Simulation and prediction of precipitation and temperature under RCP scenarios

With continuous societal development, carbon dioxide concentration in the atmosphere is continually increasing and the global climate is gradually warming (Pfister & Stocker 2021). Intergovernmental Panel on Climate Change (IPCC) reports show that the global surface air temperature increased by an average of 0.4–0.8 °C during the 20th century. The 20th century has been the century with the largest rate of surface temperature increase in the past millennium, and the 1990s have been the warmest decade in the past 100 years (IPCC 2014). According to the simulation results of different climate scenarios, it is estimated that the global mean temperature will increase by 0.2 °C every 10 years over the next 100 years and sea level may rise by 9–88 cm by the end of the 21st century (Schewe et al. 2010; Hateren et al. 2013). Although there remains great uncertainty regarding these predictions, global climate change will have a greater impact on agriculture, forestry, water resources, and coastal areas and has been recognized by an increasing number of countries and societies (Girma et al. 2020; Kendall & Spang 2020; Khan et al. 2020). Climate change is predicted to lead to more frequent droughts, extreme rainstorms, and other meteorological disasters (Nigussie & Altunkaynak 2019; Lochbihler et al. 2021). Climate change has been recognized by the international community as one of the most important global environmental problems (Bach 1998).

As for many other countries, China's environment, ecosystems, agriculture, and economy have been affected by climate change (Li et al. 2018). China's glaciers and frozen soil have been declining since 1970 (Wong 2015). On a regional scale, annual precipitation has undergone significant changes, whereby precipitation in arid and semi-arid regions in the west has increased over the past 30 years (Hua et al. 2016). As the Beijing–Tianjin–Hebei (BTH) region is China's capital economic center, future climate change trends in this region have attracted considerable attention from scholars (Guo et al. 2020). In recent years, with rapid urbanization, heavy rainfall and extreme rainfall events in flood seasons have occurred more frequently, and the social and economic losses caused by urban rainstorms and waterlogging have become increasingly serious (Zhang et al. 2014). Using the global CMIP5 (Coupled Model Intercomparison Project Phase5) combined with regional characteristics, studies into the future climate change trend and its spatial variation in the BTH region are of great significance for future regional climate change research and the selection of model driving factors (Chen & Sun 2015; Tilmes et al. 2015; Demory et al. 2020). Many scholars have investigated the application of global climate models in different regions and from different perspectives (Baker et al. 2021; Zaveri et al. 2021). Mumo & Yu (2020) studied the relationship between rainfall changes in the historical period of the CMIP5 model, observed rainfall in Kenya, and derived a suitable prediction model for Kenya. The feature selection method to develop a suitable climate model for predicting future precipitation and temperature in Pakistan. Sheehan et al. (2015) analyzed the main vegetation changes in the Pacific Northwest of the United States under one of the CMIP5 future climate models. Li et al. (2020) comparatively analyzed the aerosol load and optical depth based on CMIP5 simulation data. However, few scholars have analyzed the CMIP5 model for one of the BTH regions, which is one of China's most important economic centers. Jing et al. (2024) analyzed the dimensional structure of organized precipitation systems over the BTH region in summer. Peilan et al. (2024) analyzed the impacts of compound extreme weather events on summer ozone in the BTH region. Based on measured data from weather stations in the BTH region combined with CMIP5 data, in this study, we aimed to analyze the regional availability of 10 models under the two concentration scenarios of RCP8.5 and RCP4.5 and provide a reliable meteorological drive for predicting regional water resources.

This study was conducted in the BTH region of China. The region's total area is 216,500 km², accounting for 2.3% of the total area of China, and includes Beijing, Tianjin, and Hebei Province. The BTH region is bordered by the Taihang Mountains to the west and Bohai Bay to the east. The topography of the northwest and north is relatively high, and that of the south and east is relatively flat. The region belongs to the temperate semi-humid and semi-arid continental monsoon climate zone, with similar climatic characteristics and coexisting resource endowments. The development advantages complement each other, and the water resources system is in the same line. It is controlled by the Siberian continental air mass, cold and without snow, and prevailing northerly and northwesterly winds in winter. In spring, the region is affected by the Mongolian continental air mass; the temperature increases rapidly, wind speed is high, the climate is dry, and evaporation is high, often forming arid, windy, and sandy weather. Furthermore, the influence of oceanic air masses is relatively humid, with high temperatures and heavy rainfall. However, due to the inconsistency of the time, intensity, and range of influences of the Pacific subtropical high in summer, rainfall varies greatly, and droughts and floods occur. Autumn is generally cool, with less rainfall.

The model data were selected from five global climate models of CMIP5, with a horizontal resolution of 0.5 × 0.5°, processed by the statistical down-scaling method (Table 1). GFDL-ESM2M is a model used to study climate variability and change issues on a seasonal to centennial time scale and is composed of separate atmosphere, ocean, sea ice, and land component models, which interact through a flux coupler module. NorESM1-M is an isopycnic coordinate ocean model and advanced chemistry–aerosol–cloud–radiation interaction schemes. And it has a horizontal resolution of approximately 2° for the atmosphere and land components and 1° for the ocean and ice components. MIROC-ESM-CHEM reproduces transient variations in surface air temperatures, as well as the present-day climatology for the zonal-mean zonal winds and temperatures from the surface to the mesosphere. HadGEM2-ES is a coupled earth system model that includes earth system components as a standard. It is a unified atmospheric model developed and used by the Meteorological Administration for seamless applications ranging from short-range numerical weather forecasting to seasonal, inter-decadal, and centennial climate predictions. IPSL-CM5A-LR is a coupling model that includes coupling atmospheric circulation and ocean modules, including sea ice and marine biogeochemical models. The atmosphere has a regular horizontal grid of 39 vertical levels. The ocean model operates on irregular grids, enhanced around the Arctic and sub-arctic oceans and the equator. Five sets of global climate model interpolation and correction results were provided by the Inter-Sectoral Impact Model Intercomparison Project (ISI-MIP). They were extracted and converted by the Environmental Development Institute of the Chinese Academy of Agricultural Sciences to American Standard Code for Information Interchange (ASCII) codes. A total of 169 grid data points were extracted according to the scope of the study area. This model's climate change prediction experiment includes four typical concentration paths (RCP2.6, RCP4.5, RCP6.0, and RCP8.5). Each scenario assumes a set of greenhouse gases, aerosols, and chemically active gas emissions and concentrations.

