Beijing Intelligent Grid Temperature Objective Prediction Method (BJTM) and Verification of Forecast

Based on the daily maximum and minimum temperature forecasting products of the European Center Medium-Term Weather Forecasting (ECMWF), an intelligent temperature forecasting method have been developed for intelligent grids in Beijing(BJTM), which using the linear regression method and Kriging interpolation. The observation data of 55 appraisal stations in Beijing from January to December in 2017 were used as a test, BJTM had significantly improved the temperature forecast of the ECMWF fine-grid model. The accuracy of the daily maximum and minimum temperature forecast for 1 to 7 days had been improved respectively 25.4% and 11.2%; For the 15 national stations, the accuracy of were improved by 16.0% and 6.4%, and it was 4.3% and 4.8% higher than the forecaster. Then the stations data were interpolated into a grid data of 1km spatial resolution with Kriging interpolation. And we considered the elevation geography information to form an intelligent grid temperature forecast product in Beijing. A refined temperature forecasting method was developed based on the one-dimensional linear regression, by using the highest and lowest daily temperature forecast products of the European Center Medium-Term Weather Forecasting (ECMWF) and the 55 assessed observation stations in Beijing. The data from November 1, 2013 to December 31, 2016 were used to establish the forecasting equation, while the data from November 1, 2013 to December 31, 2016 were used as a test.