A two-dimensional cloud-resolving model is used to study the sensitivities of two microphysical schemes, a bulk scheme and an explicit spectral bin scheme, in simulating a midlatitude summertime squall line [Preliminary Regional Experiment for Storm-Scale Operational and Research Meteorology (PRE-STORM), 10-11 June 1985]. In this first part of a two-part paper, the developing and mature stages of simulated storms are compared in detail. Some variables observed during the field campaign are also presented for validation. It is found that both schemes agree well with each other, and also with published observations and retrievals, in terms of storm structures and evolution, average storm flow patterns, pressure and temperature perturbations, and total heating profiles. The bin scheme is able to produce a much more extensive and homogeneous stratiform region, which compares better with observations.However, instantaneous fields and high temporal resolution analyses show distinct characteristics in the two simulations. During the mature stage, the bulk simulation produces a multicell storm with convective cells embedded in its stratiform region. Its leading convection also shows a distinct life cycle (strong evolution). In contrast, the bin simulation produces a unicell storm with little temporal variation in its leading cell regeneration (weak evolution). More detailed, high-resolution observations are needed to validate and, perhaps, generalize these model results. Interactions between the cloud microphysics and storm dynamics that produce the sensitivities described here are discussed in detail in Part II of this paper.

The response of cloud systems to their environment is an important link in a chain of processes responsible for monsoons, frontal depression, El Nino-Southern Oscillation (ENSO) episodes and other climate variations (e.g., 30-60 day intra-seasonal oscillations). Numerical models of cloud properties provide essential insights into the interactions of clouds with each other, with their surroundings, and with land and ocean surfaces. Significant advances are currently being made in the modeling of rainfall and rain-related cloud processes, ranging in scales from the very small up to the simulation of an extensive population of raining cumulus clouds in a tropical- or midlatitude-storm environment.The Goddard Cumulus Ensemble (GCE) model is a multidimensional non-hydrostatic dynamic/microphysical cloud resolving model. It has been used to simulate many different mesoscale convective systems that occurred in various geographic locations. In this paper, recent GCE model improvements (microphysics, radiation and surface processes) will be described as well as their impact on the development of precipitation events from various geographic locations. The performance of these new physical processes will be examined by comparing the model results with observations.Specifically, the impact of different ice schemes (i.e., three-class ice scheme, four-class two-moment ice scheme) on precipitation processes are examined and compared. Spectral bin microphysics are used to investigate precipitation processes under clean and dirty environments. The coupled GCE-radiation model shows that the modulation of relative humidity by radiative processes is the main reason for the diurnal variation of precipitation in the tropics. The coupled GCE-land surface model is used to examine the impact of heterogeneous land surface characteristics (soil-vegetation) on precipitation processes. The effect of ocean flux algorithms (e.g., the TOGA COARE flux algorithm and a simple bulk aerodynamic method) on surface fluxes, environmental convective available potential energy (CAPE) and precipitation processes are compared. In addition, the coupled GCE-ocean mixed layer (OML) model is used to investigate the physical processes that affect the variation of sea surface temperature, mixed layer depth and salinity.

A basic characteristic of cloud-resolving models (CRMs) is that their governing equations are nonhydrostatic since the vertical and horizontal scales of convection are similar. Such models are also necessary in order to allow gravity waves, such as those triggered by clouds, to be resolved explicitly. CRMs use sophisticated and physically realistic parameterizations of cloud microphysical processes with very fine spatial and temporal resolution. Another major characteristic of CRMs is their explicit interaction between clouds and radiation. It is for this reason that Global Energy and Water Cycle Experiment (GEWEX) has formed the GEWEX Cloud System Study (GCSS) expressly for the purpose of studying these types of problems using CRMs. Observations can be used to verify model results and improve the initial and boundary conditions. The major advantages of using CRMs are their ability to quantify the effects of each physical process upon convective events by means of sensitivity tests (eliminating a specific process such as evaporative cooling, terrain, planetary boundary layer [PBL]), and their detailed dynamic and thermodynamic budget calculations.

The spectral representation of rain profiles observed by the Precipitation Radar (PR) of the Tropical Rainfall Measuring Mission (TRMM) provides convective and stratiform rain characteristics over the equatorial area (Takayabu, 2002). Motivated by this, we introduce a new retrieval algorithm, the spectral latent heating (SLH) algorithm, for PR. The primary difference from three algorithms in Tao et al. (2001) is that we utilize rain profiles observed by PR.

Coupling a cloud resolving model (CRM) with an ocean mixed layer (OML) model can provide a powerful tool for better understanding impacts of atmospheric precipitation on sea surface temperature (SST) and salinity (Li et al. 2000)• The objective of this study is twofold. First, by using the 3-D CRM-simulated (the Goddard Cumulus Ensemble model, GCE) diabatic source terms, radiation (Iongwave and shortwave), surface fluxes (sensible and latent heat, and wind stress), and precipitation as input for the OML model, the respective impact of individual component on upper ocean heat and salt budgets are investigated. Secondly, a two-way air-sea interaction between tropical atmospheric climates (involving atmospheric radiative-convective processes, Tao et al. 1999) and upper ocean boundary layer is also examined using a coupled 2-D GCE and OML model. Results presented here, however, only involve the first aspect. Complete results will be presented at the conference.

