Comparative simulation between with bulk and with bin type cloud microphysical scheme on non-hydrostatic regional model

There are two methods that are generally used in high-resolutional model to represent cloud microphysics. One is a bulk (moment) method that predicts variables such as cloud mixing ratio and number concentration represented by integrated values of a prescribed size-distribution function. And another is a bin (spectral) method that size-distribution functions of cloud hydrometeors are discrete-approximated by a number of size bins and are predicted by a cloud microphysics scheme. As for an influence of aerosols on clouds, significant factors are changes of cloud droplets number consideration, mean radius, and total volume, which cause a change of the radiation property and residence time of clouds. Thus, the bin method, which is able to estimate those values directly, is more adapted than a bulk method to analyze the aerosol indirect effects.

A numerical model for atmospheric dynamics used in this study is based on a multi-purpose non-hydrostatic atmospheric model developed by the Forecast Research Department of the Meteorological Research Institute and the Numerical Prediction Division of the Japan Meteorological Agency (MRI/NPD-NHM) [Saito and Kato, 1999; Saito et al., 2001]. This dynamical frame adopts a bulk-type cloud microphysical scheme [Ikawa and Saito, 1991]. Hydrometeors tracers are mixing ratios of cloud water, rain, cloud ice, snow and graupel in this bulk-type scheme. Then, we replaced it with a bin-type cloud microphysical scheme based on the Hebrew University Cloud Model [Khain and Sednev, 1995; Khain and Sednev, 1996; Khain et al., 2000]. Model tracers are size distributions of cloud condensation nuclei (CCN) and hydrometeors (water droplets, ice plate crystals, ice dendrite crystals, ice column crystals, snow flakes, graupels and hails).

In this study, we practice nest-grid simulations with bulk-type and with bin-type cloud microphysical scheme and compare these results. Numerical simulations are carried out in a region around the East China Sea within a radius of 1,400 kilometers whose center is the sea near the Kyushu region. Calculations are made from 18:00 to 24:00 of April 7th, 2003. Japan Meteorological Agency meso-analysis dataset, with horizontal grid of 10 km, vertical 20 layers, and time step of 6 hours interval, is used for initialization and nesting of dynamical variables, i.e., horizontal velocities, temperature and relative humidity. CCN data for initialization and nesting in this study are prepared from the results of a numerical model, SPRINTARS aerosol transport and radiation model (Takemura et al., 2000; Takemura et al., 2001) coupled with CCSR-NIES/AGCM (Numaguti, 1995) with a horizontal resolution of T106 and 20 vertical layers. The horizontal grid size of the model is set as 7 km (for 202 grid points) and the atmosphere up to 12 km is divided by 38 vertical layers with intervals increasing with altitude (40 m for the bottom layer to 580 m for the top layer). The time step is taken as 20 seconds.

We will indicate some simulated cloud parameters such as precipitation amount, liquid water path and ice water path. Then, component ratios of categorized hydrometeors both on bulk and on bin method will be shown also. As an example of results, simulations in the targeted case show that the precipitation amount with bin method is much larger than that with bulk method. Now we are developing a detailed radiation scheme for the bin method scheme and plan to add an analysis about radiation.