A local numerical weather prediction product revision method based on depth learning

Abstract

The invention discloses a local numerical weather prediction product revision method based on depth learning, which comprises the following steps of constructing a training data set according to the historical numerical weather prediction product and the corresponding historical station observation data; training the depth learning network model according to the training data set to obtain a revised model; extracting a data segment of a real-time numerical weather forecast product forecast field as input data and inputted into the revised model, and outputting the revised data segment throughthe revised model, and taking the outputted revised data segment as a revised result. The method of the invention utilizes the non-linear mapping ability of the depth learning network and the information extracting ability of the raster data to correct the numerical weather prediction product element value based on the actual observation data of a plurality of stations, and overcomes the problemsexisting in the prior art.

Description

technical field [0001] The invention belongs to the technical field of meteorology, in particular to a method for correcting local numerical weather forecast products based on deep learning. Background technique [0002] In the modern weather forecasting business, numerical weather prediction is becoming more and more important. Numerical weather prediction is based on the actual situation of the atmosphere, under certain initial value and boundary value conditions, through numerical calculation, to solve the fluid dynamics and thermodynamic equations, and make quantitative and objective forecasts of the future. Various analysis and forecast products obtained by numerical prediction methods are called numerical weather prediction products. [0003] With the development of numerical weather prediction, meteorological elements including temperature, pressure, humidity, and wind can obtain forecast conclusions for several hours to tens of hours in the future through numerical ...

AI Technical Summary

Problems solved by technology

[0005] Since the numerical weather prediction product is the product of theoretical calculation, in the actual use process, due to system deviation, local influence and other factors, there will be a certain deviation between the numerical weather prediction result and the actual observation value, and the station forecast result can be corrected by manual means , but the correction process requires the accumulation of a large amount of historical site observation data, which consumes a lot of manpower and cannot be corrected in a large area

Images