Lower atmospheric gravity waves (GWs) can significantly impact waves in the middle and upper atmospheres and are vital for turbulence generation. This paper puts the spotlight on the spatial–temporal variability of tropospheric GW total energy (ET) and its potential sources above four regions of interest (ROIs) gathered from high‐resolution radiosonde observations from the China Radiosonde Network during the years 2016–2019. The seasonality of ET above four ROIs shows different characteristics and is dependent on latitudes and underlying terrains, reaching its maximum identified in the winter at middle latitudes. Interestingly, the annual cycles of the maximal ET shift from 35°N in October to 25°N in March of the next year, triggered by the shift in the winter subtropical jet. Based on the random forests regressor, the jet stream between 200 and 125 hPa likely serves as the primary source for the observed GWs above the ROIs with low and middle latitudes, with relative contributions of around 60%. However, the Kelvin–Helmholtz instability between 800 and 125 hPa could be the most recognized source of GWs and contributes around 68.4% to the observed energy. During the rainy season, the ET under scenarios of convective precipitation is around 20% larger than the other. As well, as the near‐surface or low‐level wind interacts with a mountain barrier over the Tibetan Plateau region, 12.4% of the observed ET is attributed to the strength of the low‐level wind.

Differentiating local‐scale precipitation (LP) and nonlocal‐scale precipitation (NLP) and understanding their corresponding prestorm environment are of importance for accurate severe weather analysis and forecasting. However, the difference in the transient prestorm environments of LP and NLP is largely undetermined, thereby fundamentally limiting the predictability of convective storms. The present study focuses on the precursor signals of LP and their differences from those of NLP based on explicit observational analyses using high‐resolution radiosonde measurements from the China radiosonde network with a vertical resolution of 5‐m, combined with 1‐min rain gauge data from the China rain gauge network during 2013–2020. LP and NLP can be recognized by most proximity sounding parameters. For example, the upcoming LP is characterized by larger most‐unstable convective available potential energy (MUCAPE), K index, and total precipitable water (TPW), and a lower lifting condensation level (LCL), compared to the NLP scenario. By taking proximity sounding parameters as inputs and precipitation types as learning targets, a nu‐support vector machine algorithm can effectively predict LP or NLP events, achieving an overall precision of 97%. However, the precision significantly drops by approximately 20% after removing the variable of low‐level wind shear, indicating the crucial role of wind observation when predicting the LP event. Furthermore, more intensive LPs are characterized by larger MUCAPE, K index, TPW, and moist static energy, and lower LCL. Among others, MUCAPE is the most important feature, according to the gradient boosting machine algorithm.

The planetary boundary layer (PBL), the lowermost part of the troposphere, is the main place where a wide range of complicated physical and chemical processes occur involving the exchange of turbulence mixing, advection and convection of momentum, water, trace substances, and pollutants, and thus is generally thought of as being directly influenced and coupled with surface forgings (Garratt, 1994;Lee & Wekker, 2016;Stull, 1988). The timescale that PBL responds to ground surface ranges from minutes to hours (Garratt & Brost, 1981). The boundary layer height (BLH), the maximum vertical extent of PBL, is of critical importance to the dispersion of air