Statistical downscaling of hydrometeorological variables using general circulation model output

Empirical relationships between daily hydrometeorological variables for a catchment in Nagano prefecture, Japan and three indices of regional atmospheric circulation are examined with a view to assessing their use in General Circulation Model (GCM) downscaling. The indices (vorticity, flow strength and angular direction of airflow) were calculated by using daily grid-point sea-level pressure data derived from: (a) the National Centers for Environmental Prediction, NCEP Re-analysis data set (1979-1995); and (b) the UK Meteorological Office, Hadley Centre coupled ocean-atmosphere model (HadCM2SUL) for two periods indicative of present (1980-1999) and future greenhouse gas plus sulfate aerosol forcing (2080-2099). Statistical models of the surface variables were then "forced" by using the three airflow indices obtained from HadCM2SUL. The differences between the NCEP and HadCM2SUL "present" downscaled variables were generally greater than those arising between the downscaling of the two GCM airflow scenarios. The lack of change in downscaled surface variables between the 1980-1999 and 2080-2099 data was attributed to the low sensitivity of atmospheric circulation patterns in HadCM2SUL to greenhouse gas forcing. (C) 1998 Elsevier Science B.V.