A Diagnostic Study of Numerical Weather Prediction Systematic Forecast Errors in Heat

It is proved, that NWP systematic forecast errors in that are due to the differential equivalent potential temperature variation processes between the model atmosphere and real atmosphere during the period of validity. And it is also proved that these forcing factors which evoke the thermal variation can be split into various linear terms according to the non-acceleration theorem in a primitive equation. By applying this scheme to diagnose the forecast products of the T42L9 model (the operational model of NMC in China) in January and in July 1992, it was indicated that the model has the ability of forecasting thermal reasonable zonal mean, but there are still great errors in zonal belts, especially in the upper troposphere and low stratosphere levels in mid-latitude as well as near surface. The results of analysis by employing this scheme give reasons for the systematic forecast errors and the potential way of improving the model. And it is also shown that this scheme by applying non-acceleration theorem is an efficient tool to diagnose the NWP models physical processes.