2
Transition from Research to Operations in Weather Satellites and Numerical Weather Prediction
Weather sensitive sectors of society include those involved with energy generation, agriculture, forestry, fisheries, construction, recreation, tourism, transportation and navigation, public utilities (energy purchase, electricity distribution, and capacity planning), retail trade distribution and stocking, finance, insurance and re-insurance, recreation, and real estate (NRC, 1998a). The utility of this information is diverse—it impacts military operations and staging; commercial airline scheduling, operations and flight planning; space launch scheduling; agriculture crop selection, planting, cultivation, and harvest timing; water resource management; and a wide range of commercial industries that schedule outdoor activities (e.g., construction, transportation). Information about extreme weather, especially that which puts life and property at risk, is essential for all sectors, but particularly for the emergency management, preparedness, and disaster relief communities. Both the public and private sectors exhibit a growing demand for accurate information about and prediction of extreme weather and climate events.
To meet this demand the federal government organizes the nation's weather prediction responsibilities into two related areas. First, operations are the basis for production and dissemination of official forecasts and warnings. (Operational services are also divided between public sector predictions, both civilian and military, and private sector, value-added dissemination and prediction services.) Second, research, systems development, and technology development and implementation are
supported to improve the skill of weather forecasts. These activities are in some cases tightly coupled to operational efforts, while others have a weaker connection. Research is carried out in both federal laboratories and universities. Research and development, whether in federal laboratories or universities, is largely supported by the federal government, and amounts to about $500 million each year (OFCM, 1998).
A challenge facing the meteorological community is to reduce impediments that limit the efficient transfer of weather and climate research findings into improved operational forecast capabilities that will benefit a wide group of users in the United States and countries that rely on U.S. forecasting leadership. To place the report's subsequent discussion of numerical weather prediction and satellite technology into a broader context, this chapter discusses the relation of research and operations in the meteorological community.
The operational forecast system is responsible for collecting and assembling data and for using that data, in conjunction with models, to produce forecast products in a timely fashion. Consequently, the system encompasses many elements, from the instruments on land, on the ocean surface, in the atmosphere, and in space, to the computational resources required to create, display, and disseminate the products. All elements of the system can be improved, and both the private sector and the academic research and development communities can contribute to that improvement. There are prohibitive impediments to improvement, however. Because the system is operational and is required to provide service 24 hours a day, 7 days a week, it cannot be taken off line for improvements to be made. Therefore, testing of changes in parallel to the operational forecasts is an essential precursor to adopting improvements to ensure that the changes introduced do not degrade the forecast skill. Yet EMC lacks the computational resources to conduct parallel testing.
IMPROVING THE FORECAST SYSTEM
Improvements in observational capabilities, assimilation techniques, and forecast models have the potential to increase the timeliness and accuracy of the forecast products. Both private sector and academic research and development organizations have driven such improvement in the past, and can be expected to do so in the future. These organizations do not have to remain on line to provide an operational flow of
observational data and forecast products; therefore they are free to concentrate on advancing the state of the art, raising effectiveness and skill in observing, assembling observations, and making predictions.
Strong interaction between the research and operational communities can improve the transition process. If the research community produces new science, one would expect opportunities to improve operations to exist. But without effective transitions from research or a dialogue between research and operations about system performance, improvements to the skill of the operational forecast system will be slow. The potential forecast skill based on current research understanding, state-of-the-art sensors, and computers is expected to be higher than that of the operational system. Verification of forecast skill and ongoing dialogue about performance should guide operational practices toward improvement.
Key issues for an operational forecast system are to ensure that transitions do indeed result in improvements and that the effort required for the transition is not disruptive. Improvements have been noted in the past such as the development of numerical weather prediction (1948), the first climate forecast of warm sea surface temperature in the tropical Pacific (1986), and the realization that chlorofluorocarbons photodissociate in the stratosphere depleting the ozone layer (1974). Feasibility must be demonstrated for the entire operational process, and production of additional climate and weather information must be accompanied by considerations of its dissemination, use, and impact. Impact may be related to demand, which depends on visibility of the information and recognition of its value. In certain cases (e.g., human health) public perceptions of hazards might result in demands for new products even though the ability to accurately forecast and quantify health hazards may be limited. The private sector also is capable of visualizing new products that, when advertised to the public, will generate demand.
Transition to Operations
If the new information has sufficient value, then a transition to operational status is desirable. The major challenges in accomplishing such a transition are institutional. Observations, modeling and prediction, and information dissemination to users should be tightly linked, and financial support of the operational system requires long-term commitment.
The transition plan should be sensitive to the research community that developed the system. Ideally, the new system should serve a dual purpose by continuing to serve the needs of the research community and continuing to promote advances in data analysis and use. To that end, the research community should be involved as advisors, and sufficient resources should remain available for continued exploration.
Not only should the operational structure include sufficient commitment to maintain critical observations, modeling, prediction, and information dissemination, but it should also include a continued interaction with the research community to promote the opportunities to maintain a state-of-the art capability6. The continued dialogue between the research and operational communities is needed to guarantee that the latest techniques and current knowledge are available to the operational services. This linkage will ensure that the latest research continues to be available for operational use. The operational services will be able to keep abreast of the latest research, to continuously assess the observing and simulation systems, to persistently determine how the information can be improved, and to consistently interface with the user community in the design of new useful products. The dialogue should include interested members of the research and user communities, and ongoing surveys should continually assess changing user needs.
GUIDELINES FOR THE TRANSITION PROCESS
BASC has selected the following criteria as key to an effective transition for the field of weather and climate prediction:
-
A strong research program, including understanding of the role of the operational community in the broader context of the weather prediction system (NRC, 1998a, 1999c).
-
A healthy infrastructure for transition. The forecasting system needs an observation, technology, and modeling capability that serves as a foundation for research and permits the demonstration of the potential for useful new products without drawing resources away from the operational forecast system. There is a need for a long-term commitment of adequate resources to maintain the research or operational programs. Mechanisms should be developed to enable continuous development and
6 |
For a discussion of the changing paradigm for research to operations transition, see “Beyond Basic and Applied” (Pielke and Byerly, 1998). |
-
maintenance of state-of-the-art capabilities (Recommendation 4, NRC, 1999c).
-
Strong interface with the user community (Pielke and Byerly, 1998).
-
International observation and data access partnerships (NRC, 1998a; Recommendation 8, NRC, 1999c).
-
Continuous evaluation processes of each of the components of the weather prediction system as well as its subcomponents (NRC, 1998a).
In order to focus attention on the strengths and weaknesses of current U.S. efforts, the following two chapters apply the principles contained in these guidelines to two critical aspects of the U.S. weather and climate forecasting enterprise.