Modeling insights and scientific interactions stimulated by DWFE and DTC

The Developmental Testbed Center (DTC)'s Winter Forecast Experiment (DWFE), conducted from January 10 through March 31, 2005, produced not only daily, 48-hour CONUS forecasts at 5 km resolution by two WRF model cores (NMM and ARW) but also a mechanism to promote interactions among human participants. These participants contributed, often synergistically, to find ways to derive value from these unprecedented model forecasts.

DWFE participants included the co-directors and small scientific and programming staff of the DTC, scientists at NCAR, and scientists at NOAA's Forecast Systems Laboratory (FSL), all in Boulder, CO; a university visitor to the DTC at NCAR; the WRF Program Coordinator in Camp Springs, MD; operational forecasters at several National Weather Service (NWS) Forecast Offices; and staff at NCEP in Camp Springs, MD. All but the NWS forecasters and NCEP staff interacted at various times in person in Boulder. The NWS forecasters contributed observations about DWFE model performances during periodic telephone conferences and via Web-based survey forms. NCEP staff helped set up DWFE and responded to questions about model configuration and performance by phone and email once it was underway.

The participants brought to bear a variety of backgrounds, biases, and tools to the task of interpreting DWFE forecasts. This collective experience helped participants to identify technical problems in several DWFE modeling system components (including the initialization, model, and verification software) and raise scientific questions about the configuration and formulation of the models, which some participants subsequently have pursued in greater depth.

The mechanisms of human interaction created by DTC to support DWFE were crucial to making practical and scientific insights possible and should lead to practical improvements in future high-resolution operational forecasting efforts.