Impact of microphysical assumptions on the intensity and the structure of simulated hurricanes: Can satellite observations help determine the optimal set of microphysical assumptions?

Improving forecasting of hurricane intensity and track remains a significant challenge for the research and operational communities and is the goal of the Hurricane Forecasting Improvement Project (HFIP) recently launched by NOAA in collaboration with NASA.

Recent studies indicate that the hurricane inner core dynamics, dominated by convective and microphysical processes, and potential oceanic feedback, might play a crucial role in determining the storm's intensity, structure and size. To analyze and accurately predict the Tropical Cyclone (TC) evolution requires properly reflecting these small scale complex interactions by using: i) high-resolution models; ii) realistic physical parameterizations.

Many factors determine a tropical cyclone's intensity. Ultimately, though, intensity is dependent on the magnitude and distribution of the latent heating that accompanies the hydrometeor production during the convective process. Hence, the microphysical processes and their representation in hurricane models are of crucial importance for accurately simulating hurricane intensity and evolution. The accurate modeling of the microphysical processes becomes increasingly important when running high-resolution models that should properly reflect the convective processes in the hurricane eyewall.

We study the impact of microphysical assumptions on the structure and the intensity of the simulated hurricanes. In particular we compare and contrast the members of high-resolution ensemble WRF model simulations of Hurricane Rita (2005). The members of the ensemble include simulations with three different microphysical schemes and eight different Particle Size Distribution (PSD) assumptions within one of the microphysical schemes.

Today there is still significant uncertainty in the PSDs of the precipitating systems – they vary geographically, as a function of the type of precipitation (convective versus stratiform) and also as a function of the aerosol loading of the atmosphere. In turn, the PSD assumptions affect the hydrometeor growth, phase changes and fallout, resulting in modulation of the precipitation efficiency of the simulated storm and their ice versus liquid partitioning.

In this presentation we will compare and contrast the thermodynamic and hydrometeor structure of the different simulations in an attempt to understand how the microphysical assumptions affect the storm intensity, vertical structure, size and track.

Furthermore, the PSD assumptions have a very strong impact on the simulated remotely sensed characteristics (radar reflectivity and microwave brightness temperatures). This sensitivity impacts negatively the accuracy of satellite retrievals of precipitation. At the same time, the strong PSD sensitivity provides an opportunity to use comparisons between observed and forward simulated radiometric signatures of precipitation to determine what PSD assumptions result in most realistic simulated hurricanes. Such an approach provides a promising alternative to the more common model evaluation based on comparison of modeled and satellite-retrieved geophysical quantity (e.g. rain rate or near surface wind speed).

To help determine the best set of assumptions, we use the geophysical WRF model fields as input to instrument simulators to produce microwave brightness temperatures and radar reflectivity at the TRMM (TMI and PR) frequencies and polarizations. We also simulate the surface backscattering cross-section at the QuikSCAT frequency, polarizations and viewing geometry. We use satellite observations from TRMM and QuikSCAT to determine those parameterizations that yield a realistic forecast and those parameterizations that do not.

To facilitate hurricane research, we have developed the JPL Tropical Cyclone Information System (TCIS) - the satellite and data analysis component of the NASA's TC-IDEAS (Tropical Cyclone – Data Exchange and Analysis system) that is being developed jointly by teams at the Marshall Space Flight Center (MSFC) and the Jet propulsion Laboratory (JPL). TCIS includes a comprehensive set of multi-sensor observations relevant to the large-scale and storm-scale processes in the atmosphere and the ocean. In this presentation, we will illustrate how the TCIS can be used for hurricane research.

Our results indicate that multiparameter satellite observations can help discriminate between simulations with different microphysical assumptions. In particular, assuming hydrometeor distributions with larger number of smaller particles results in model simulations with radiometric signatures that compare more closely to observations. Such results will have impact on hurricane forecasting in two ways: i) providing guidance as to the optimal set of physical parameterizations to be used; ii) improving the data assimilation outcome by designing model forecasts whose radiometric signatures are close to the observed ones, thus increasing the relative importance of the observations during the assimilation.

Furthermore, the improved understanding of what the PSD characteristics are will lead to decrease in the uncertainty of satellite retrievals of precipitation which very often use model-derived retrieval databases that reflect the microphysical assumptions used by the models.

The work described here was performed at the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration.