Linking regional climate simulations and hydrologic models for climate-change impact studies: a data processing framework

Abstract

Extended Abstract Quantifying the effects of future changes of climatic variables such as precipitation and temperature is a key challenge in assessing the impact of climate change on hydrological systems and their ecological status. A common method to estimate future climate-change impacts on hydrology involves climate variables such as temperature and precipitation from global or regional climate models in combination with hydrological models (Teutschbein and Seibert, 2012). For watershed-based studies the resolution of global (approximately 100 to 250 km) and regional (approximately 25 to 50 km) climate models is, however, not appropriate for resolving climate features at the catchment scale (Salathe, 2003). Although regional climate models provide data closer to the watershed scale, simulations of present conditions do not often agree with climate observations. This means that such regional climate model simulations do not always accurately simulate daily precipitation or temperature within a watershed and have often shown significant biases (e.g., Varis.2014). According to Teutschbein and Seibert (2012) the reason for such biases can be related to systematic model errors caused by discretization and spatial averaging within grid cells. This means that the local climate within a watershed may be influenced by topographic features and thus the output of a climate model grid cell does not necessarily represent a climate gauge within a watershed. Thus, these simulations might not be directly useful for assessing hydrological impacts at the watershed scale (Teutschbein and Seibert, 2012). One recommendation is to use bias correction methods. These methods use a transformation algorithm to adjust simulated climate data by identifying possible biases between observed and simulated climate variables. The computed biases are taken for correcting both present and future climate model runs. This means that bias correction methods are assumed to be stationary and that the correction parameters for present climate conditions are considered to be valid for future conditions.