通过定量分析候尺度气候要素的可预报性上限，可为提升短期气候预测准确性与服务防灾减灾提供重要的科学依据。基于非线性误差增长理论和非线性局部Lyapunov指数方法，利用观测资料对中国候尺度气温与降水的可预报性上限进行了定量估计，并进一步分析了其空间分布的季节变化特征。结果显示，候尺度气温可预报性上限在夏、秋季节相对较高，冬、春季节相对较低，整体上显著高于降水。中国大部分地区气温可预报性上限介于15~25 d，华南部分地区可达30 d以上，呈现出显著的纬向梯度和季节差异。降水可预报性上限总体位于10~20 d，空间异质性更为突出。全年平均而言，湿态下降水可预报性上限最高，常态次之，干态最低。与气温不同，各状态下降水可预报性上限的季节差异相对较小。非线性局部Lyapunov指数能够有效刻画有限时间内误差的非线性增长过程，为理解候尺度气温降水可预报性上限及其区域差异提供动力学视角。

By quantitatively analyzing the Predictability Limit (PL) for subseasonal climate variables, this study provides an important scientific basis for improving the accuracy of short‑range climate forecasts and for supporting disaster prevention and mitigation services. Based on the theory of nonlinear error growth and the nonlinear local Lyapunov exponent, we used observational data to estimate the PL for subseasonal temperature and precipitation over China, and further examined the seasonal variations in their spatial distributions. The results show that: \mathrm{①} The annual mean PL of subseasonal temperature exhibits pronounced spatial heterogeneity across China. Over most regions, the PL ranges from 15 to 25 days and displays a north‑low-south‑high pattern. The coastal areas and southern parts of South China form the most prominent high‑value zones, where the predictability horizon can exceed 35 days. \mathrm{②} The PL of subseasonal temperature shows significant seasonal characteristics. Predictability is relatively low in spring and winter, and higher in summer and autumn. In spring, the PL is generally lower and its spatial distribution is comparatively uniform; in summer, spatial differences in PL are strongest, exhibiting a zonal “high in the north and south, low in the middle” structure. In autumn, PL reaches its highest values among the four seasons, exceeding 30 days over most of the country. In winter, an inverse pattern to that of autumn occurs, with lower values in northern and southern regions and higher values in the mid‑Yangtze River basin and southwestern China. \mathrm{③} The PL of subseasonal precipitation is overall lower than that for temperature. Unlike temperature, the seasonal differences in precipitation predictability are relatively small across different hydrological states. For both annual and seasonal averages of PL, the wet‑state PL is highest, followed by the normal state, with the dry state being lowest. The middle and lower reaches of the Yangtze and Yellow Rivers constitute low‑PL regions in the dry state but high PL regions in the wet state, indicating a clear dependence of PL on precipitation state. This study quantitatively reveals the spatial distribution patterns and seasonal variation characteristics of the predictability limit for subseasonal temperature and precipitation in China, thereby providing a dynamical perspective for deepening the understanding of these predictability limits and their regional differences.