A Cluster-Based Method for Precipitation Diurnal Variation Type Division and Spatial Distribution Extraction

Abstract

The invention discloses a method for classifying and extracting spatial distribution of precipitation diurnal variation types based on clustering, which includes calculating and obtaining a 24-dimensional vector corresponding to each precipitation index of each grid unit based on multi-year hourly grid precipitation data; Standardize the obtained 24-dimensional precipitation daily change data; use different clustering data, and cluster the standardized precipitation daily change data based on the KMeans algorithm; analyze the clustering results to determine the optimal number of clusters; merge the clusters Similar classes in the results, the classification scheme of daily precipitation changes is obtained, and the average daily change characteristics of each class are calculated, and the positions of the grid cells contained in each type in the clustering results obtained after merging constitute the spatial distribution map of precipitation daily change characteristics . The invention breaks through the problems of strong subjectivity in region determination and poor accuracy of extracted precipitation diurnal variation characteristics in the existing research on precipitation diurnal variation, and has broad application prospects in the fields of precipitation forecasting, meteorological disaster early warning, and the like.

Description

technical field [0001] The invention belongs to the related fields of precipitation forecasting and meteorological disaster early warning, and relates to a method for classifying daily variation types of precipitation and extracting spatial distribution based on clustering. Background technique [0002] Affected by factors such as solar radiation, the underlying surface, and topography, the distribution of precipitation in 24 hours a day is not uniform, but shows certain regularity (for example, some areas have more rain at night), and there are significant regional differences. The classification and spatial distribution information of precipitation diurnal changes can help to detect the precipitation formation mechanism in the region, verify and improve the weather forecast model, and assist in the early warning of meteorological disasters, which has important theoretical and application value. [0003] Existing precipitation diurnal variation type classification and spati...

AI Technical Summary

Problems solved by technology

[0005] Purpose of the invention: In order to overcome the problems of strong subjectivity in regional determination and poor accuracy of extracted precipitation diurnal variation characteristics in the current research on precipitation diurnal variation, provide a method for the division of precipitation diurnal variation types and spatial distribution extraction based on clustering. The data mining method of "and above" obtains a relatively objective and easy-to-understand spatial distribution pattern of precipitation diurnal variation from the continuously distributed grid precipitation data.

Images