Monthly runoff prediction method for drainage basin cascade reservoir system

Abstract

The invention relates to the field of long-term runoff prediction in drainage basins. According to the technical scheme, a monthly runoff prediction method for a drainage basin cascade reservoir system comprises the following steps: S1, obtaining data, and according to the obtained data, calculating CFS lattice point monthly rainfall and drainage basin actually-measured daily and lunar surface average rainfall; S2, fitting a regression relationship between the monthly rainfall of the CFS grid point in the drainage basin in different prediction periods and the actually measured lunar surface rainfall of the drainage basin, and using the relationship for correction of CFS prediction to obtain the lunar surface rainfall of the drainage basin in the next nine months; S3, carrying out time-space downscaling on the predicted lunar surface rainfall of the drainage basin; S4, constructing a distributed hydrological model SWAT suitable for the cascade reservoir system; and S5, predicting the monthly reservoir inflow of each reservoir of the cascade reservoir system by using an SWAT model constructed by predicting the daily scale rainfall of the sub-basin obtained in the S3. The method can predict medium-and-long-term monthly runoff of each reservoir section, and solves the problem of medium-and-long-term runoff prediction of cascade reservoirs in drainage basins.

Description

Technical field [0001] The present invention relates to the field of long-term runoff prediction in the basin, and in particular to a lattice regular basis for the Numerical prediction product CFS and SWAT model coupling. Background technique [0002] The long-term monthly laundering of complex cascade reservoir systems is difficult, and it is very important for the development of flood control and drought, and the reservoir scheduling scheme is very important, so there is an urgent need for a more accurate prediction method. The current medium and long-term moon runoff is mainly divided into a mathematically statistical pathway and the land coupling pathway. Compared with the relevant mathematical statistical pathway, the land-enhancement pathway is based on rainfall - runoff relations, and the long-term rainfall prediction introduces the hydrological model. It has made a prediction of long-term runoff, and its physical mechanism is more clear, and it is more likely to be due to...

AI Technical Summary

Problems solved by technology

[0003] At present, the more recognized rainfall prediction mainly comes from numerical weather prediction products, but most of them have the problems that the accuracy decreases with the increase of the forecast period and the temporal and spatial distribution of rainfall does not match the input of the hydrological model. Therefore, when applying long-term numerical prediction products, it is necessary to improve Accuracy and temporal and spatial precipitation scales of rainfall forecasts over longer forecast periods, i.e. post-processing

There are many correction methods for medium and long-term rainfall series, mainly mean correction method, quantile regression correction method, etc., but these methods are all based on an untested "same distribution" assumption (the measured rainfall sequence and the simulated predicted rainfall The probability distribution of the rainfall sequence is the same) a

Images