It is well known from previous research that significant differences exist amongst reanalysis products from different institutions. Here, we compare the skill of NCEP-R (reanalyses by the National Centers for Environmental Prediction, NCEP), ERA-int (the European Centre of Medium-range Weather Forecasts Interim), JCDAS (the Japanese Meteorological Agency Climate Data Assimilation System reanalyses), MERRA (the Modern Era Retrospective-Analysis for Research and Applications by the National Aeronautics and Space Administration), CFSR (the Climate Forecast System Reanalysis by the NCEP), and ensembles thereof as predictors for daily air temperature on a high-altitude glaciated mountain site in Peru. We employ a skill estimation method especially suited for short-term, high-resolution time series. First, the predictors are preprocessed using simple linear regression models calibrated individually for each calendar month. Then, cross-validation under consideration of persistence in the time series is performed. This way, the skill of the reanalyses with focus on intra-seasonal and inter-annual variability is quantified. The most important findings are: (1) ERA-int, CFSR, and MERRA show considerably higher skill than NCEP-R and JCDAS; (2) differences in skill appear especially during dry and intermediate seasons in the Cordillera Blanca; (3) the optimum horizontal scales largely vary between the different reanalyses, and horizontal grid resolutions of the reanalyses are poor indicators of this optimum scale; and (4) using reanalysis ensembles efficiently improves the performance of individual reanalyses.

Pronounced hygric seasonality determines the regional climate and, thus, the characteristics of rain-fed agriculture in the Peruvian Callejón de Huaylas (Cordillera Blanca). Peasants in the Cuenca Auqui on the eastern slopes above the city of Huaraz attribute recently experienced challenges in agricultural production mainly to perceived changes in precipitation patterns. Statistical analyses of daily precipitation records at nearby Recuay (1964 to 2013) and Huaraz (1996 to 2013) stations do not corroborate the perceived changes. Either insufficient temporal resolution of available precipitation records or other environmental and sociopolitical factors impacting traditional farming methods may be the reason for the lack of concordance between the two information sources investigated in this study.

This study presents a statistical downscaling (SD) method for high-altitude, glaciated mountain ranges. The SD method uses an a priori selection strategy of the predictor (i.e., predictor selection without data analysis). In the SD model validation, emphasis is put on appropriately considering the pitfalls of short observational data records that are typical of high mountains. An application example is shown, with daily mean air temperature from several sites (all in the Cordillera Blanca, Peru) as target variables, and reanalysis data as predictors. Results reveal strong seasonal variations of the predictors' performance, with the maximum skill evident for the wet (and transitional) season months January to May (and September), and the lowest skill for the dry season months June and July. The minimum number of observations (here, daily means) required per calendar month to obtain statistically significant skill ranges from 40 to 140. With increasing data availability, the SD model skill tends to increase. Applied to a choice of different atmospheric reanalysis predictor variables, the presented skill assessment identifies only air temperature and geopotential height as significant predictors for local-scale air temperature. Accounting for natural periodicity in the data is vital in the SD procedure to avoid spuriously high performances of certain predictors, as demonstrated here for near-surface air temperature. The presented SD procedure can be applied to high-resolution, Gaussian target variables in various climatic and geo-environmental settings, without the requirement of subjective optimization.

