Oscillation (MJO) in the Beijing Climate Center Climate System Model version 2 with a medium resolution (BCC-CSM2-T159) and a high resolution (BCC-CSM2-T382). On the basis of the original deep convection scheme, a modified scheme is suggested, which involves the transport processes of deep convective cloud water. The liquid cloud water that is detrained is transferred horizontally to its neighboring grids, and a portion of the cloud water that is horizontally transported is allowed to be transported downward into the lower troposphere. Both BCC-CSM2-T159 and BCC-CSM2-T382 with the modified deep convection scheme perform better than that used the original deep convection scheme in reproducing the major features of the MJO, such as its spectrum, period, intensity, eastward propagation and life cycle. Further analysis shows that those pronounced improvements in the MJO features in both BCC-CSM2-T159 and BCC-CSM2-T382 with the modified scheme are caused by transport processes of deep convective cloud water. The modified deep convection scheme enhances moisture and energy exchange from the lower troposphere to the upper troposphere around convective cloud, and promotes the convergence of moisture in the lower troposphere to the east of the MJO convection center, and then induces eastward propagation of the MJO. The comparisons between the coupled experiments and their corresponding experiments following Atmospheric Model Intercomparison Project (AMIP) simulations indicated that atmosphere–ocean interactions are also important to improve MJO simulations in the models.

This study evaluates the ability of 23 climate models from phase 6 of the Coupled Model Intercomparison Project (CMIP6) in simulating extreme climate events over China. The multimodel ensemble (MME) performs better than most individual models in reproducing the climatological mean distribution of all extreme indices. The MME can reproduce well the climatological mean distributions of five extreme climate indices over China, including annual total precipitation (PTOT), maximum consecutive 5-day precipitation (RX5), simple daily intensity (SDII), maximum daily maximum temperature (TXX), and minimum daily minimum temperature (TNN), with Taylor skill scores exceeding 0.7. SDII and TXX are the most skilful precipitation and temperature extreme indices simulated by the MME, respectively. The MME has relatively lower skill in simulating the climatological mean distribution of warm days (TX90P) and cold nights (TN10P) over China. Future projections of these extreme climate indices by the end of the 21st century are explored with the MME under the SSP1-2.6, SSP2-4.5, and SSP5-8.5 scenarios. The PTOT and RX5 in northwestern China are all projected to increase by more than 30% under SSP5-8.5. R20 is projected to increase by 4-5 days over southeastern China under SSP5-8.5. There are fewer (more) consecutive dry days over north China (south China), with a change of 5 days under SSP5-8.5. The extreme temperature indices, including TX90P, TXX, and TNN, all increase with time and higher SSP scenarios. The three indices increase by 40-55%, 4-6 degrees C, and 4-7 degrees C under SSP5-8.5 over east China, respectively. The TN10P decreases by more than 6% over east China. The changes in these extreme indices under SSP1-2.6 and SSP2-4.5 are similar to those under SSP5-8.5 but with a smaller magnitude. Large uncertainties still exist in the future projections, especially under the high SSP scenarios.

The reliability of the near-land-surface air temperature (LSAT) projections from the state-of-the-art climate-system models that participated in the Coupled Model Intercomparison Project phase six (CMIP6) is debatable, particularly on regional scales. Here we introduce a method of constructing a constrained multi-model-ensemble (CMME), based on rejecting models that fail to reproduce observed LSAT trends. We use the CMME to constrain future LSAT projections under the Shared Socioeconomic Pathways 5–8.5 (SSP5–8.5) and 2–4.5 (SSP2–4.5), representing the high and intermediate scenarios. In comparison with the “raw” (unconstrained) CMIP6 multi-model ensemble (MME) mean, the impact of the observation-based constraint is less than 0.05 o C 100 years −1 at a global scale over the second half of 21 st century. However, the regional results show a wider range of positive and negative adjustments, from -1.0 o C 100 years −1 to 1 o C 100 years −1 under the SSP5–8.5 scenario. Although amplitude under SSP2–4.5 is relatively smaller, the CMME adjustment is similar to that under SSP5–8.5, indicating the scenario independency of the CMME impact. The ideal 1pctCO2 experiment suggests that the response of LSAT to carbon dioxide (CO 2 ) forcing on regional scales is responsible for the MME biases in the historical period, implying the high reliability of CMME in the 21 st century projections. The advantage of CMME is that it goes beyond the idea of “model democracy” assumed in MME. The unconstrained CMIP6 MME may be overestimating the risks of future warming over North America, but underestimating the risks over Asia.

This study evaluates the decadal prediction skill of 13 forecast systems in predicting winter precipitation over Eurasia, contributing to the Decadal Climate Prediction Project of the Coupled Model Intercomparison Project Phase 6. Northeast Asia stands out as a region with improved decadal prediction skill for forecast years 2–5 due to the initialization. Observations show anticyclonic and cyclonic wind anomalies over the North Pacific and Northeast Asia, respectively, with southwesterly flow to the east of Northeast Asia. Ten forecast systems reproduce such circulation anomalies favoring abundant winter precipitation in Northeast Asia. The significant positive (negative) correlations between the detrended Northeast Asian precipitation (NEAP) and AMV (PDO-like) time series are reproduced by seven (nine) forecast systems. However, most forecast systems underestimate the correlation between the NEAP and the AMV, and have relatively low skill in predicting the PDO. Further improvements in these aspects will help to improve the decadal prediction skill of winter precipitation over Northeast Asia. The multi-model ensemble (MME) is able to reproduce both links of NEAP with AMV and PDO-like variability. The MME demonstrates significant skill and outperforms the individual forecast systems in predicting the NEAP for all 4-year averaged periods in the range of 1–8 years, demonstrating the benefits of using the ensemble mean of multiple models.

