Estimating the local predictability of heatwaves in south China using the backward nonlinear local Lyapunov exponent method

Heatwaves can have significant negative socioeconomic impacts and damage ecosystems. Consequently, accurate forecasts of such events are of great importance. In this study, the backward nonlinear local Lyapunov exponent (BNLLE) is used to investigate the local predictability of summer (June, July, and August) heatwave events over south China (SC). The predictability of the summer 2-m temperature (T2m) was quantified first. The results show that the summer T2m has a high predictability, and its predictability limit can reach 12 days, which is close to the upper limit of atmospheric predictability. Nine major heatwave events that occurred over SC were analysed, and their error dynamics and local predictabilities were studied using the BNLLE method. The local predictability limits covered a wide range, varying from 4 to 12 days. Analysis of the forecast error growth and associated rates revealed that the nine major heatwave events showed different error dynamics, thereby leading to the different local predictability limits. In addition, the regional dynamical information associated with these nine heatwave events was investigated and the error dynamics were found to have a heterogeneous spatial distribution. That is, both the error growth and their rates in the northern regions were larger than those in the southern regions. The regional dynamical information associated with the summer T2m also showed the same pattern. Therefore, the northern regions of SC are sensitive to error growth and limit the heatwave forecast skill.

Automated summary

In this study, the local predictability of summer heatwave events in south China was investigated using the BNLLE method. The results showed the high predictability of summer temperature and the wide range of local predictability limits for major heatwave events. The error dynamics and regional dynamical information revealed the heterogeneous spatial distribution of error growth and limitations in heatwave forecasting in the northern regions of south China.