Precipitation space-time prediction method and electronic equipment

Abstract

The invention discloses a rainfall space-time prediction method and electronic equipment. The method comprises the following steps: acquiring rainfall observation data, acquiring rainfall space forecast factors, and acquiring prior distribution of spatio-temporal variation parameters; establishing a generalized linear regression model according to the rainfall observation data and the rainfall space forecast factors; constructing a Gaussian process model; according to the generalized linear regression model, the Gaussian process model and the prior distribution of the spatial-temporal variation parameters, constructing a Bayesian hierarchical model, and determining a posterior probability distribution form of the spatial point precipitation under the observation condition; and according to the posterior probability distribution form, determining a spatial point precipitation prediction result. According to the method, the precipitation space forecasting factors capable of reflecting the space change and the space-time variation parameters are fused into the generalized linear regression model, and the Bayesian hierarchical model is constructed by using the precipitation observation data of the discrete precipitation observation stations, so that the precipitation time dynamic change can be effectively represented, and the precipitation related parameter space variation process can be reflected.

Description

technical field [0001] The invention relates to the field of precipitation spatiotemporal prediction, in particular to a precipitation spatiotemporal prediction method and electronic equipment. Background technique [0002] The spatial distribution characteristics of precipitation are the main controlling factors affecting runoff simulation and a series of other hydrological problems. Among them, the uncertainty of runoff process simulation mainly comes from the uncertain precipitation input, and the most sensitive factor of flood prediction is the spatial variation of precipitation. However, it is difficult to fully understand the temporal and spatial distribution characteristics of precipitation from discrete rainfall station observations. Robust spatial interpolation techniques are needed to use known data to obtain distributed precipitation data to support regional hydrology, water resource analysis, and efficient use and management of water resources. . At the same ti...

AI Technical Summary

Problems solved by technology

[0004] However, there are many "zero values" in daily and hourly precipitation, which makes the spatial interpolation of hourly and daily precipitation more difficult, and the static interpolation method is difficult to reflect the temporal dynamic characteristics and dependencies of precipitation

Images