The Second Prospectus Development Team (PDT-2) of the U.S. Weather Research Program was charged with identifying research opportunities that are best matched to emerging operational and experimental measurement and modeling methods. The overarching recommendation of PDT-2 is that inputs for weather forecast models can best be obtained through the use of composite observing systems together with adaptive (or targeted) observing strategies employing both in situ and remote sensing. Optimal observing systems and strategies are best determined through a three-part process: observing system simulation experiments, pilot field measurement programs, and model-assisted data sensitivity experiments. Furthermore, the mesoscale research community needs easy and timely access to the new operational and research datasets in a form that can readily be reformatted into existing software packages for analysis and display. The value of these data is diminished to the extent that they remain inaccessible. The composite observing system of the future must combine synoptic observations, routine mobile observations, and targeted observations, as the current or forecast situation dictates. High costs demand fuller exploitation of commercial aircraft, meteorological and navigation [Global Positioning System (GPS)] satellites, and Doppler radar. Single observing systems must be assessed in the context of a composite system that provides complementary information. Maintenance of the current North American rawinsonde network is critical for progress in both research-oriented and operational weather forecasting. Adaptive sampling strategies are designed to improve large-scale and regional weather prediction but they will also improve diagnosis and prediction of flash flooding, air pollution, forest fire management, and other environmental emergencies. Adaptive measurements can be made by piloted or unpiloted aircraft. Rawinsondes can be launched and satellites can be programmed to make adaptive observations at special times or in specific regions. PDT-2 specifically recommends the following forms of data gathering: a pilot field and modeling study should be designed and executed to assess the benefit of adaptive observations over the eastern Pacific for mesoscale forecasts over the contiguous United States; studies should be done over the western Atlantic and Caribbean-Gulf of Mexico regions, particularly during hurricane season; and enhanced observations should be implemented for the mountainous western states and for the Mississippi and Missouri River Valleys. Data sensitivity tests and observing system simulation experiments (OSSEs) are important tools for gauging the benefit of existing or proposed observing systems. OSSEs involve only model predictions and are essentially self-contained. Data sensitivity tests require the full consideration of modeling infrastructure, namely, observation ingest quality control, objective analysis, and numerical prediction. Sensitivity tests involving both wind and moisture profiles are particularly needed to determine their impact on improved precipitation forecasts. New variational analysis techniques are capable of assimilating so-called proxy observations. These techniques should be fully exploited. Diabatic initialization should be addressed through the assimilation of satellite cloud data and very high resolution WSR-88D radar measurements into very high resolution models with sophisticated cloud microphysics. Success in this area should improve quantitative precipitation forecasts in the first few (model) hours. There is a pressing need to better understand the interaction of moist convection with large-scale flow. One key is better characterization of the impact of precipitation formation and evaporation on the fluxes of mass, momentum, and heat in moist convection. Humidity measurement in precipitating downdrafts is a crucial measurement, which currently cannot be made reliably. The capabilities of polarization-diversity radar should be explored in a quasi-operational context to determine whether WSR-88D radars should be upgraded. Progress in quantitative precipitation forecasting is impeded by poorly resolved and inaccurate water vapor measurements. Further improvements in numerical weather prediction demand improved monitoring of Earth surface characteristics so that spatial and temporal variations in air–surface fluxes are realistically simulated. Over land, priority should be given to the coupling of mesoscale meteorological models with hydrological models and to routine assimilation of surface (soil, moisture, and plant) characteristics. Improved air–sea fluxes are essential to proper modeling of marine cyclogenesis. The most important, practical ocean measurements include sea surface temperature, thermocline depth, wave spectra, and ice coverage and thickness.
Read full abstract