Testing the applicability of physiographic classification methods toward improving precipitation phase determination in conceptual models

Abstract

Abstract Regions with a large percentage of precipitation occurring near freezing experience high percentages (>10%) of misclassified precipitation events (rain versus snow) and necessitate efforts to improve precipitation phase determination schemes through the use of more accurate surface air temperature thresholds (Trs). Meteorological data from 169 sites in Scandinavia were used to test the applicability of using physiographic categories to determine Trs. Three classification methods involving varying degrees of automation were evaluated. The two automated methods tested did not perform as well as when tested on a smaller region, showing only 0.16% and 0.20% reduction in error. A semi-manual method produced the largest average reduction in misclassified precipitation (0.53%) across all sites. Further refinement of classification criteria for mountain and hill stations showed that at mesoscales (>5 km), maximum elevation is a better predictor of Trs (0.89% average reduction in error) than terrain relief (0.22%), but that relief becomes increasingly important at microscales (0.90%). A new method for categorizing mountainous stations based on upslope or downslope air movement increased the average reduction in error up to 0.53%. These results provide a framework for future landscape classification methods and confirm the importance of microscale topography for determining Trs in alpine regions.