IMPACT OF CUMULUS INITIALIZATION ON THE SPINUP OF PRECIPITATION FORECASTS IN THE TROPICS

In order to ameliorate the precipitation spinup problem (prediction models' inability to produce realistic precipitation rates at the beginning of the forecast period), the impact of a tropical initialization procedure on precipitation forecasts has been investigated. The procedure is divided into the three components: 1) application of diabatic nonlinear normal-mode initialization (NNMI), 2) modification of the initial divergence by incorporation of satellite-imagery data, and 3) modification of the moisture and temperature fields by Donner's cumulus-initialization scheme. Numerical experiments were conducted by running 10.5-h forecasts (42 time steps) starting from various initial conditions after application of some combination of the three initialization components. A triangular-42 version of the National Center for Atmospheric Research (NCAR) Community Climate Model-1 global spectral model (CCM1) and its associated NNMI package were used. The results of a case study from the reanalyzed First GARP (Global Atmospheric Research Program) Global Experiment (FGGE) level III data show that 1) even if a good estimate of diabatic heating rates were available, diabatic NNMI alone would not solve the spinup problem, 2) the adjustments of moisture and temperature using the cumulus initialization are essential to ameliorate the spinup problem, and 3) the divergence adjustment, assisted by satellite-imagery data, is beneficial when used in conjunction with the cumulus initialization and diabatic NNMI procedures.