The evaluation of models in general is a nontrivial task and can, due to epistemological and practical reasons, never be considered complete. Due to this incompleteness, a model may yield correct results for the wrong reasons, i.e., via a different chain of processes than found in observations. While guidelines and strategies exist in the atmospheric sciences to maximize the chances that models are correct for the right reasons, these are mostly applicable to full physics models, such as numerical weather prediction models. The Intermediate Complexity Atmospheric Research (ICAR) model is an atmospheric model employing linear mountain wave theory to represent the wind field. In this wind field, atmospheric quantities such as temperature and moisture are advected and a microphysics scheme is applied to represent the formation of clouds and precipitation. This study conducts an in-depth process-based evaluation of ICAR, employing idealized simulations to increase the understanding of the model and develop recommendations to maximize the probability that its results are correct for the right reasons. To contrast the obtained results from the linear-theory-based ICAR model to a full physics model, idealized simulations with the Weather Research and Forecasting (WRF) model are conducted. The impact of the developed recommendations is then demonstrated with a case study for the South Island of New Zealand. The results of this investigation suggest three modifications to improve different aspects of ICAR simulations. The representation of the wind field within the domain improves when the dry and the moist Brunt–Väisälä frequencies are calculated in accordance with linear mountain wave theory from the unperturbed base state rather than from the time-dependent perturbed atmosphere. Imposing boundary conditions at the upper boundary that are different to the standard zero-gradient boundary condition is shown to reduce errors in the potential temperature and water vapor fields. Furthermore, the results show that there is a lowest possible model top elevation that should not be undercut to avoid influences of the model top on cloud and precipitation processes within the domain. The method to determine the lowest model top elevation is applied to both the idealized simulations and the real terrain case study. Notable differences between the ICAR and WRF simulations are observed across all investigated quantities such as the wind field, water vapor and hydrometeor distributions, and the distribution of precipitation. The case study indicates that the precipitation maximum calculated by the ICAR simulation employing the developed recommendations is spatially shifted upwind in comparison to an unmodified version of ICAR. The cause for the shift is found in influences of the model top on cloud formation and precipitation processes in the ICAR simulations. Furthermore, the results show that when model skill is evaluated from statistical metrics based on comparisons to surface observations only, such an analysis may not reflect the skill of the model in capturing atmospheric processes like gravity waves and cloud formation.

The coarse grid spacing of global circulation models necessitates the application of downscaling techniques to investigate the local impact of a changing global climate. Difficulties arise for data-sparse regions in complex topography, as they are computationally demanding for dynamic downscaling and often not suitable for statistical downscaling due to the lack of high-quality observational data. The Intermediate Complexity Atmospheric Research (ICAR) model is a physics-based model that can be applied without relying on measurements for training and is computationally more efficient than dynamic downscaling models. This study presents the first in-depth evaluation of multiyear precipitation time series generated with ICAR on a 4×4 km² grid for the South Island of New Zealand for an 11-year period, ranging from 2007 to 2017. It focuses on complex topography and evaluates ICAR at 16 weather stations, 11 of which are situated in the Southern Alps between 700 and $\mathrm{2150}\mathrm{m}\mathrm{m}.\mathrm{s}.\mathrm{l}$ (m m.s.l refers to meters above mean sea level). ICAR is assessed with standard skill scores, and the effect of model top elevation, topography, season, atmospheric background state and synoptic weather patterns on these scores are investigated. The results show a strong dependence of ICAR skill on the choice of the model top elevation, with the highest scores obtained for 4 km above topography. Furthermore, ICAR is found to provide added value over its ERA-Interim reanalysis forcing data set for alpine weather stations, improving the median of mean squared errors (MSEs) by 30 % and up to 53 %. It performs similarly during all seasons with a MSE minimum during winter, while flow linearity and atmospheric stability are found to increase skill scores. ICAR scores are highest during weather patterns associated with flow perpendicular to the Southern Alps and lowest for flow parallel to the alpine range. While measured precipitation is underestimated by ICAR, these results show the skill of ICAR in a real-world application, and may be improved upon by further observational calibration or bias correction techniques. Based on these findings ICAR shows the potential to generate downscaled fields for long-term impact studies in data-sparse regions with complex topography.

This study explores the potential of different predictor strategies for improving the performance of regression-based downscaling approaches. The investigated local-scale target variables are precipitation, air temperature, wind speed, relative humidity, and global radiation, all at a daily time scale. Observations of these target variables are assessed from three sites in close proximity to mountain glaciers: 1) the Vernagtbach station in the European Alps, 2) the Artesonraju measuring site in the tropical South American Andes, and 3) the Mount Brewster measuring site in the Southern Alps of New Zealand. The large-scale dataset being evaluated is the ERA-Interim dataset. In the downscaling procedure, particular emphasis is put on developing efficient yet not overfit models from the limited information in the temporally short (typically a few years) observational records of the high mountain sites. For direct (univariate) predictors, optimum scale analysis turns out to be a powerful means to improve the forecast skill without the need to increase the downscaling model complexity. Yet the traditional (multivariate) predictor sets show generally higher skill than the direct predictors for all variables, sites, and days of the year. Only in the case of large sampling uncertainty (identified here to particularly affect observed precipitation) is the use of univariate predictor options justified. Overall, the authors find a range in forecast skill among the different predictor options applied in the literature up to 0.5 (where 0 indicates no skill, and 1 represents perfect skill). This highlights that a sophisticated predictor selection (as presented in this study) is essential in the development of realistic, local-scale scenarios by means of downscaling.