Close Window

### Final Report

SECTION 1: PROJECT OBJECTIVES AND ACCOMPLISHMENTS

The main objective was to perform a real-time forecast experiment to predict 12-h precipitation amounts for Puerto Rico. The goals were to develop a stand-alone statistical model to compare with the current guidance products for issuing quantitative precipitation forecasts (QPFs). The purpose was to increase our understanding of tropical rainfall predictability in the presence of rugged terrain.

Tasks included: (1) develop climatologies, (2) perform regionalizations, (3) develop a statistical forecast model, and (4) perform and verify a real-time forecast experiment.

Based on the tasks outlined in the proposal and the tasks accomplished, the project was successful.

James Elsner was responsible for directing the project including data analysis and decision making. He was assisted by Matthew Carter. Mr. Carter performed most of the calculations. His PhD dissertation provides a synopsis of the real-time forecast experiment including results and lessons learned.

Shawn Bennett was responsible for organizing the real-time forecast experiment. It was necessary for Mr. Bennett to emphasis to his forecasters the importance of their participation in the experiment. His expertise was consulted in regards to weather factors conducive to heavy rainfall events over the island.

Lessons learned: Scientific results from this cooperative project have shown the utility of climatology for precise specification of rainfall probabilities. The problem is that climatologies of this sort are not readily available to the forecasters, or are not available in a useful format. Results also show the diurnal variability in the relationship between quantitative rainfall probabilities and topography. Topography explains a significant portion of the geographic distribution of probabilities during the late afternoon and evening.

The probabilistic quantitative precipitation forecast experiment has provided the WFO with the objective scientific evidence and reasoning for the prevalence of heavy rainfall in certain sectors of the island. This knowledge replaces subjective reasoning and hand waving arguments used to forecast precipitation location and quantity in Puerto Rico. This work improved the issuance of QPFs from WFO San Juan as forecasters have begun to integrate this new scientific evidence into their daily forecast operation. In addition, this work provides a sound basis and launching point for future scientific research.

SECTION 2: SUMMARY OF UNIVERSITY/NWS EXCHANGES

Our COMET research was integrated into the CITM/SOO workshop held at FSU during the second week of May 1998. The workshop theme was directed at the University link with the NWS Offices in Florida. Our COMET work was highlighted with a joint seminar (with Shawn Bennett) in the section on QPF, with an emphasis on knowledge transfer from Puerto Rico to Florida. The workshop was used to establish a collaboration with the NWS Office in Mobile, AL. These activities were not directly funded by COMET nor were they part of the original proposal.

In April of 1999, the PI was invited to the Second Caribbean Climate Outlook Forum to give a talk on hurricane activity. The workshop provided a forum for discussion on seasonal precipitation forecasts for the Caribbean, including Puerto Rico. The participants, represented various nations of the region, issued a consensus three-month seasonal forecast of rainfall.

A Junior Forecaster from WFO San Juan, Mr. Andy Roche competed analysis of rainfall data collected from rain gages and WSR-88D derived rainfall estimates for specific heavy rainfall events. He reduced the data and developed the contouring and plotting routines using GEMPAK. His studies were inspired and designed to be complementary to the COMET Cooperative Project. His studies were funded under a NOAA/NWS University Assignment to Florida State University. Mr. Roche finished course work necessary for a graduate degree in meteorology under this program.

The project funded a joint paper submitted to the {\it Journal of Hydrology} concerning the forecast experiment.

SECTION 3: PRESENTATIONS AND PUBLICATIONS

1. Carter, M. M., 1999: Interannual variability of rainfall in Puerto Rico. Preprints, 23rd Conf. on Hurricanes and Tropical Meteorology, Dallas, TX, Amer. Meteor. Soc. 551--552.

2. Carter, M. M., 1999: A Quantitative Precipitation Forecast Experiment For Puerto Rico", PhD Dissertation, Department of Meteorology.

3. Carter, M. M., J. B. Elsner, and S. Bennett, 2000: A quantitative precipitation forecast experiment for Puerto Rico. Submitted.

SECTION 4: SUMMARY OF BENEFITS AND PROBLEMS ENCOUNTERED

BENEFITS:

The principal benefit to the FSU has been the infusion of knowledge concerning forecasting precipitation over a tropical island.

This real-time forecast project provided a focus for understanding the relationship of precipitation forecasts in support of user services. Although we have yet to complete the project, we envision that the knowledge generated will provide San Juan forecasters with an enhanced awareness of subjective biases inherent in heavy rainfall forecasts. This will lead to more useful forecasts to the public thereby saving lives.

The study sponsored by this COMET Cooperative Project is a first of its kind for Puerto Rico. The probabilistic quantitative precipitation forecast experiment has provided the WFO with the objective scientific evidence and reasoning for the prevalence of heavy rainfall in certain sectors of the island. This knowledge replaces the heretofore subjective reasoning and hand waving arguments used to forecast precipitation location and quantity in Puerto Rico.

This work has improved the QPF from WFO San Juan as forecasters have begun to integrate this new scientific evidence into their daily forecast operation. In addition, this work provides a sound basis and launching point for future scientific research.

PROBLEMS:

During the project period, the PI transferred his professorship to the Department of Geography. This created a bit of down time in terms of collaboration between the PI and San Juan.

Perhaps the most challenging event was the turn over in the SOO position. Shawn Bennett moved to WFO Brownsville and Rachel Gross took over as SOO at WFO San Juan. These challenges were overcome via maintaining close and open communication between the SOOs and Dr. Elsner at FSU. Shawn consulted with both FSU and San Juan on the Project when needed.

The landfall of destructive hurricane Georges during September of 1998 shifted the priority of the experiment. This caused a slight delay in the implementation of the forecast at the San Juan office.

The location of Puerto Rico and the slowness of Internet connections for receipt of model data etc, made collaboration more challenging. It was handled by judicious scheduling of communications and calculations.