Method for improving forecasting precision of short temporary rainfall

Abstract

The invention belongs to the field of short temporary rainfall forecasting, and particularly relates to a method for improving the precision of short temporary rainfall forecasting, which comprises the following steps: collecting a continuous radar echo image for weather nowcasting, and converting the radar echo image into a tensor; performing superposition processing on the tensor through a four-layer MFI-LSTM network to obtain a network output tensor at the current moment; converting the network output tensor at the current moment into a corresponding radar echo map; acquiring short temporary rainfall forecast information from the newly acquired radar echo map; the MFI-LSTM network is composed of a multi-scale feature interaction module, a convolution LSTM gating mechanism and a memory feature interaction module. On the basis of the LSTM model, two brand-new structures of the multi-scale feature interaction module and the memory feature interaction module are added, so that interaction of feature information of input data, a hidden state and a memory unit is enhanced; the problem that the final radar echo intensity prediction is inaccurate due to prediction error accumulation in the radar echo extrapolation task of the original LSTM can be prevented.

Description

technical field [0001] The disclosure belongs to the field of short-imminent precipitation forecasting, and in particular relates to a method for improving the accuracy of short-imminent precipitation forecasting. Background technique [0002] Short-term precipitation forecasting refers to the forecasting of precipitation intensity in local areas within a relatively short period of time in the future (such as 0-2 hours), which can help staff in various industries such as meteorology, disaster prevention and mitigation, agriculture, and transportation to prepare countermeasures in advance. Preventing or mitigating social and economic losses caused by heavy rainfall, storms and other severe weather has high practical significance. [0003] At present, meteorological services such as short-term precipitation forecasting, lightning nowcasting, and typhoon disaster monitoring and early warning mainly rely on radar echo extrapolation. The traditional methods of radar echo extrapo...

AI Technical Summary

Problems solved by technology

[0006] In view of the deficiencies in the prior art, the purpose of this disclosure is to provide a method for improving the accuracy of short-imminent precipitation forecast, which solves the problem in the prior art that the prediction of radar echo intensity is not accurate enough and affects the accuracy of short-imminent precipitation forecast

Images