Large-scale underground water reserve remote sensing dynamic monitoring and driving factor quantitative splitting method

Abstract

The invention provides a large-scale underground water reserve remote sensing dynamic monitoring and driving factor quantitative splitting method, which comprises the following steps of: acquiring regional land water reserve change based on GRACE gravity satellite inversion, and performing calculation to obtain a large-scale regional underground water reserve change result; and based on a multi-source information parameter regression model, establishing a coupling relation model between the driving factors and the regional underground water reserve change result, and determining the contribution degree of each driving factor. According to the method, the multi-collinearity problem among the driving factors is solved, the regression model coefficient reasonability is improved, meanwhile, the observation value variance is reduced, the model fitting precision is improved, and the contribution degree of the main driving factors causing regional underground water reserve changes can be reflected more accurately.

Description

technical field [0001] The invention relates to the technical field of hydrological ecological protection, in particular to a method for remote sensing dynamic monitoring of large-scale groundwater reserves and quantitative splitting of driving factors. Background technique [0002] The launch of GRACE and GRACE-Followon gravity satellites provides an effective way for remote sensing monitoring of large-scale groundwater storage changes. Traditional groundwater level monitoring methods include observation well method, remote sensing image extraction method, numerical simulation method, etc. The above methods are limited by the number of observation wells, can only extract shallow soil water, and have high data requirements, which require a lot of time. Manpower and material resources make it difficult to accurately obtain large-scale or reflect the actual situation of regional groundwater storage changes. The GRACE gravity satellite has the characteristics of uniform global...

AI Technical Summary

Problems solved by technology

Among them, the partial least squares method is used as an unbiased estimate, and there is no deviation in the observed value, but the variance is large, which makes the observed value and the real value differ greatly; while the ridge regression analysis abandons the unbiasedness of the least squares method, adding a Deviation value, to reduce the standard error, so that the regression coefficient is more realistic, but at the same time the accuracy is also reduced

Images