On the relationship between the variability of catchment hydroclimate and physiography, and the uncertainty of runoff generation hypotheses