Clark-Hall Modeling Studies

The three-dimensional, non-hydrostatic anelastic meteorological model described by Clark (1977), Clark and Hall (1991), and Clark et al. (1996), exploits features such as two-way interactive grid nesting and vertically-stretched terrain-following coordinates. The model uses a bulk parameterization for both the liquid and ice phase. The liquid phase is parameterized according to a modified version of the Kessler (1969) scheme, with Simpson and Wiggert's (1969) autoconversion formula. In this scheme, liquid water exists as cloudwater and rainwater. The ice phase parameterization uses the Koenig and Murray (1976) ice microphysical scheme. The Koenig-Murray formulation treats two types of ice particles: pristine ice (Ice A) - ice crystals initially formed by heterogeneous nucleation or ice splinter processes due to riming, and ice particles (Ice B) - also called graupel and initially formed by the freezing of raindrops or the interaction of Ice A particles with raindrops. This ice scheme is described, in detail, by Bruintjes et al. (1994). In summary, the model carries microphysical variables of water vapor, cloud water mixing ratio, rain water mixing ratio, and number concentration and mixing ratio for two types of ice particles.

Case I: 10 January 2001 (Winter Case)

Meteosat Satellite Image from 9 Jan 2001
at 0655 UTC

Meteorological Conditions

A larger frontal system passed through the UAE region over a two-day period (January 8-10). A satellite photograph of the cloud system on January 9 during the seeding experiment is shown in the figure to the left. At the time of the photograph, convective clouds are evident over the Gulf, the northern parts of the UAE and over the mountains of eastern UAE and into Oman. The main part of the system is located over Iran. The UAE is usually on the southern fringes of these systems as shown in this case

This case was characteristic of the synoptic systems that pass through the region during the winter, the frequency varying from year to year, and sometimes bringing rain to the UAE. In the region ahead of the synoptic disturbance, scattered low level cumulus clouds were observed. Warm, moist air rose and moved to the northeast as it was drawn into the cyclonic system approaching the UAE from the northwest late on 9 January. The system itself arrived in a sequence of wave disturbances. These disturbances appeared as a sequence of bands of convective elements within a larger ascending cloudy region.


The results of this case study are currently unavailable. They will be posted as soon as possible.

Animations of Results

9 January 2001 MM5 Forecast

Case II: 31 July 2001 (Summer Case)

Meteosat Satellite Image from 31 Jul
2001 at 1000 UTC

Meteorological Conditions

The case notes recorded on 31 July 2001 indicate there was widespread convection capped by an inversion with strong rain showers. Early storms developed over and west of mountains, but later persistent cells (possibly tied to the sea breeze front) produced rain along the coast (Jebel Ali) and and eventually southwest of Al Dhafra. The research aircraft sampled these storms. Larger storms formed even later (sunset and beyond) that pelted Abu Dhabi with wind and rain, and continued to form and propagate northeastward. Flow interactions appeared to be the forcing for the storms. Here we describe research simulations designed to study and understand the mechanisms for precipitation formation in this situation.

Results - Sounding Initialization

Two distinct flows were produced. First, easterly monsoon flow leaks through canyons and lower mountain ridges, providing moisture for all the day's convection. In the morning, the Oman Mountains provide orographic lift to the easterly monsoon flow, which initiated shallow clouds onshore and convective clouds along the peaks. Toward midday, the mature, but shallow, convective clouds produced some rain that evaporated, cooling the air to make it negatively buoyant and causing downdrafts that led to weak surface gusts. Also, the heating of the land relative to the water led to a sea breeze onset at 9 am LST that pushed inland. These flows (easterly gap flow and northwesterly sea breeze) collided leading to additional areas of convergence and lift producing even stronger and deeper convective updrafts along the interface of the flows over the northern coastal cities where rainwater formed and fell from the clouds.

