Visualization and Validation of NWP Cloud and other Fields with Simulated Weather Imagery

Weather and All-Sky Imagery - A Comparison of Observed and Local Analysis and Prediction System (LAPS) Simulated Clouds A Visualization Procedure (VP) has been developed at the ESRL's Global Systems Division to construct simulated imagery of the atmosphere. The VP uses ray-tracing and other tools to efficiently simulate various aspects of radiative transfer using analyzed or forecast model prognostic variables in clear (moisture) or polluted air (aerosols), clouds (cloud liquid and ice), and precipitation (rain, snow, and ice hydrometeors). From arbitrary vantage points and under day or night-time lighting conditions the VP renders polar or cylindrical projection images of the atmosphere and the underlying terrain. VP weather imagery can be compared with camera images captured with fish-eye (all-sky) or wide-angle lenses to assess the fidelity and improve the performance of Numerical Weather Prediction (NWP) analysis and forecast systems. The VP is tested with an NWP system that produces very rapid update (15 mins or more frequent) and very high resolution (500 m or finer) analyses and forecasts (Local Analysis and Prediction System - LAPS). The cloud analysis of LAPS uses satellite (including IR and 1-km resolution visible imagery, updated every 15-min), METAR, radar, and aircraft observations along with a first guess forecast to produce 3-D fields of cloud and hydrometeor variables. The largely sequential data insertion procedure of today's LAPS is being updated with a 3/4DVAR cloud analysis module that in the future will be used both in LAPS and other fine scale data assimilation systems. Comparing analyses with day-time and night-time camera images under cloudy, precipitating, and clear/polluted air conditions, LAPS simulated VP imagery can realistically reproduce rainbows, twilight sky colors and other atmospheric phenomena. Since camera images are not yet used as observational input in LAPS, subjective and quantitative comparisons of high resolution observed and simulated weather imagery provide a highly valuable opportunity or “acid test” to assess the quality of NWP cloud analyses. VP weather imagery can also readily convey NWP information about current and forecast weather in an easily perceivable, visual form to both scientific and lay audiences. To fill observational gaps at the small-scales, using the VP with variational assimilation techniques, observed camera images will also be assimilated in a 3/4DVAR cloud analysis module, useful in systems such as LAPS, GSI, and various global models.