Publications

Shi, MX; Liu, XG; Fan, P; Zhang, W; Gao, W (2023). Evaluation and application analysis of kilometer-scale convective parameters derived from a statistical downscaling method over Central China. CLIMATE DYNAMICS, 61(9-10), 4563-4586.

Abstract
Thunderstorms are usually accompanied by heavy precipitation and other disasters. Therefore, it is very important to study the thermodynamic and dynamic mechanisms that control the occurrence and development of storms. However, the relatively coarse resolution of the reanalysis may limit the evaluation and application of its derived convective parameters at regional and local scales, especially in high mountains. Given that the effect of grid averaging on ERA5 can be well-reflected in the monthly mean, the paper proposes a methodological framework to achieve the statistical downscaling of convective parameters (SDCP). After the convection-related variables are decomposed, we introduce a multi-scale geographically weighted regression (MGWR) model and take the Normalized Difference Vegetation Index (NDVI) and the Digital Elevation Model (DEM) as explanatory variables to improve the spatial resolution of monthly-scale convective parameters. The convective parameters chosen for the downscaling experiments are closely related to thunderstorm occurrence, involving instability, moisture, and wind shear. ERA5 and SDCP datasets' quality is evaluated based on measurements available from 11 radiosonde stations over 10 years. We also explore the climatology of thunderstorm environments and the corresponding trends. The results show that the accuracy of the SDCP data obtains additional added value in complex-terrain areas. Compared with ERA5, the downscaled results have a finer spatial structure. A strong but spatially and scale-varying relationship exists between the monthly mean convective parameter and the resampled NDVI and DEM. Considering the very high predictive ability of the MGWR model, the performance of the SDCP algorithm mainly depends on the accuracy of the ERA5 data. Our study demonstrates the possibility of high spatial resolution land surface features observed by remote sensing satellites in improving the spatial resolution of thermodynamic parameters.

DOI:
10.1007/s00382-023-06820-9

ISSN:
1432-0894