Validation of a 9-km WRF dynamical downscaling of temperature and precipitation for the period 1980–2005 over Central South Chile

• Published in: Theoretical and applied climatology Vol. 143; no. 1-2; pp. 361 - 378
• Main Authors: Fernández, Alfonso, Schumacher, Vanúcia, Ciocca, Isabella, Rifo, Andreaw, Muñoz, Ariel A., Justino, Flavio
• Format: Journal Article
• Language: English
• Published: Vienna Springer Vienna 01.01.2021; Springer; Springer Nature B.V
• Subjects: Analysis; Aquatic Pollution; Atmospheric Protection/Air Quality Control/Air Pollution; Atmospheric Sciences; Bias; Cell size; Chile; Climate; Climate science; Climatic changes; Climatic conditions; Climatology; Clustering; Confidence; data collection; Earth and Environmental Science; Earth Sciences; Earth system science; latitude; Meteorological research; Numerical weather forecasting; Original Paper; Precipitation; simulation models; Statistical analysis; Temperature; Temporal variations; Waste Water Technology; Water Management; Water Pollution Control; weather; Weather forecasting; Weather stations; Chile; Dynamical downscaling; Central South Chile; Temperature; Precipitation
• DOI: 10.1007/s00704-020-03416-9
• ISSN: 0177-798X

In this paper, we evaluated a dynamical downscaling produced for Central South Chile (32°S–38°S) relative to climatic conditions between 1980 and 2005. Assessing the skill of dynamical downscaling relative to the present climate is key to determine the degree of confidence on regional climatic projections. We used the Weather Research and Forecasting model to simulate that period at ~ 9 km grid-cell size, forced by the bias-corrected Community Earth System Model. Results indicated that the dynamical downscaling adequately reproduced spatio-temporal features of the climate within the region. Temperature showed a positive bias at the annual scale while the opposite occurred for precipitation. The bias varied when the comparison was performed relative to a gridded product or instrumental records from weather stations. At the monthly scale, the model failed to capture long-term trends relative to the gridded dataset while reproducing spatial patterns, especially for temperature. We found a generally statistically significant spatial clustering of the monthly mean bias that can support implementation and application of dynamical downscaling and bias-correction methods that account for the distinct climatic features of the study area. In particular, the strip 34°S–35°S presented features that are coincident with previous findings suggesting this latitude to be a boundary between different climate regimes north and south. According to our results, we assert that this dynamical downscaling is comparable with other available databases and thus can be utilized in future studies as an additional and independent source of analysis, contributing to a balanced appraisal of climate scenarios for policymaking within the region.