Mid- and long-term forecasting method of groundwater level change based on support vector machine binary data assimilation

Abstract

The invention discloses a mid- and long-term forecasting method for groundwater level changes based on support vector machine binary data assimilation, belonging to the field of groundwater forecasting in the subject of hydrology and water resources. The nonlinear regression method of vector machine is used to construct a prediction model for groundwater level change; the driving factors at the forecast time are input to predict the groundwater level change at this time; when there is groundwater level observation data at the forecast time, the set of driving and output elements is assimilated and corrected to construct a univariate data assimilation Model; update and input the driving factors at the forecast time, and output the forecast value of univariate data assimilation; perform data assimilation correction again on the forecast value of univariate data assimilation to obtain the forecast value of binary data assimilation. This method is based on machine learning theory and uses data assimilation technology to integrate big data to forecast changes in groundwater level, which can be directly applied to mid- and long-term forecasting and optimal allocation of groundwater resources.

Description

technical field [0001] The invention belongs to the field of groundwater forecasting in the subject of hydrology and water resources, in particular to a medium and long-term forecasting method for groundwater level changes based on support vector machine binary data assimilation. Background technique [0002] As the largest freshwater resource in the world [Shiklomanov and Rodda, 2003], groundwater resources occupy an important position in the global water cycle [Taylor et al., 2013]. Groundwater, as the main source of drinking water, maintains the sustainable development of the ecological environment [IPCC, 2014]. With population growth and climate change, groundwater demand has increased dramatically [Konikow, 2011; Wada et al., 2011; Wada et al., 2012]. With the exploitation of groundwater, groundwater aquifers are facing a crisis of tightening [Sun, 2013]. Understanding the spatial and temporal distribution of groundwater and making mid- and long-term forecasts for gro...

AI Technical Summary

Problems solved by technology

At present, there is no relevant research on the use of binary data assimilation technology based on support vector machines in the medium and long-term forecasting of groundwater level changes

Images