For the RCP4.5 scenario, which is another climate scenario under government intervention, with total radiative forcing stabilized at 4.5 W/m² in 2100, atmospheric CO₂ concentration increased to 538 ppm, decreasing and N₂O increased to 372 ppb. The total global population reached 9 billion, and then began to decrease. In addition, the use of renewable energy and carbon capture systems and the use of fossil fuels continue to decrease. Carbon storage increases and greenhouse gas emissions also significantly decrease with the increase of forest area. Due to the implementation of the afforestation policy and increased crop yield per unit area, RCP4.5 is the only emission mode with reduced arable land. And RCP4.5 is an intermediate stable path relative to the RCP2.6 and RCP8.5 emission scenarios closer to actual development. The RCP8.5 scenario assumes the largest population, low technological innovation rate, and slow energy improvement; it has long-term high energy demand and high greenhouse gas emissions without climate change response policies. The temperature and radiation (i.e., sunshine hours) of this scenario, wind speed, and relative humidity are all affected by greenhouse gas emissions and atmospheric aerosols. Considering the above reasons, the RCP4.5 and RCP8.5 emission scenarios are selected for analysis in this study.

From the overall data, it can be concluded that the CMIP5 model does not show a satisfactory correlation in estimating the interannual variability of future temperature and rainfall in the BHT region of China. The Mann–Kendall trend test shows an annual downward trend in the observed annual precipitation; however, the CMIP5 model shows an overall upward trend. This question also shows that for a certain region, it is not advisable to use the data evaluated by a single model to study future climate change trends. Rather, different aspects of the model should be summarized, and rational use should be made.

The error of the CMIP5 model in simulating the observed rainfall and temperature was associated with the convective parameterization scheme. The purpose of designing these schemes is to show the average effect of convection on the model grid, which is transferred to the grid scale as momentum, temperature, and humidity increase (West et al. 2007). According to Hohenegger et al. (2008), these convection schemes were developed for tropical and coarse-resolution models, and are therefore not suitable for temperate regions and high-resolution models. This requires the modeling team to reparameterize the convection scheme to manage the uncertain convection activity in the atmospheric column, resulting in unrealistic results in the model (Kendon et al. 2012).

For regional climate change, global climate models are used to predict, with low resolution and greater uncertainty, when applied in a smaller area. The uncertainty of climate change projections is due to many factors, including the uncertainty of emission scenarios, the natural variability of the climate system, and the uncertainty of the climate model itself (Ito et al. 2020). The uncertainty of greenhouse gas emission levels is an important source of uncertainty in future climate predictions and is related to the uncertainty of future technological development and socioeconomic policies (Cai et al. 2019). There are three main sources of uncertainty in climate prediction: The uncertainty of the scenario (i.e., the uncertainty of future greenhouse gas emissions); the uncertainty of the natural variability within the climate system; and the uncertainty that characterizes the climate process (i.e., the uncertainty of the structural framework).

Based on the CMIP5 model output of historical simulation and future prediction under two RCP scenarios, this study analyzed the future changes and related uncertainties of rainfall and temperature in the BHT region on an annual average scale. The results showed that different statistical indicators lead to different regional matching degrees in the model. The models were inconsistent at different time scales. Compared with the interannual variation, the CMIP5 historical model is more consistent with the observed value in simulating the mean annual temperature variation.

A comparative analysis of 10 climate models under the two scenarios conducted through the Taylor chart algorithm recommended by CMIP combined with historical observation data in the BHT region showed that the MIROC-ESM-CHE model is more consistent with the evaluation of future precipitation capacity and that the HadGEM2-ES model is correct. The future RCP4.5 temperature evaluation is more consistent, and the MIROC-ESM-CHEM model is more consistent with the future RCP8.5 temperature evaluation. This provides an important meteorological driving factor for the analyses of future water resources in the study area.

Analyses of future climate change trends in the study area showed that multiple models have relatively small differences in predicting future temperatures. For the annual average, the increase under RCP8.5 is greater than that under RCP4.5, indicating that radiative forcing has a more significant impact on temperature. The spatial temperature distribution in the study area maintains a strong consistency with the topography and geomorphology, and the temperature in the plain area is relatively high, indicating that urbanization is one of the critical factors affecting regional temperature.

We acknowledge the reviewers and editors for their patience and valuable advice on improving the quality of this paper and for teaching us how to write papers of higher quality. Financial support for this work was provided by the Scientific Research Project of Colleges and Universities in Anhui Province (2023AH051181), the Scientific Research Foundation for High-level Talents of Anhui University of Science and Technology (No:2022yjrc47), the Academician Workstation in Anhui Province of Anhui University of Science and Technology (No: 2022-AWAP-07), and the Independent research project of State Key Laboratory of Simulation and Regulation of Water Cycle in River Basin (NO. SKL2022ZD02) References.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.