Tropical Cyclone Product Development

In an effort to improve the tropical cyclone formation probability guidance product, large-scale vertical motion will be added as an additional screening parameter, and its effect on the skill of the forecast measured.  The large scale vertical motion field will come from a Q-vector form of the omega equation valid over the entire sphere, which will use the GFS model fields as input.  The omega equation is solved using a vertical normal mode and a spherical harmonic transform.  A test forcing was created such that the computed omega could be compared to the analytical omega.  The results show good agreement.  (J. Dostalek)

Omega computed from test forcing.  The x axis is latitude in degrees and the y axis is pressure in hPa.

Analytical omega corresponding to test forcing.  The x axis is latitude in degrees and the y axis is pressure in hPa.

A proposal entitled “Global Tropical Cyclone Formation Probability Product” was accepted for FY2010.  The research will expand the extent of the product from the Atlantic and North Pacific basins to also include the southern hemisphere and Indian Ocean basins.  (J. Dostalek, A. Schumacher)

The NPROVS (NOAA Product Validation System) graphical software was installed at the RAMM Branch.  The software will be used to assess the quality of tropical temperature and moisture retrievals from polar-orbiting satellites.  The assessment will be based not only on overall performance with respect to radiosondes and dropsondes, but also on the ability to capture atmospheric characteristics such as low-level moisture, trade-wind inversions, stability indices, and tropopause level.  The included figure shows the ProfileDisplay interface, part of the NPROVS graphical software.  In addition, a proposal (“Validation of Satellite-Based Thermodynamic Retrievals in the Tropics”) for the continuation of this work for an additional year was accepted. (J. Dostalek)

Interface for ProfileDisplay, part of the NPROVS graphical software.

The Dvorak Technique for estimating tropical cyclone intensity has been validated versus aircraft-influenced best track intensities.  This method continues to be an important operational tool for making such estimates.  Finding suggest that biases are a function of intensity, intensity trend, translation speed, latitude and size (pressure of outer closed isobar), as shown in the figure below, whereas errors are primarily a function of intensity.  The goal of such research is to help identify the causes of existing biases and possibly reducing the errors associated with operation tropical cyclone intensity estimates.  A journal article is being prepared.  (J. Knaff)

Figure 1.  Average biases associated with the Dvorak technique coming from two agencies SAB (left) and TAFB (right).  The comparisons are homogeneous.  The composites were constructed using intensity trend (first row), latitude (second row), translation speed (third row) and ROCI (fourth row).  Stratifications are provided below.  

The Multi-platform Tropical Cyclone – Surface Wind Analysis (MTC-SWA) was successfully ported to computers systems at the National Satellite Operations Facility.   The data ingest and software is now running and producing all the products and graphics in real-time.  Software locations have been supplied for review, and paper work requesting archival of products being prepared. (J. Knaff)

An updated version of the Wind Radii CLIPER Model was developed for the Joint Typhoon Warning Center (JTWC).  The modification allows for the production of wind radii forecasts in the North Indian Ocean and the entire Southern Hemisphere.   Climatological coefficients were adapted from other basins to accomplish this task.  The request was made by JTWC because this model is the only tropical cyclone wind radii guidance available at JTWC and such capabilities did not exist in these basins. (J. Knaff)

The track and intensity forecast error distributions utilized by NHC’s operational wind probability model were updated to include data from the 2008 season and remove data from the 2003 season. The model uses the previous 5 years of errors in a Monte Carlo algorithm to estimate the probabilities of 34, 50 and 64 kt wind out to 5 days, and a number of operational products are derived from the output.  A version is also run by the Joint Typhoon Warning Center (JTWC). These new distributions were provided to NHC and JTWC for use during the 2009 hurricane and typhoon seasons. (J. Knaff)

A simple GOES water vapor advection scheme was developed for potential use in a project to extend the Tropical Cyclone Formation Probability (TCFP) product time range from 24 to 48 hours.  Verification shows that the advected water vapor brightness temperatures (BT) match the observed BT better than persistence (the method the current TCFP uses) out to 42 hour over a one month sample of GOES-E & simulated ABI data, however the R2 values drop off significantly after 6-12 hours (Figure 1).  This suggests that the extended TCFP will have to rely more heavily on GFS forecast-based parameters and have less reliance on satellite-based convective parameters than the current product.  Work continues on development of the longer-range TCFP product.  (A. Schumacher, M. DeMaria)

Figure 1.  Verification of a simple GOES-E water vapor brightness temperature advection scheme and persistance with simulated ABI water vapor brightness temperatures at forecast times 0, 6, …, 48 h and from 8 S to 51 N at 37.5 W (longitude at midpoint between GOES-E and MSG).

Data collection needed to extend the TCFP product to the Indian Ocean and S. Hemisphere, hence making it global tropics product, is complete.  Both northern and southern hemisphere geostationary satellite water vapor imagery from Meteosat-5&7 and Meteosat-9, southern hemisphere scans from GOES-W and MTSAT-1R, and Indian Ocean and S. Hemisphere best tracks have been obtained and algorithm development is currently underway. (A. Schumacher, M. DeMaria, J. Knaff)

Return to the current RAMMB CIRA Quarterly Report