Abstract: A high-resolution weather research forecasting model (WRF) was used to simulate and analyze 14 short-time summer precipitation events that occurred in the Shanxi Province in the past 5 years. Based on this, the conditional nonlinear optimal perturbation (CNOP) method was used to identify the initial sensitive areas of these precipitation events. Then, the common sensitive areas were determined based on the overlap degree of the initial sensitive areas of all events. Finally, an encrypted observation layout was designed based on the common sensitive areas mentioned above, and its effectiveness was verified through observation system simulation experiments (OSSE). The results indicated that the WRF model can simulate these 14 short-time summer precipitation events, and significant differences in the sensitive areas were identified using the CNOP method during these events. When the scope of sensitive areas was expanded, the overlap between sensitive areas across different events increased. Grid points with overlap exceeding 50% were selected as common sensitive areas. The results showed that these common areas were mainly located in Datong City in the northwest of Shanxi, Lvliang City in the west, and Taiyuan and Jinzhong areas in the central region. Based on the aforementioned common sensitive areas, two intensive observation networks, one within and one outside the sensitive areas, were designed separately. The OSSEs demonstrated that the intensive observation layout within sensitive areas significantly improved precipitation forecasting skills. Furthermore, an intensive observation network with a resolution of up to 500 m was designed within these sensitive areas. Subsequent OSSEs confirmed that assimilating high-resolution intensive observation data from the sensitive areas identified at lower resolution effectively enhanced the forecasting skill for short-time summer precipitation in Shanxi. This indicates that the sensitive areas derived at low resolution are applicable at high resolution and can guide the layout design of encrypted observations at high resolution.