To address the challenge of graded straight-line wind warnings in fine short-term forecasting and nowcasting, this study utilizes polarimetric S-band Doppler weather radar data together with 1-h maximum wind observations to analyze 40 convective wind events triggered by isolated severe storms (reflectivity not lower than 60 dBZ) in the spring of 2024. The aim is to extract radar precursors and early warning indicators. These storms are categorized into two groups based on the maximum surface wind gust: the strong-wind group (Force 9-13) and the non-strong-wind group (not higher than Force 7). The results indicate that the radar signatures of strong-wind events resemble those of severe-wind-producing supercells, with the majority exhibiting persistent mesocyclones and other supercell characteristics. This indicates that the isolated storms producing strong straight-line winds are primarily supercells, with significantly higher reflectivity, vertically integrated liquid water content, mid-altitude radial convergence intensity and storm motion speed than those of the non-strong-wind group. Notably, the descent of reflectivity core is a common feature of storms and cannot distinguish the intensity of straight-line winds. About 40% of the severe wind events show a Z_DR trough 42 min before the surface wind gusts, indicating that the strong downdrafts induced by the phase changes (e.g., melting and evaporation) of hail or graupel particles represent one of the formation mechanisms for straight-line winds. A relatively continuous Z_DR trough can serve as a precursor for strong straight-line winds. This study identifies the radar precursors and warning indicators, as well as their associated lead times, for isolated severe storms that produce severe straight-line winds in spring. These findings provide a scientific basis for the warning and objective algorithm design of severe straight-line winds.