The National Weather Service (NWS) operational responsibilities for the tropical and subtropical Atlantic are focused around the National Centers for Environmental Prediction (NCEP), namely the Hydrometeorological Prediction Center (HPC) in Camp Springs, Maryland, and the Tropical Prediction Center/National Hurricane Center (TPC/NHC) in Miami, Florida, as well as forecast offices at Miami, Florida, and San Juan, Puerto Rico. Historically, there has been only limited face-to-face interchange at the forecaster level between the national centers and NWSFO San Juan. Coordination visits by lead forecasters from NWSFO San Juan to these national centers can improve information exchange and comprehension and help to identify and utilize new forecast technology and techniques. In general, cross-visits can improve communication and knowledge of the strengths and weakness that each office possesses.
The benefits of such coordination visits were realized when NWSFO San Juan sent three lead forecasters to visit HPC and SPC during late June and early July 1996. These visits provided the ideal forum to address operational concerns, become personally acquainted with a number of NCEP supervisory and operations personnel, and strengthen the relationship between these national centers and NWSFO San Juan in an effort to improve tropical analysis and forecasting.
Three lead forecasters--Ron Block, Scott Stripling, and Roberto Garcia--spent their initial one and one-half days at NCEP/HPC. Given San Juan's location and the purpose of this assignment, logically we focused on interaction with NCEP's International/Tropical Desk. During the initial afternoon, Michael Davison, chief of this desk, oriented us to its history, functions, and personnel. Its genesis is traced to the desire of several South American countries to better identify and forecast climate anomalies spawned by the 1989 El Niño event. As a result, the WMO began sponsoring Caribbean and Latin American meteorologists at the International Desk, generally for four months. Their training focuses on the analysis of raw data coupled with an emphasis on the interpretation of computer model performance in the tropics and a total immersion in NCEP technology. Generally four to five guest meteorologists participate per training session, and their performance has been instrumental in improving model precipitation fields across the tropics and subtropics.
We observed how the visiting WMO meteorologists were daily divided into the Northern and Southern Hemisphere teams. Much of this time was spent on an analysis of satellite imagery and/or hand analysis of raw data with N-TRANS and MCIDAS the primary tools utilized. By late morning, team members briefed the others present on their interpretations, stressing model comparison versus satellite in an attempt to determine the best model of the day. Feedback was provided by Mr. Davison and the San Juan forecasters; model agreement was reached; and after final edits, the daily tropical bulletin was disseminated.
During the morning of day two, the WMO meteorologists conversed with us on various aspects of their positions, both at NCEP and in their home countries. This provided some insight on the unique limitations encountered at each site. Of particular note to San Juan is the use by the WMO meteorologists of our state forecast discussions (SFD) as a primary briefing tool. They expressed some difficulty interpreting this product due to the utilization of NWS rather than WMO contractions, as well as the use of unfamiliar terminology. They encouraged us to limit, and if possible, eliminate contractions, stating that similar concerns are echoed by other tropical meteorologists. We learned from Mr. Davison which models perform best in the tropics, especially across the northeast Caribbean. From his experience, the ECMWF and the MRF rate best during the summer and winter, respectively.
The morning ended with the daily weather briefing performed in tandem by the senior duty meteorologist, who discussed model comparison, and the senior analyst who briefed on satellite interpretation and the dynamic and synoptic patterns of the day. We were impressed by their utilization of computer graphics presented on a big screen television.
The afternoon was spent interacting with other scientists and desks. Dr. Peter Kaplen, chief modeler, discussed present and future model trends with particular reference to the northeast Caribbean. He agreed that the ECMWF performs best, while the principal tasks of modelers remain improving all model accuracy beyond 24 hours. At the NESDIS desk (collocated with NCEP), the Dvorak technique for estimating precipitation was demonstrated, and it was applied to tropical storm Arthur then brewing near the North Carolina coast. Likewise, we viewed the decision tree, along with real-time data, for developing satellite precipitation estimates (SPENES). The need for routing SPENES prepared for San Juan during adverse weather was discussed. Both NESDIS and we agreed that this product and subsequent quality control would benefit both sites and should be pursued.
The HPC SOO, Dr. Russell Schneider, and several researchers discussed NCEP operations and limitations. Marine Branch forecasters displayed their model output, suite of products, and types of quality control; and they expressed a desire for more real-time data from the Caribbean. NESDIS researchers discussed available resources and research trends. Finally, the quantitative precipitation forecasters presented their methods of analysis and technology for issuing a QPF.
