Runoff estimation in ungauged catchments is probably one of the most basicand oldest tasks of hydrologists. This long-standing issue has receivedincreased attention recently due to the PUB (Prediction in Ungauged Basins)initiative. Given the challenges of predicting runoff for ungaugedcatchments one might argue that the best course of action is to take a fewrunoff measurements. In this study we explored how implementing such aprocedure might support predictions in an ungauged basin. We used a numberof monitored Swedish catchments as hypothetical ungauged basins where wepretended to start with no runoff data and then added different sub-sets ofthe available data to constrain a simple catchment model. These sub-setsconsisted of a limited number of single runoff measurements; in other wordsthese data represent what could be measured with limited efforts in anungauged basin. We used a Monte Carlo approach and predicted runoff as aweighted ensemble mean of simulations using acceptable parameter sets. Wefound that the ensemble prediction clearly outperformed the predictionsusing single parameter sets and that surprisingly little runoff data wasnecessary to identify model parameterizations that provided good results forthe "ungauged" test periods. These results indicated that a few runoffmeasurements can contain much of the information content of continuousrunoff time series. However, the study also indicated that results maydiffer significantly between catchments and also depend on the days chosenfor taking the measurements.
展开▼
