Multi-Scale Performance Assessment of State-of-the-Art Reference Grids for Precipitation over an Alpine Catchment in Northern Italy.

Abstract:

Precipitation plays a key role in the hydrological cycle and serves as a prominent variable in the meteo-hydrological modeling chain. Accurate assessment of precipitation in this context is of utmost importance to model hydrological quantities. Reference data sets for precipitation are furthermore required to evaluate and bias correct climate model simulations in order to better represent current regional conditions. Uncertainties increase when moving to higher spatial resolution, including small-scale processes, and more complex terrain.

In this study a high resolution reference data set derived from over 150 observation stations is applied to evaluate the performance of state-of-the-art reference data sets over the mountainous Adige catchment (~ 12,000 km²), located in Northern Italy with elevations up to 3800 m. A variety of reference data sets is available at various spatial resolutions ranging from several km to climate model scale at 2°. These data sets stem either directly from observations or re-analysis simulations, with or without assimilated precipitation, or are estimated by remote sensing approaches. Comparison is performed at various spatial resolutions corresponding to those of the applied precipitation data sets. Furthermore, an intercomparison of these is conducted at the corresponding resolutions, where higher resolved data sets are aggregated to the target resolution. This procedure allows for a thorough analysis without penalizing global data sets at coarse resolution.

The main objective of this study hence is to address and quantify uncertainties of these coarse data sets. Additionally, whether they perform acceptable on catchment scale and can be applied for bias correction and model calibration, or if regional high resolution data sets outperform these data sets significantly.