Improving boundary conditions of a numerical weather prediction model for the Maltese Islands

Abstract

Several aspects of modern society have come to depend on accurate and regular weather forecasts which allow them to make strategic and informed decisions, in order to preserve and maintain their assets. Due to the extent of the calculations involved in meeting the accuracy and quality requirements, NWP models are used. One of the most commonly used NWP systems is the WRF model. However, the default boundary conditions for the Maltese Islands contain data irregularities that limit the accuracy of the model output. Therefore, the aim of this study was to improve these boundary conditions by revising the land use categories and identify any significant improvement to weather forecasts made thereafter. The land use categories, which followed the USGS 24-category Land Use Categories, were significantly improved in terms of category type and distribution. To determine the effectiveness of these improved boundary conditions, the precipitation and temperature forecasts were generated with the default and improved land use categories. These were then compared against the meteorological observations where an overall improvement in forecast accuracy was identified.