Real-Time, High-Resolution, Three-dimensional Cloud and Wind Data Assimilation Technology for the Battlespace Environment

Introduction: To meet the U.S. Navy's increasing needs for accurate and detailed descriptions and predictions of the 3-D battlespace atmospheric environment, a high-resolution data assimilation system using remotely sensed data is under development at the Marine Meteorological Division of the Naval Research Laboratory. This system fuses remotely sensed data, particularly a combination of radar and satellite data, which possess information about the 3-D dynamical and hydrological structures of the atmosphere with high spatial and temporal resolutions, especially over oceans or in data-denied areas where conventional meteorological information is limited.

Data Assimilation System: The new data assimilation system uses a variational approach, together with advanced data fusion techniques, to retrieve 3-D cloud and wind fields from radar observations, geostationary satellite data, and surface observations in real-time. Also, the data assimilation system has the capability to use data from multiple radars to enlarge data coverage and to improve data resolution and accuracy.

The cloud analysis algorithm uses data fusion technologies to merge brightness temperatures and cloud albedo from geostationary satellite data, radar reflectivity from Doppler radar observations, and cloud observations from surface reports to retrieve 3-D cloud information.1 The Navy's Coupled Ocean/Atmosphere Mesoscale Prediction System (COAMPSTM )* model forecasts are used as background fields for cloud analyses. Figure 1 gives an example of 3-D clouds produced by this system from a severe storm case on May 10, 2003, when a squall line developed and moved from north to south across the Washington, DC, and Norfolk, Virginia, areas. In this case, data from three radars in the storm area provided very fine structures of the storms that were missing from the background fields, especially at lower levels; satellite data and surface observations gave good estimations of cloud top and base heights.

FIGURE 1
3-D clouds (with water mixing ratio isosurface of 0.15 g/kg, blue color represents clouds below 700-hPa pressure level) of a severe storm system on the east coast on May 10, 2003, retrieved from radar, satellite and surface observations, and horizontal winds at 500-m level above the ground.

A variational approach is used to retrieve the 3-D winds from radar observations of radial velocities.2 By minimizing a simplified cost function, this method forces the analysis fields toward radar observations and the first guess fields from model forecasts, and at the same time, applies the radial momentum and mass continuity equations as constraints to ensure that the retrieved wind fields satisfy the dynamic laws that govern atmospheric flow. After wind retrieval, temperature and pressure perturbations are also computed to match the retrieved winds to keep the dynamic balance between the wind and mass fields. Extensive testing of the radar wind retrieval system has been accomplished. Figure 2 gives an example of retrieved horizontal winds from an isolated storm in North Carolina on July 24, 2002. In this case, data from the Doppler radar in Morehead City of North Carolina were used in the analysis. Strong lower level convergence near the storm center that was not apparent in the background fields can be seen clearly in Fig. 2.

FIGURE 2
Radar observed reflectivity (shaded area, dBZ) of an isolated storm on July 24, 2002, and retrieved horizontal winds (with maximum vector representing 17 m/s) at 750-m level above the ground.

Products and Applications: Several environmental parameters critical to Navy warfighters and their missions can be directly derived from the 3-D clouds and winds produced by the data assimilation system. As an example, cloud ceiling derived from the 3-D clouds in Fig. 1 is given in Fig. 3. In this figure, cloud base height less than 1000 m can be seen in several locations with the lowest cloud base height of less than 500 m. Verifications against surface observations show notable improvements in the accuracy of the derived cloud base heights over those from the background fields, especially in convective regions. Experiments have also been conducted to assimilate the retrieved 3-D clouds and winds into the high-resolution COAMPS model, and the results have showed some improvement in short-term theater-scale weather prediction. Verification of the radar radial velocity derived from the first-hour forecasts of 3-D winds with radar data assimilation against radar observations for the storm case in Fig. 1 shows an increase of 0.182 in correlation and a decrease of 1.59 (m/s) in RMS error over the control forecast. The high-resolution winds from both the data assimilation system and the COAMPS model forecast are input into chemical/biological dispersion models, which are used for assessing contamination avoidance and decontamination strategies. The technology was demonstrated during Fleet Battle Experiment — Juliet with products providing up-to-date, detailed information to tactical decision makers about the 3-D atmospheric battlespace conditions. This work, though focused on battlespace environmental applications, establishes a scientific framework for using radar-derived meteorological information in multiple nowcasting and numerical weather prediction applications.

FIGURE 3
Cloud ceiling derived from the 3-D clouds in Fig. 1 with red color indicating regions with cloud base heights less than 500 m.

Acknowledgments: The authors thank Dr. Qin Xu of the National Severe Storm Laboratory for his assistance in the system development; and Mike Frost of CSC and Larry Phegley of NRL for their help in acquiring real-time radar and satellite data.

[Sponsored by NRL and ONR]