Within the climate change community the uncertainty originating from differences in climate model sensitivity is well documented and global climate models are consistently compared using a range of future emission scenarios. For climate change impact analyses usually only a single impact model is used however also the use of impact models results in uncertainty. Within WaterMIP for the first time we systematically compares simulation results of land surface models and global hydrological models for both future and historic time series. Within WaterMIP all models are run at 0.5 degree spatial resolution for global land areas using a standerdized land mask. In the first stage models were run for a 15-year simulation period (1985-1999) using a newly developed global sub-daily meteorological data set (WATCH Forcing Data, WFD). The data are derived from the ERA-40 reanalysis product via sequential interpolation to half-degree resolution, elevation correction and monthly-scale adjustments based on CRU (corrected-temperature, diurnal temperature range, cloud-cover) and GPCC (precipitation) monthly observations combined with new corrections for varying atmospheric aerosol-loading. To simulate the impact of climate change different global hydrological and land surface models were forced with bias corrected output from 3 GCMs using 2 emission scenarios. For this a new statistical bias correction methodology was developed, which can be used for correcting climate model output to produce internally consistent fields that have the same statistical intensity distribution as the observations. Already for the 1985-99 periods There were a considerable spread in simulation results between the different models for example simulated evapotranspiration was between 415 and 586 mm year-1 (61,000 to 86,000 km3), and simulated runoff ranged from 290 to 457 mm year-1 (43,000 to 67,000 km3). For the future period our initial results show that depending on the bias correction and hydrological model used the sign of the climate change input can change. Especially, the method for calculating (potential) evaporation cause large difference between models in the simulated climate change impacts. Our results show that when studying climate change impacts there is also considerable uncertainty caused by the choice of the hydrological model. We suggest that for climate change impact studies not only multiple climate models should be used in but also multiple impact models
|Publication status||Published - 2010|
|Event||AGU Fall Meeting 2010, San Francisco, California, USA - |
Duration: 13 Dec 2010 → 17 Dec 2010
|Conference||AGU Fall Meeting 2010, San Francisco, California, USA|
|Period||13/12/10 → 17/12/10|
Ludwig, F., Haddeland, I., Clark, D., Voss, F., van Vliet, M. T. H., Hagemann, S., & Kabat, P. (2010). Multi-model estimate of the historic and future global water balance: A model intercomparison using multiple global hydrological models and multiple climate models. Abstract from AGU Fall Meeting 2010, San Francisco, California, USA, .