SECTION 6. INTENSITY FORECASTS AND FORECAST PROCEDURES

Intensity forecast procedures vary between forecast agencies both in what parameters are emphasized and how the forecasts are developed. The following sections describes procedures used to forecast TC intensity.

JTWC provides analyses and forecasts of maximum one-minute sustained surface wind speed (i.e., the maximum intensity), and the 35, 50, and 100 kt wind radii out to 72 hours when applicable.

The procedure begins when the Dvorak technique (1984, see also Appendix B) is used to determine the current intensity of a TC and as a first guess for the intensity forecast (Guard, et. al., 1992). The Typhoon Duty Officer (TDO) then adjusts the forecast after evaluating climatology and the synoptic situation.

An interactive conditional climatology scheme allows the TDO to define a situation similar to the system being forecast in terms of location, time of year, current intensity, and intensity trend. Synoptic influences such as the location of major troughs and ridges, and the position and intensity of the TUTT all play a large part in intensifying or weakening a TC.

JTWC incorporates an experimental intensity analysis and forecast checklist (Appendix D). Such criteria as upper-level outflow patterns, neutral points, sea surface temperatures, enhanced monsoonal or cross-equatorial flow, and vertical wind shear are evaluated for their tendency to enhance or inhibit normal development. These criteria are incorporated into the tropical cyclone intensity forecast process through locally developed rules of thumb.

In addition to climatology and synoptic influences, the first guess is modified for interactions with land, with other TCs, and with extratropical features. Satellite cloud imageries help to assess the potential for fast development, rapid intensification, and time of peak intensity.

Climatological and statistical methods are also used to assess the potential for rapid intensification (Mundell, 1990). JTWC also has a statistical typhoon intensity forecast model, STIFOR, that provides intensity forecasts out to 72 hours for 12-hour intervals (Chu, 1994). The STIFOR model is an adaptation of the statistical hurricane intensity forecast (SHIFOR) model (Jarvinen and Neumann, 1979) which is used by NHC. The Typhoon Intensity Prediction Scheme (TIPS) has been developed for the western Pacific Ocean tropical cyclone intensity forecasts (Fitzpatrick, 1997). The predictors of TIPS are digitized satellite data, wind shear, persistence, climatology and sea surface temperatures. One would expect that TIPS provides better intensity forecast than STIFOR does.

NHC provides intensity analysis and forecasts in terms of maximum one-minute sustained surface wind speed (i.e. the maximum intensity), the 34, 50, and 64 kt wind speed radii for 12, 24, and 36 hour forecasts, and the 50 kt wind speed radii for 48 and 72 hour forecasts.

According to Sheets (1990), intensity forecasts are developed at NHC using empirical synoptic pattern recognition, satellite and aircraft data, and two statistical models, SHIFOR (Jarvinen and Neumann, 1979) and SHIPS (DeMaria and Kaplan,1994), the latter still being classified as an experimental model.

All of the empirical techniques are centered on a qualitative assessment of changes in the environment and environmental flow patterns affecting the TC. SHIFOR is a blend of climatology and statistics with the heaviest weight on climatology near 72 hours. The NHC forecasters use all of these techniques to build the intensity forecast, which is mutually dependent on the motion forecast. This information is provided to the field forecast offices and field users who may amplify the information based on local conditions.

One of the most important tasks assigned to forecasters is providing recommendations to area and task force commanders who set conditions of readiness. In order to generate the recommendations, forecasters must assimilate a tremendous amount of data in a short period of time. The following four-step procedure is recommended when making TC intensity forecasts.

Step 1: Pre-deployment. Forecasters are advised to understand the mission, and to review the Hurricane Havens and Typhoon Havens Handbooks (U.S. Navy 1976, and 1982), and the applicable Fleet Operation Orders that govern the ship's operations on deployment.

The handbooks indicate which ports are considered storm havens and which are not, and the maximum wind speeds that constitute sortie criteria. This information should be provided to operations officers and commanding officers during the pre-deployment briefings and prior to any expected port visits.

Step 2: Analysis. When a TC develops or moves into the forecast area, plot the warning and analyze the current synoptic environment to determine key synoptic features affecting the intensification/dissipation trend. This analysis should center on verification of the intensification and dissipation rate issued by the forecast center.

One suggested approach for analyzing the current synoptic environment is to plot 48 hours worth of past storm positions on the corresponding numerical analysis (e.g., NOGAPS 250-hPa analysis). Then, locate synoptic features within 30 degrees longitude and latitude of the TC by paging through the analyses. Try to identify the synoptic features that affect TC intensity. The synoptic features that affected TC intensity in the recent past may also affect it in the near future.

Next, determine intensity trends using intensity observations available over the past 24 to 48 hours. Each intensity observation should be compared with the previous observations from the same site and same type of observation platform. This is necessary because each fixing agency has its own unique method to determine TC intensity thaat depends on available equipment, data, and operating procedures.

Step 3: Forecast Verification. Verify the TC intensity forecast in the official forecast for accuracy and significant departures from previous forecasts. First, plot the official intensity forecasts on the verifying prognostic charts from the numerical model (e.g., the NOGAPS 250-hPa, 48-hour prognostic charts). Next, note the location with respect to major synoptic features identified during the previous step. An initial assessment of forecast confidence can now be made. Table 6.8 shows some examples of confidence value assignments.

If a low confidence scenario is identified during the intensity forecast verification step, try to understand the reasons for the official forecast trend. A review of the synoptic and prognostic discussions issued by forecast centers should shed some light on the situation. Many times the forecast center is moving from one primary forecast scenario to another (e.g., a maximum intensity of typhoon to a super typhoon or vice versa). In this case, the prognostic discussions should describe the alternate scenario.

Step 4: Forecast Recommendations. Brief the forecast intensity, but be sure that the customers also understand the TC intensity forecast errors. The intensity forecast has a mean absolute error of approximately 11 kt at 24 hours, 18 kt at 48 hours, and 24 kt at 72 hours (see Table 6.3). However, 72-hr errors of 50 kt or greater do occur over a 72-hour period several times each year.

In connection with this, consider a situation where ships must sortie from Guam when the winds are expected to exceed 50 kt sustained over the island. If a TC, forecast to be 45 kt as it crosses Guam, actually intensifies to 100 kt, ships remaining in port may be damaged. When the intensity forecast includes errors inherent in the forecast, users can make informed decisions, such as whether or not to sortie the ships in the case described above.

For the forecaster in harm's way, short-term loss of communications and other equipment failures frequently occur. This, however, does not preclude the requirement to make forecast recommendations. The forecaster who follows the procedures above will be able to make reasonable forecast recommendations to the on-scene commanders.

***** End of SECTION 6 *****

Revision date: 10 July 1998

Main

Chapter 6

Section 5 Appendix A

Section 6