Which method was used to bias adjust Global and Regional climate variables

CORDEX and CMIP5 variables are bias adjusted using the Distribution Based Scaling method (DBS; Yang et al., 2010) versus the global reference dataset HydroGFD2.0 (Berg et al., 2018). Both bias adjustment method and global reference dataset are developed by the Swedish Meteorological and Hydrological Institute (SMHI).

The DBS method is a parametric quantile-mapping variant. This type of methods fit a statistical distribution to the cumulative distribution function and use those fitted distributions to conduct the quantile-mapping. For precipitation, the number of wet-days is first corrected by applying a wet-day threshold. Then, the single (double) gamma distribution (see Yang et al., 2010) is fitted to the daily rain data in case there are more than 25 (500) rain days. Below 25 rain days, the mean precipitation intensity is corrected instead of a distributional correction. For temperature, the normal distribution is fitted to the data. Temperature corrections were done conditional on the wet/dry state of the corresponding precipitation time series. The seasonal variations in the biases were represented by monthly parameter windows for precipitation and a moving window of 31 days width for temperature.

There is some post-processing in place for the data to be suitable for hydrological impact modeling.  Bias adjustment of daily mean/maximum/minimum temperature using quantile mapping can in some cases lead to inconsistencies. For e.g., the daily maximum (minimum) temperature could be lower (higher) than the daily mean temperature. If such inconsistencies occur, daily minimum and maximum temperatures are adjusted in such a way that the anomalies with respect to the daily mean temperature are in line with the climatological anomalies for the particular day in the seasonal cycle. This means, for e.g., that an inconsistency occurring on June 25 will be adjusted using the climatological anomalies for June 25, estimated by a moving window. Also, the adjustment is done conditional on the wet/dry state of the corresponding precipitation series. The climatology of the anomalies was derived from the HydroGFD2.0 global reference data set.

The DBS method is a parametric quantile-mapping variant. This type of methods fit a statistical distribution to the cumulative distribution function and use those fitted distributions to conduct the quantile-mapping. For precipitation, the number of wet-days is first corrected by applying a wet-day threshold. Then, the single (double) gamma distribution (see Yang et al., 2010) is fitted to the daily rain data in case there are more than 25 (500) rain days. Below 25 rain days, the mean precipitation intensity is corrected instead of a distributional correction. For temperature, the normal distribution is fitted to the data. Temperature corrections were done conditional on the wet/dry state of the corresponding precipitation time series. The seasonal variations in the biases were represented by monthly parameter windows for precipitation and a moving window of 31 days width for temperature.

There is some post-processing in place for the data to be suitable for hydrological impact modeling.  Bias adjustment of daily mean/maximum/minimum temperature using quantile mapping can in some cases lead to inconsistencies. For e.g., the daily maximum (minimum) temperature could be lower (higher) than the daily mean temperature. If such inconsistencies occur, daily minimum and maximum temperatures are adjusted in such a way that the anomalies with respect to the daily mean temperature are in line with the climatological anomalies for the particular day in the seasonal cycle. This means, for e.g., that an inconsistency occurring on June 25 will be adjusted using the climatological anomalies for June 25, estimated by a moving window. Also, the adjustment is done conditional on the wet/dry state of the corresponding precipitation series. The climatology of the anomalies was derived from the HydroGFD2.0 global reference data set.

References

Berg, P., Donnelly, C., and Gustafsson, D.: Near-real-time adjusted reanalysis forcing data for hydrology, Hydrol. Earth Syst. Sci., 22, 989–1000, https://doi.org/10.5194/hess-22-989-2018, 2018

 

Yang, W., Andréasson, J., Graham, P. L., Rosberg, J. and Wetterhall, F.: Distribution-based scaling to improve usability of regional climate model projections for hydrological climate change impacts studies, Hydrology Research, 41 (3-4): 211–229, https://doi.org/10.2166/nh.2010.004, 2010