Project title: Feature-based Validation of mirs soundings for Tropical Cyclone Analysis and Forecasting

PROJECT OBJECTIVES:

The purpose of this work is to determine the utility of the Microwave Integrated Retrieval System (MIRS) retrievals for a specific region (the Tropics) and for a specific application (tropical cyclone analysis) to complement the broader Cal/Val activities at STAR. For the 2012 Atlantic hurricane season, the MIRS soundings generated from both the AMSU (Advanced Microwave Sounding Unit) instrument aboard the NOAA and MetOp satellites, and from the ATMS (Advanced Technology Microwave Sounder) instrument aboard Suomi-NPP will be analyzed. The validation of the MIRS soundings will consist of standard error analysis of the thermodynamic profile in addition to comparing their performance in four products relevant to tropical cyclone analysis and forecasting: Maximum Potential Intensity (MPI), vertical velocity profile from an entraining plume model, the Multiplatform Satellite Surface Wind Analysis, and a statistical intensity and wind radii estimate. Comparisons will be made not only between retrievals from the AMSU and the ATMS instruments, but also between the AMSU retrievals generated from MIRS and AMSU retrievals generated from an older, statistical algorithm. Dropsondes from tropical cyclone reconnaissance flights will provide the “ground truth” measurements.
This project supports the following NOAA mission goals: Weather and Water.

PROJECT ACCOMPLISHMENTS: Past Fiscal Year by Objective:

Code, including Fortran programs, Linux shell scripts, Python scripts, and IDL procedures have been written to perform the collocation among three data sources: dropsondes, the statistical AMSU retrievals, and the MIRS AMSU retrievals. Data for the 2012 Atlantic hurricane season has been collected, and the collection of the data for the 2013 season is nearing completion. Figure 1 shows the temperature statistics among the dropsondes and the two AMSU retrieval techniques for hurricanes Ernesto and Isaac. For these collocations, the satellite retrieval was within 50 km and 1 hour of the dropsonde. Both the bias (dashed) and root mean square error (solid) show that the MIRS retrievals are generally more accurate in the lower and middle levels, but that the statistical technique performs better at the upper levels. A temperature correction which would produce more accurate retrievals from the statistical technique at lower levels, however, was not applied in this instance. Additionally, the number of matchups at each level, given on the right-hand side of the figure is rather small. More reliable statistics will be produced as more cases are added to the analysis.

Figure 1. Matchup statistics of AMSU temperature retrievals using a statistical technique (red) and the MIRS algorithm (green) for hurricanes Ernesto and Isaac of the 2012 Atlantic season. The bias is shown in dashed lines and the root mean square error in solid lines. The number of collocations at each level is given on the right.

PRESENTATIONS / PUBLICATIONS