In this simulation the sea breeze pushed inward, moving the areas of convection inland. This seems to exaggerate the strength of the modeled sea breeze because radar observations show these clouds formed by colliding boundaries remained closer to the coast. The exact location of the interface of the colliding boundaries depends on the timing and relative strength of the sea breeze and convective downdrafts, which in turn depends on the land-gulf temperature difference, the convective potential of the orographic cloud (i.e., whether moisture is available and the degree of instability in the atmospheric profile), and the amount of rain evaporated in their downdrafts (which depends in part on the drop size distribution and relative humidity and depth of the sub-cloud layer).

Results - MM5 Initialization

The Clark-Hall model simulation was begun at 6 am on the 31st (2 Z) carried out with three nested domains (10 km, 3.3 km, and 1.1 km horizontal grid spacing) for 13 hours (until 15 Z). The simulation showed that four distinct flows affecting the UAE must be considered: 1) moist, easterly flow off the Gulf of Oman, culminating in a strong easterly gap flow through the low-lying ridges at the northern end of the Oman Mountains, 2) a building northwesterly sea breeze from the Arabian Gulf, moist throughout a very shallow stable later 3) a weak, dry south - southeasterly land breeze, which becomes disorganized as solar heating warms the UAE interior, and 4), cool, evaporatively driven downdrafts from convective clouds over the ridges of the Oman Mountains, spreading westward.

Animations of Results

31 Jul 2001 MM5 forecast

3-D Visualization [rotating] of Cloud Water (yellow) andRain (red) Mixing Ratios from 0642 to 1442 UTC {sounding initialization}

3-D Visualization [non-rotating] of Cloud Water (yellow) and Rain (red) Mixing Ratios from 0642 to 1442 UTC {sounding initialization}

Horizontal Visualization of Wind Vectors and Cloud Water (orange) Mixing Ratios from 0642 to 1442 UTC {sounding initialization}

Horizontal Visualization of Wind Vectors, Water Vapor (white) and Cloud Water (red) Mixing Ratios from 0600 to 1800 UTC {MM5 initialization}

New Cloud Microphysical Parameterization Scheme

As part of the entire research project, a new microphysical parameterization scheme has been developed that will allow us to test the impact of aerosols (including those either pollutant or seeding agent in nature). The basic requirement of the scheme is that it should be of sufficient detail and breadth to allow model simulations that reproduce the basic character and evolution of naturally occurring clouds and cloud systems in the project area while treating aerosols in enough detail that hygroscopic seeding can be realistically simulated.

We determined that the scheme should include the following features:

  • Nucleation of both natural and artificial aerosols are treated directly, with explicit prediction of supersaturation.
  • Treatment of the warm rain process are of sufficient detail that the development of drizzle and rain is a function of the cloud droplet distributions produced from the activated aerosols.
  • Six hydrometeor classes are used; including cloud water and rain to define the liquid water spectra while the ice particle spectra is divided into four classes - ice crystals, snow, graupel and hail.
  • Two moments of the size distribution, number concentration and mixing ratio, are predicted for each hydrometeor class, with the particle size distributions given by gamma distributions.

The new microphysical parameterization scheme developed for the model has been assembled, coded, and undergone initial testing over the past year. It includes all of the desired features listed above and is described in detail below. Each hydrometeor class interacts with water vapor and with the other hydrometeor classes through a series of idealized representations or parameterizations of the physical processes. These processes are the nucleation of cloud condensation nuclei, condensation/evaporation, collision/coalescence, including drop breakup, both homogeneous and heterogeneous nucleation of ice crystals, deposition/sublimation, collision/aggregation, accretion, freezing, melting and shedding, and ice multiplication via the rime splintering mechanism.

The treatment of the nucleation of CCN is based on Cohard et al. (1998). Warm rain processes are based on the treatment of Cohard and Pinty (2000), who have extended the work of Ziegler (1985) which, in turn, was based on Berry and Reinhardt (1974) and Long (1974). The treatment of ice is similar to the double-moment four-class ice scheme developed by Brad Ferrier of NASA (Ferrier, 1994), although our new formulation does not allow for mixed-phase particles (i.e. ice particles with water coatings)

Read about MM5 real-time forecasts