Previous analyses on the estimates of water vapor and cloud-related feedbacks in the tropics usually use observations over the Earth Radiation Budget Experiment （ERBE） period （1985-89）. To examine the sample dependence of previous estimates, the authors extend the analysis to two additional periods： 1990-94 and 1995-99. The results confirm our hypothesis, i.e., the values of the feedbacks depend on the period of data coverage. The differences in the feedbacks from cloud radiative forcings （CRFs） estimated from the three periods are particularly significant. Two possible causes for these differences are proposed. First, a regime behavior in the CRFs-Sea Surface Temperature Anomaly （SSTA） rela- tionship over the cold tongue region is revealed： when SSTA is below -0.5℃, the CRF anomalies are insensitive to the SSTA; when the SSTA is between -0.5℃ and 2.0℃, the CRF anomalies are positively correlated with the SSTA; however, when the SSTA exceeds 2.0℃, the CRF anomalies decrease with the SSTA. This regime behavior is due to the regime behavior of cirrostratus and deep convective clouds. Second, the CRFs-SSTA relationship is regulated by remote forcings. Warming of the far eastern equatorial Pacific would reduce the water vapor convergence over the central Pacific by weakening the trade wind over the southeastern Pacific, thereby reducing the feeding of moisture to the convective flow. The results suggest that CRFs-SSTA relationships during ENSO events are nonlinear and strongly depend on the magnitude and the spatial distribution of the SSTA.
The vertical structures of atmospheric temperature anomalies associated with El Nio are simulated with a spectrum atmospheric general circulation model developed by LASG/IAP （SAMIL）. Sensitivity of the model’s response to convection scheme is discussed. Two convection schemes, i.e., the revised Zhang and Macfarlane （RZM） and Tiedtke （TDK） convection schemes, are employed in two sets of AMIP-type （Atmospheric Model Intercomparison Project） SAMIL simulations, respectively. Despite some deficiencies in the upper troposphere, the canonical El Nio-related temperature anomalies characterized by a prevailing warming throughout the tropical troposphere are well reproduced in both simulations. The performance of the model in reproducing temperature anomalies in ＂atypical＂ El Nio events is sensitive to the convection scheme. When employing the RZM scheme, the warming center over the central-eastern tropical Pacific and the strong cooling in the western tropical Pacific at sea surface level are underestimated. The quadru-pole temperature anomalies in the middle and upper troposphere are also obscured. The result of employing the TDK scheme resembles the reanalysis and hence shows a better performance. The simulated largescale circulations associated with atypical El Nio events are also sensitive to the convection schemes. When employing the RZM scheme, SAMIL failed in capturing the classical Southern Oscillation pattern. In accordance with the unrealistic anomalous Walker circulation and the upper tropospheric zonal wind changes, the deficiencies of the precipitation simulation are also evident. These results demonstrate the importance of convection schemes in simulating the vertical structure of atmospheric temperature anomalies associated with El Nio and should serve as a useful reference for future improvement of SAMIL.
The spectral version 1.1 of the Flexible Global Ocean–atmosphere–land System （FGOALS1.1-s） model was developed in the State Key Laboratory of Numerical Modeling for Atmospheric Sciences and Geophys- ical Fluid Dynamics at the Institute of Atmospheric Physics （LASG/IAP）. This paper reports the major modifications to the physical parameterization package in its atmospheric component, including the radiation scheme, convection scheme, and cloud scheme. Furthermore, the simulation of the East Asian Summer Monsoon （EASM） by FGOALS1.1-s is examined, both in terms of climatological mean state and interannual variability. The results indicate that FGOALS1.1-s exhibits significant improvements in the simulation of the balance of energy at the top of the atmosphere： the net radiative energy flux at the top was 0.003 W m-2 in the 40 years fully coupled integration. The distribution of simulated sea surface temperature was also quite reasonable, without obvious climate drift. FGOALS1.1-s is also capable of capturing the major features of the climatological mean state of the EASM： major rainfall maximum centers, the annual cycle of precipitation, and the lower-level monsoon circulation flow were highly consistent with observations in the EASM region. Regarding interannual variability, simulation of the EASM leading patterns and their relationship with sea surface temperature was examined. The results show that FGOALS1.1-s can reproduce the first leading pattern of the EASM and its close relationship with the decaying phase of the ENSO. However, the model lacked the ability to capture either the second major mode of the EASM or its relationship with the developing phase of the ENSO.
