Downscaling of climate indicators to 10x10 m resolution

Climate impact research sometimes needs climatological input data at a much higher spatial resolution than that of typical meteorological data sets (e.g. reanalysis, regional climate models, gridded observational data sets). We contribute climate data to a large interdisciplinary project “FORSITE II”, where forest site classification maps are produced for three provinces of Austria under the influence of climate change. The climate data involves many climate indicators relevant for vegetation with a spatial resolution of 10x10 m, leading to about 400 million grid cells for the research domain. Complex climate indicators such as potential evapotranspiration calculated with the Penman-Monteith equation need temperature, wind speed, humidity, and radiation as input.

Most meteorological variables are highly correlated with elevation, especially on a regional scale. Thus, for the downscaling to 10 m, a local vertical gradient of the climate indicator is calculated. Combining the vertical gradient with the difference of the topography on the fine and coarse grid yields the climate indicator on the high-resolution grid with a clear imprint of the fine topography.

Radiation-based indicators require a more advanced approach: For realistic shading effects of the topography, a solar radiation model is used to calculate potential direct radiation on both the coarse and the 10x10 m topography on each day of the year. Indirect radiation is directly downscaled with the difference of the sky view factors between coarse and fine resolution. These two downscaling methods are then applied to actual daily indirect and direct radiation, which was provided on a 100x100 m grid by GeoSphere Austria.