Data-driven scale extrapolation: estimating yearly discharge for a large region by small sub-basins

摘要

Large-scale hydrological models and land surface models are so far the onlytools for assessing current and future water resources. Those modelsestimate discharge with large uncertainties, due to the complex interactionbetween climate and hydrology, the limited availability and quality of data,as well as model uncertainties. A new purely data-driven scale-extrapolationmethod to estimate discharge for a large region solely from selected smallsub-basins, which are typically 1–2 orders of magnitude smaller than thelarge region, is proposed. Those small sub-basins contain sufficientinformation, not only on climate and land surface, but also on hydrologicalcharacteristics for the large basin. In the Baltic Sea drainage basin, bestdischarge estimation for the gauged area was achieved with sub-basins thatcover 5% of the gauged area. There exist multiple sets of sub-basinswhose climate and hydrology resemble those of the gauged area equally well.Those multiple sets estimate annual discharge for the gauged areaconsistently well with 6 % average error. The scale-extrapolation methodis completely data-driven; therefore it does not force any modelling errorinto the prediction. The multiple predictions are expected to bracket theinherent variations and uncertainties of the climate and hydrology of thebasin.