Root Mean Square Difference between the nine ensemble member change anomalies of the seasonal mean near-surface (2m) temperature for the 90% percentile for 2066 - 2095 relative to 1976-2005, for the JJA season, under the RCP 8.5 pathway

Root Mean Square Difference for seasonal (JJA) mean near-surface (2m) temperature (°C) change from the 90% percentile projected for 2036-2065, relative to present (1976 - 2005), under the RCP 8.5 pathway for the southern African region. To generate the image, nine coarse General Circulation Models (GCM) are downscaled to a finer spatial resolution (0.44°x 0.44°) using the Rossby Centre regional model (RCA4) forcing its lateral boundaries. The model simulated daily temperature averages, which are used to generate projections of seasonal change. The projections are generated using the high (RCP8.5) pathway which associates CO2 concentrations of approximately 950ppm by the year 2100. The associated RMSD it calculated and shows the uncertainty range of the projected model simulated residual values, and gives a relative perspective of spatial areas associated with higher and lower projection uncertainties.

本数据集表征南非区域在典型浓度路径8.5（Representative Concentration Pathway 8.5，RCP8.5）情景下，相对于基准期（1976-2005年），2036-2065年90%分位数预测的夏季（June-July-August，JJA，即6、7、8月）平均近地表（2米）气温变化的均方根差（Root Mean Square Difference，RMSD）。
为生成该可视化图像，研究采用罗斯比中心区域模式（Rossby Centre regional model，RCA4）对9套粗分辨率大气环流模式（General Circulation Models，GCM）进行降尺度处理，将其空间分辨率提升至0.44°×0.44°，并通过侧边界强迫驱动模式运行。该模式模拟的日平均气温被用于生成季节尺度的气候变化预测。本次预测采用高排放情景RCP8.5，该情景预计到2100年大气二氧化碳浓度将达到约950ppm。
所计算得到的对应均方根差（RMSD）可反映模式模拟残差预测值的不确定性范围，并可直观展现不同空间区域的预测不确定性高低分布特征。