A priori selection and data-based skill assessment of reanalysis data as predictors for daily air temperature on a glaciated, tropical mountain range

Abstract

Abstract. This study presents an empirical-statistical downscaling (ESD) method for high-altitude, glaciated mountain sites. In the ESD model validation emphasis is put on appropriately considering the pitfalls of small observational data records that are typical of high mountains. An application example is shown with daily mean air temperature time series on a glaciated mountain range (Cordillera Blanca, Peru) as target variables, and an ensemble of reanalyses air temperature time series as "a priori" predictor (i.e. a predictor selected without preceding data analysis). Results reveal strong seasonal variations of the predictor's performance. With increasing data availability, the skill tends to increase. Similarly for lower temporal resolutions, the skill increases. Applied to a choice of different atmospheric reanalysis predictor variables, the ESD model identifies only air temperature and geopotential height as significant predictors with regard to local-scale air temperature variability. Accounting for natural periodicity in the data is vital in the ESD procedure to avoid spuriously high performances of certain predictors, which is demonstrated for 2 m air temperature versus air temperature in the pressure level close to the mountain station site.