The final two days were spent at NCEP'S Tropical Prediction Center/National Hurricane Center in Miami. After meeting with Mr. Jerry Jarrell, Deputy Director, to open day one, we were introduced to Dr. Lixion Avila, the hurricane forecaster on duty who oriented us to the functions of his position. Fortunately (for us, at least), tropical storms Arthur (Atlantic) and Alma (Pacific) were intensifying this day, which provided an ideal opportunity for a timely hands-on observation and training in regard to NHC operations. The decision-making processes inherent in the issuance of a watch and warning and the methods of product quality control were noted. We were impressed with the thoroughness with which Dr. Avila analyzed all model and raw data, interfaced with emergency managers and the media, and utilized the hurricane hotline to receive input and coordinate with NCEP and the relevant field offices. One surprise was the discovery that the hurricane forecaster does not normally have support staff to assist in some of the more time-consuming functions. In fact, his only direct assistance during our visit was borrowing the Doppler focal point, Colin McAdie from the NOAA/ERL Hurricane Research Division, to provide fixes on Arthur. From Mr. McAdie we received hands-on training on how to determine the exact location of a hurricane eye.
During the afternoon, we talked with the NHC SOO, Dr. G. Jiing, who provided input on the new tropical weather discussion (TWDAT) format. Despite overall approval, reservations were expressed regarding: (1) the elimination of the subcategory of Puerto Rico, replaced by the more general Caribbean; and (2) the use of the TWDAT issuance time rather than the more timely former satellite precipitation message (SIM) time. Discussions with Tropical Analysis and Forecasting Branch (TAFB) meteorologists broached the topic of the paucity of upstream data for San Juan and the need for greater guidance from NHC forecasters.
The second day began by observing the decision-making process coordination among Dr. Avila, Mr. Jarrell, and Dr. Steven Lyons (chief of TAFB) regarding the naming of Alma. The ultimate decision rested with the hurricane forecaster, as everyone deferred to his experience. During the late morning, Dr. Jiing discussed the evolution and genesis of easterly waves. He presented several contrasting theories of wave formation and interjected several reasons why waves are incorrectly identified, suggesting some rules-of-thumb that should facilitate our analysis and forecast of these systems at the NWSFO.
During the afternoon, we toured TAFB facilities, observing how each of the three desks prepared their products. This included the analyst and marine and aviation forecasters who prepare the TWDAT, offshore forecast and tropical aviation products, respectively. Dr. Lyons--who doubled as marine forecaster--discussed marine guidance, including a new PC-based program for wind/wave forecasting and a second one for tropical cyclones. We took these back to San Juan for installation. Personnel from the Hurricane Research Division introducedtheir library and other research facilities. When we queried them about generating more publications in tropical meteorology, they indicated that most of their time and resources are committed to improving NOAA technology, for example N-trans. The computer support team demonstrated how to extract time-sections from N-trans sounding data, valuable for case studies being pursued at San Juan.
Impressions and Conclusions
This long overdue visit to the NCEP/HPC and NCEP/TPC/NHC was deemed a success by all involved. We were cordially received and permitted to interface with all the available personnel and desks. Everyone gained a greater appreciation for each other's mission, resources, and limitations. We discussed the role of San Juan as the front line field office for Atlantic Basin tropical meteorology and how we can assist these centers in their mission. Such an exchange is important to facilitate future inter-office communication, coordination, and timely exchange of forecast information.
Several highlights and themes pervaded our visits. At every desk meteorologists were performing some type of hand-analysis of raw data. The lesson reiterated was that despite the flourish of state-of-the-art modernization technology, the time-worn process of examining real-time data remains an imperative precursor to all other forms of analysis and guidance. We were surprised that the hurricane forecaster did not possess a retinue of support staff, having to personally execute even the most mundane functions.
The mission and accomplishments of the International Desk are impressive. Most San Juan forecasters are mid-latitude trained with relatively limited tropical experience. We agree with NCEP personnel that they would benefit considerably by being exposed to the expertise and technology available at the International Desk. At the same time, the NWSFO operational experience would enhance the skills of NCEP forecasters and provide an added dimension to the training of the WMO meteorologists. We recommend that to the extent possible San Juan forecasters engage in more frequent visits, and for longer intervals (at least one week), to become immersed in their training.
This was an ideal opportunity for San Juan forecasters to visit and interchange with the relevant national centers. The authors, at least, experienced for the first time the advantages and frustrations of the people on the other end of the telephone line. Many concerns were expressed which can only facilitate inter-office exchanges and improve our joint mission. Other forecasters are encouraged to take advantage of any opportunity to visit other offices.
Many people at NCEP, TPC, and San Juan contributed to this visiting forecaster program, although only a few are mentioned in this summary. We are particularly indebted to Mike Davison of NCEP and Dr. Lixion Avila and Dr. J. G. Jiing of TPC for their time and efforts in making us feel like part of the team. Shawn Bennett (SOO, NWSFO San Juan) was instrumental in turning this trip into a reality and in providing excellent input into this paper. Israel Matos (MIC, NWSFO San Juan) was also very supportive of this program and also supplied excellent suggestions to improve this presentation. Robert Garcia, lead forecaster and trip participant, added input.