泰勒图（Taylor diagram）

泰勒图：泰勒图(Taylor,2001)常用于评价模型的精度，常用的精度指标有相关系数，标准差以及均方根误差(RMSE)。一般而言，泰勒图中的散点代表模型，辐射线代表相关系数，横纵轴代表标准差，而虚线代表均方根误差。泰勒图一改以往用散点图这种只能呈现两个指标来表示模型精度的情况。从更广义地来讲，泰勒图可以延展到需要用二维平面呈现三维数据的应用场景。这一点与三元图有异曲同工之妙。

功能：相关/分布描述

案例1

案例2

这里只描述了两种指标，归一化标准差和相关系数
A Taylor diagram provides a statistical summary of how well patterns match and is widely used to compare the outputs from models with observations1. Supplementary Figure 1 shows a comparison of the globally averaged seasonal time series (climatology for 1980 2005) from models and the GRACE observations (climatology for 2003 2012). First, the correlation coefficients for the radial line denote the relationship between GRACE and the models, indicating similarities on seasonal time scales. Second, the normalized standard deviations (standard deviation of the model divided by that of the observations) on the x- and y-axes indicate if the variance of the model is larger or smaller than those in the observations. The points inside the dashed curved line mean that the model simulations have lower variability than the GRACE observations. By contrast, the models with normalized standard deviation are higher than one has higher variability than the GRACE observations. For example, GISS observations have more than two times the standard deviation of the GRACE observations because of a larger seasonal cycle. The reasons for the biases in the model could be from simulated atmospheric forcing from the configurations, such as soil type, layers, and depths, or parameterizations in different land surface models. On combining the correlation and standard deviation, the points closer to the reference point on the x-axis show a better agreement, making the ensemble mean of the CESM-LE one of the realistic models representing terrestrial water storage. Supplementary Figure 1 shows that the CESM agrees with observations for the simulated seasonal cycle of total water storage.