Generating reliable meteorological data in mountainous areas with scarce presence of weather records: The performance of MTCLIM in interior British Columbia, Canada

 We tested the weather extrapolator MTCLIM for three sites in an arid region.  We provided a detailed validation of MTCLIM with statistical and graphical analysis.  MTCLIM acceptably extrapolates temperatures even from short weather data series.  MTCLIM performance for precipitation is acceptable only for monthly values.  MTCLIM creates acceptable weather records if high daily accuracy is not needed. Abstract Climate models have an important role in biometeorological research in mountainous areas where few, dispersed and relatively short data records are the norm. Weather-extrapolator models are a possible solution and we tested the performance of the mountain microclimate simulation model (MTCLIM, a meteorological point data extrapolator) in three arid sites in southern interior British Columbia, representing a gradient of available meteorological information. Measures of several goodness-of-fit indices (Pearson's correlation coefficient, coefficient of determination, mean error, mean absolute error, modeling efficiency and Theil's inequality coefficient) and equivalence tests showed that MTCLIM simulated temperature better than precipitation and performed inside the accuracy requirements for our dendrochronological studies for both variables even with short data series. Histograms showed that predicted daily Tmax and Tmin in this arid area had some seasonal biases, probably influenced by the presence of a large water body nearby. In long-term ecological and dendrochronological studies, temporal changes of climate variables at monthly or yearly scales are usually more important than their absolute values at daily scale, and in the present study the histograms of observed and data simulated by MTCLIM at those scales were similar. Therefore, we conclude that MTCLIM can extrapolate reliable weather data for use in ecological studies in arid mountainous terrain, provided there is an adequate weather record at the reference station and proper information to calculate the input parameters: lapse rates and precipitation isohyets.