The National Center for Atmospheric Research (NCAR) AutoNowCaster (ANC) provides short-term (0-60 minute), time- and space-specific nowcasts of thunderstorm intensity using fuzzy-logic techniques. The ANC has been successfully deployed at multiple, land-based sites to produce automated nowcasts in operational environments. The ANC uses a variety of input data that includes GOES satellite imagery, Doppler radars (like the SPY-1/TEP), numerical model output, surface and upper air data, among others. The ANC has had considerable development and testing on continental convection. The ANC has also been successfully tested on its ability to nowcast oceanic convection (Kessinger et al, 2001).
For this paper, the performance of the ANC is evaluated with the 15 May 2000 case. In oceanic regions, radars have a limited ability to detect boundary layer convergence zones due to the absence of insects. These convergence zones are typically important precursor indicators for the initiation of convection. In Kessinger et al (2001), the convergence zone position was manually entered into the ANC because the TEP could not detect it. For naval operations, automation is critical. Therefore, the radar-based techniques used to locate the convergence zones have been modified such that satellite imagery can be used instead. Additional techniques are under development to better utilize the GOES satellite imagery within the ANC. This is desired since satellite imagery can extend the ANC region past the TEP maximum range. These additional techniques include using output from the Naval Research Laboratory (NRL) cloud classification algorithm in various ways.
Kessinger, C., R. Roberts, M. Chapman, M. Meister, and T. Tsui, 2001: Thunderstorm nowcasting in an oceanic environment: A feasibility study. Preprints, 30th International Conference on Radar Meteorology, American Meteorological Society, Munich, Germany, 19-25 July 2001, 258-260.
Supplementary URL: http://www.rap.ucar.edu/staff/kessinge/Sat03_Kessinger.pdf