Clustering-based daily precipitation variation type classification and spatial distribution extraction method

Abstract

The invention discloses a clustering-based daily precipitation variation type classification and spatial distribution extraction method. The method comprises: on the basis of many years of hour-by-hour grid precipitation data, carrying out statistics to obtain a 24-dimensional vector corresponding to each precipitation index in each grid unit; standardizing the obtained 24-dimensional precipitation daily variation data; with different clustering data, clustering the standardized precipitation daily variation data based on a KMeans algorithm; analyzing clustering results and determining an optimal cluster number; merging similar classes in the clustering results to obtain a daily precipitation variation classification scheme, calculating each class of average daily variation characteristic,and forming a spatial distribution map of daily precipitation variation characteristics by positions of grid units included by all types in the obtained clustering result after merging. According tothe invention, problems of high subjectivity of area determination and poor accuracy of the extracted daily precipitation variation characteristic in the existing daily precipitation variation study are solved; and the clustering-based daily precipitation variation type classification and spatial distribution extraction method have the broad application prospects in fields of precipitation forecasting and 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