River basin runoff prediction method and device

Abstract

The invention provides a river basin runoff prediction method and device. The method comprises the following steps of establishing the basic database of a river basin and dividing into rate periodic basic data and verification period basic data; acquiring at least two runoff prediction models; adjusting model parameters in the at least two runoff prediction models; acquiring at least two pieces ofrunoff prediction data corresponding to the runoff verification period basic data; calculating a precision evaluation index between runoff measured data and each piece of runoff prediction data; calculating the credibility score value of each runoff prediction model corresponding to the runoff precision evaluation index; and using a bayesian model weighted average method to carry out weighted average on the runoff prediction data and the credibility score value, and acquiring the runoff prediction value of the river basin. In the invention, the plurality of runoff prediction models are used to carry out common prediction on the river basin runoff, the model parameters are optimized, prediction precision is effectively increased, weighted average is performed on the prediction results of the plurality of runoff prediction models, and prediction precision is increased and reliable credibility is possessed.

Description

technical field [0001] The invention relates to the technical field of watershed runoff prediction, in particular to a method and device for predicting watershed runoff. Background technique [0002] Accurate forecasting of river runoff is the basis and key to determining river governance strategies, layout and application of runoff control projects, and rational allocation of runoff resources. Under the coupled driving effect of nature and man-made, the water and sediment content of rivers has changed significantly, and the relationship between runoff and sediment has undergone major adjustments. Significant changes in river runoff and sediment content have led to a series of new problems such as severe shrinkage of downstream river channels and a significant reduction in river channel flood and sediment transport capacity, seriously threatening the flood control safety of river basins. [0003] Scientific and effective selection of runoff trend prediction methods is an im...

AI Technical Summary

Problems solved by technology

At present, the research on watershed runoff prediction mainly adopts hydrological model simulation methods, such as SWAT model, MIKE model, USLE model, etc. However, due to the large differences in the internal mechanisms of the above methods to predict runoff changes, the factors affecting runoff changes considered by different methods are different. The focus is different. In addition, runoff is a dynamic system, and changes in environmental factors increase the uncertainty of runoff change predictions, resulting in low accuracy, poor reliability, and difficulty in guaranteeing the credibility of runoff prediction results from different existing methods. It is difficult to provide strong support for accurate prediction of the future runoff of rivers, and it cannot meet the basic needs of river governance practices.

Images