The incorporation of hourly GOES data in a surface heat flux model and its impacts on operational temperature predictions in bodies of water

Accurate information on cloud cover and surface heat flux is critical to the success of surface water model predictions. To predict temperature, a numerical model requires surface heat flux values which are calculated based on cloud cover and other meteorological variables. Existing surface heat flux formulations overestimate the values due to underestimation of cloud cover and predicted surface water temperatures are correspondingly higher than that observed from measured data. Lack of adequate spatial and temporal resolution of cloud data and improper representation of cloud cover in the interpolation scheme are the two major sources hypothesized to cause the surface heat flux over prediction. To correct this problem, satellite-derived cloud cover data are utilized to compute the net surface heat flux. Since the Geo-stationary Operational Environmental Satellite (GOES) provides extremely high spatial and temporal resolution of data (hourly frequency at 1 kilometer), correct cloud cover information can be obtained to provide more accurate surface heat flux estimation. The purpose of this study is to determine the effect of satellite-derived cloud cover data on surface heat flux estimates and to evaluate the resulting heat flux on Lake Erie surface water temperature predictions. In this research, more than 3600 GOES-8 satellite images are analyzed to derive cloud cover information. The performance evaluation is based on the results from two different simulations: a two-week cloud cover sensitivity test on surface heat flux and a season long simulation starting from mid May to the end of October, 1995. Traditional statistical measures and Empirical Orthogonal Function (EOF) analysis were used as evaluation tools. We found that adequately robust cloud cover results in a 50% reduction in the magnitude of solar radiation and GOES-8 derived cloud cover yields more accurate cloud cover information and better surface heat flux estimation. Therefore the model predicted Lake Erie surface water temperatures will drop 2 to 3 degrees C in average and compare much more favorably to observed values.