In view of the shortage of rainfall data in small and medium watersheds and the difficulty of accurate flood forecasting, we proposed to use linear cointegration and set pair analysis (SPA) to interpolate and transfer the observational rainfall data under typhoon rain and non-typhoon rain, and carried out the research on flood forecasting of Tuo Ying Reservoir.
The results show that: (1) Based on the typhoon rain cointegration calculation results, it is found that there is a cointegration relationship between rainfall stations. The Nash coefficients of the cointegration simulation results and the measured rainfall are both above 0.85, and the correlation coefficient reaches 0.90. The main rain peak and the raining error are small. The calculation results are good and the integration theory can be used for data interpolation when short of rainfall data. (2) The non-typhoon rain rainfall sequence is decomposed by supplementary ensemble empirical mode (CEEMD) and SPA, and the optimal plan for data migration is drawn up according to the comprehensive connection degree, which significantly improves the effectiveness and accuracy of the migration of rainfall data, and improves flood forecasting accuracy. The method proposed in this paper provides a new idea for the interpolation and transfer of rainfall data in other regions with both typhoon storm and non-typhoon storm.