摘要
A right annual cycle is of critical importance for a model to improve its seasonal prediction skill. This work assesses the performance of the Grid-point Atmospheric Model of IAP LASG （GAMIL） in retrospective prediction of the global precipitation annual modes for the 1980 2004 period. The annual modes are gauged by a three-parameter metrics： the long-term annual mean and two major modes of annual cycle （AC）, namely, a solstitial mode and an equinoctial asymmetric mode. The results demonstrate that the GAMIL one-month lead prediction is basically able to capture the major patterns of the long-term annual mean as well as the first AC mode （the solstitial monsoon mode）. The GAMIL has deficiencies in reproducing the second AC mode （the equinoctial asymmetric mode）. The magnitude of the GAMIL prediction tends to be greater than the observed precipitation, especially in the sea areas including the Arabian Sea, the Bay of Bengal （BOB）, and the western North Pacific （WNP）. These biases may be due to underestimation of the convective activity predicted in the tropics, especially over the western Pacific warm pool （WPWP） and its neighboring areas. It is suggested that a more accurate parameterization of convection in the tropics, especially in the Maritime Continent, the WPWP and its neighboring areas, may be critical for reproducing the more realistic annual modes, since the enhancement of convective activity over the WPWP and its vicinity can induce suppressed convection over the WNP, the BOB, and the South Indian Ocean where the GAMIL produces falsely vigorous convections. More efforts are needed to improve the simulation not only in monsoon seasons but also in transitional seasons when the second AC mode takes place. Selection of the one-tier or coupled atmosphere-ocean system may also reduce the systematic error of the GAMIL prediction. These results offer some references for improvement of the GAMIL seasonal prediction skill.
A right annual cycle is of critical importance for a model to improve its seasonal prediction skill. This work assesses the performance of the Grid-point Atmospheric Model of IAP LASG （GAMIL） in retrospective prediction of the global precipitation annual modes for the 1980 2004 period. The annual modes are gauged by a three-parameter metrics： the long-term annual mean and two major modes of annual cycle （AC）, namely, a solstitial mode and an equinoctial asymmetric mode. The results demonstrate that the GAMIL one-month lead prediction is basically able to capture the major patterns of the long-term annual mean as well as the first AC mode （the solstitial monsoon mode）. The GAMIL has deficiencies in reproducing the second AC mode （the equinoctial asymmetric mode）. The magnitude of the GAMIL prediction tends to be greater than the observed precipitation, especially in the sea areas including the Arabian Sea, the Bay of Bengal （BOB）, and the western North Pacific （WNP）. These biases may be due to underestimation of the convective activity predicted in the tropics, especially over the western Pacific warm pool （WPWP） and its neighboring areas. It is suggested that a more accurate parameterization of convection in the tropics, especially in the Maritime Continent, the WPWP and its neighboring areas, may be critical for reproducing the more realistic annual modes, since the enhancement of convective activity over the WPWP and its vicinity can induce suppressed convection over the WNP, the BOB, and the South Indian Ocean where the GAMIL produces falsely vigorous convections. More efforts are needed to improve the simulation not only in monsoon seasons but also in transitional seasons when the second AC mode takes place. Selection of the one-tier or coupled atmosphere-ocean system may also reduce the systematic error of the GAMIL prediction. These results offer some references for improvement of the GAMIL seasonal prediction skill.
基金
Supported by the National Natural Science Foundation of China under Grant No. 40605022
the "973" Project of the Ministryof Science and Technology of China under Grant No. 2006CB403600
the Special Research Program for Public Welfare(Meteorology) under Grant No. GYHY200706005