The East Asian summer monsoon (EASM) is unique among monsoon systems that it features meridional evolution of the summer monsoon. In this study, we evaluate the performances of a Variable-Resolution Community Earth System Model (VR-CESM) regionally refined over eastern China (14 km) in reproducing the seasonal evolution of EASM precipitation over China. Compared with reference datasets, VR-CESM shows better performance than the corresponding globally uniform coarse-resolution model CESM (quasi-uniform 1 degrees), especially over western China where complex local topography exists. The northward monsoon migration is closely related to low-level southerly flows and vertical moisture advection, which are more reasonably simulated in VR-CESM. The four critical timings of the EASM (monsoon onset, withdrawal, peak, and duration) are also better captured in VR-CESM than in CESM. The corresponding spatial Pearson correlation coefficients of the four critical timings with respect to reference datasets are about 0.1 higher in VR-CESM than those in CESM. Both models are most accurate in simulating monsoon onset and least accurate at simulating the monsoon peak. The overestimated zonal thermal contrast in CESM is responsible for the earlier monsoon onset and excessive precipitation in September over the Yangtze River valley. Finer resolution in VR-CESM, especially over the Tibetan Plateau (TP), appears to be a main factor in simulating better zonal thermal contrast and seasonal evolution of the EASM.

The surface air temperature (SAT) trend on the Tibetan Plateau (TP) was 3.45°C 100 years −1 from 1961 to 2014. The multi‐model ensemble (MME) of 33 coupled models participated in the Coupled Model Intercomparison Project phase six (CMIP6) was about 1°C 100 years −1 lower than the observation. Although MME generally shows better skill in reproducing the distribution of SAT trend over TP than most of the CMIP6 models, its performance is greatly degraded by a small group of models, about 12% on average, with large biases. In this paper, the constrained multi‐model ensemble (CMME) based on a certain observation‐based threshold is used to constrain future projections of the SAT trend over TP. Compared with the MME results, the improvements in CMME are mainly over the eastern plateau in historical simulation and are relative to the reduction of the model biases to carbon dioxide (CO 2 ) forcing. Under the high‐emission SSP5‐8.5 scenario, SAT increases significantly over the entire TP. The constraint of CMME on the MME is mainly over the eastern plateau with a difference of 0.5°C 100 years −1 , about 6% of the MME results. Under the intermediate‐emission scenario SSP2‐4.5, the effect of CMME is relatively smaller, but the corresponding spatial distribution is similar to that under the SSP5‐8.5 scenario. The CMIP6 models tend to underestimate the warming trend projections over the water source regions in the northeastern plateau and should be noticed.

The experimental data of ongoing CMIP6（Coupled Model Intercomparison Project Phase 6） are widely used to study the mechanism of climate change and provide technical support for the assessment report of the Intergovernmental Panel on Climate Change（IPCC）.With more types of model experiments and more complex climate model,the amount of CMIP experimental data are also increasing rapidly.Therefore,Beijing Climate Center（BCC） has established Earth System Grid Federation（ESGF） data node to share experimental data of BCC CMIP6.BCC has three latest versions of models to participate in the project through model development in recent years.The hardware of the platform adopts a distributed storage architecture and is deployed in the demilitarized zone（DMZ） of China Meteorological Administration,which provides a strong guarantee for its network access rate and security.The data processing module mainly checks the integrity,processes the original model output and adopts the climate model output rewriter（CMOR） software to standardize the format.Thematic real-time environmental distributed data services data server is used for local storage management and data sharing,publishing metadata to ESGF index node for unified data retrieval.The data storage directory adopts hierarchical management structure with self-describing information to realize hierarchical and classified storage of different elements in different experiments.To ensure the security of data sharing,the platform is optimized based on ESGF security framework in addition to physically adding replica storage,and the needs of easy access are also considered.Totally,190 TB experimental data of BCC CMIP6 have been released and shared since the establishment of the platform.The platform has provided important technical support for BCC to participate in the CMIP6,and it has also supported scientific research in the fields of climate change simulation and prediction,weather and climate extremes,global warming and human activities.Subsequent work will provide continuous data services to the CMIP and can be extended to other related model comparison programs.It is also important to further improve the capabilities of customized data sharing services.

The El Niño-Southern Oscillation (ENSO) ensemble prediction skills of the Beijing Climate Center (BCC) climate prediction system version 2 (BCC-CPS2) are examined for the period from 1991 to 2018. The upper-limit ENSO predictability of this system is quantified by measuring its “potential” predictability using information-based metrics, whereas the actual prediction skill is evaluated using deterministic and probabilistic skill measures. Results show that: (1) In general, the current operational BCC model achieves an effective 10-month lead predictability for ENSO. Moreover, prediction skills are up to 10–11 months for the warm and cold ENSO phases, while the normal phase has a prediction skill of just 6 months. (2) Similar to previous results of the intermediate coupled models, the relative entropy (RE) with a dominating ENSO signal component can more effectively quantify correlation-based prediction skills compared to the predictive information (PI) and the predictive power (PP). (3) An evaluation of the signal-dependent feature of the prediction skill scores suggests the relationship between the “Spring predictability barrier (SPB)” of ENSO prediction and the weak ENSO signal phase during boreal spring and early summer.