Latent heating profiles associated with three TOGA COARE active convective episodes (December 10-17 1992; December 19-27 1992; and February 9-13 1993) are examined using the two-dimensional version of the Goddard Cumulus Ensemble (GCE) Model, and retrieved by using the Goddard Convective and Stratiform Heating (CSH) algorithm. The following sources of rainfall information are input into the CSH algorithm: Special Sensor Microwave Imager (SSM/I), shipborne radars and the GCE model. Diagnostically determined latent heating profiles are calculated using 6 hourly soundings used for validation.The GCE model simulated rainfall and latent heating profiles are in excellent agreement with those estimated by soundings. In addition, the typical convective and stratiform heating structures (or shapes) are well captured by the GCE model. Radar measured rainfall is smaller than that estimated by the GCE model and SSM/I in both December convective episodes. SSM/I derived rainfall is more than the GCE model simulated for the December 19-27 and February 9-13 periods, but it is in excellent agreement with the GCE model for the December 10-17 period. The GCE model estimated stratiform amount is about 50 % for December 19-27, 42 % for December 11-17 and 56 % for the February 9-13 case. These results are consistent with large-scale analyses. Accurate estimates of stratiform amount are needed for good latent heating retrieval. A higher (lower) percentage of stratiform rain can imply a maximum heating rate at a higher (lower) altitude. The GCE model always simulates more stratiform rain (10 to 20 %) than the radar for all three convective episodes. The SSM/I derived stratiform amount is about 37 % for December 19-27, 48 % for December 11-17 and 41 % for the February 9-13 case.Temporal variability of CSH algorithm retrieved latent heating profiles using either the GCE model simulated or radar estimated rainfall and stratiform amount is in good agreement with that diagnostically determined for all three periods. However, less rainfall and a smaller stratiform percentage estimated by radar resulted in a weaker (underestimated) latent heating profile, and a lower maximum latent heating level compared to those determined diagnostically. Rainfall information from SSM/I can not retrieve individual convective events due to poor temporal sampling. Nevertheless, this study suggests that a good rainfall retrieval from SSM/I for a convective event can lead to a good latent heating retrieval.Sensitivity testing has been performed and the results indicate that the SSM/I derived time averaged stratiform amount may be underestimated for December 19-27. Time averaged heating profiles derived from SSM/I, however, agree well with those derived by soundings for the December 10-17 convective period. The heating retrievals may be more accurate for longer time scales, provided there is no bias in the sampling.An appropriate selection of latent heating profiles from the CSH look-up table is important. Sensitivity tests addressing this issue have been performed.

A prognostic cloud prediction scheme, designed for large-scale models, has been incorporated into a Single Column Model (SCM) and used to simulate the cloud cluster properties observed during the 19-26 December, 1992 Tropical Oceans Global Atmosphere (TOGA)-Coupled Ocean Atmosphere Response Experiment (COARE). Results from the SCM simulations have been compared with simulated profiles obtained from the Goddard Cumulus Ensemble Model (GCEM).Observed large-scale advective temperature, water vapor, and surface fluxes have been used as forcings to run the SCM and the GCEM. Results indicate that the SCM produces mixed profiles of cooling/warming and drying/moistening in the vertical, which are highly sensitive to the prescribed surface fluxes. Errors in the temperature and moisture profiles simulated by the SCM are about +/-3 K and +/-3 g kg(-1), while those from the GCEM are approximately -2 K and 1 g kg(-1) at most levels. The SCM produced -10 % errors in the relative humidity above 700 mb and -30 % at the surface, while in the GCEM, the errors were about 10 to 15 % at most levels. The distributions of precipitation rates are fairly well simulated, but the high cloud fractions are slightly underestimated by the SCM as compared to the GCEM. The cloud liquid water is underestimated, but the ice contents are slightly overestimated by the SCM. Results show that the SCM has been able to simulate the distributions of temperature, moisture, and precipitation rates fairly well as compared to the GCEM and other GEWEX Cloud System Study (GCSS) model intercomparison products.Sensitivity studies have been carried out to investigate the implications of different physical processes in the SCM. Results indicate that the interactions between various physical processes are nonlinear, and a mere substitution of the heating and moistening profiles from the GCEM may not be able to reproduce the observed temperature and moisture values by the SCM. The SCM has also been used to simulate the distributions of large-scale cloud cluster properties and their diurnal variation during the disturbed and suppressed periods of convection. Results show that the diurnal variations of simulated values are in agreement with many observational studies conducted by different authors over the TOGA-COARE.

ABSTRACT Deep convection in the tropical atmosphere,is simulated using the weak temperature gradient (WTG) approximation,applied to the Goddard Cumulus Ensemble Model (GCEM). The model is run in two spatial dimensions with periodic boundary conditions over uniform sea surface temperature (SST) with no mean vertical shear. 60-day experiments are run in which the SST is varied, with the horizontal mean free tropospheric temperature held close to the same profile, which is derived from a radiative-convective equilibrium (RCE) simulation. The large-scale vertical velocity is computed,as that required to balance the total heating. This vertical velocity is then used to advect moisture in the moisture equation. No large-scale horizontal moisture advection is used. The quasi-steady statistics of the last 30 days of each simulation are analyzed. As SST is increased, the mean precipitation rate increases nonlinearly, with something of an “S” shape with respect to SST. Surface latent heat flux does not increase with SST, implying that the precipitation variations are not controlled in any simple way by surface fluxes in these simulations. RCE occurs in the WTG model at an SST greater than that used in the RCE simulation (run in standard, not WTG mode) from which the horizontal mean temperature profile was derived. The time mean vertical profiles of relative humidity, moist static energy and large-scale vertical velocity give the impression of a bifurcation, with two distinct populations depending on whether the SST is above or below a threshold value, 29.5