Correcting bias in rainfall inputs to a semi-distributed hydrological model using downstream flow simulation errors

Abstract

The focus of this study is on bias correcting semi-distributed rainfall inputs into a hydrological model applied in the Okavango River basin in southern Africa, where there are very few local observations and heavy reliance is placed on global rainfall datasets. While the hydrological model, before rainfall bias correction, is able to represent the broad characteristics of the sub-basin streamflow responses, as demonstrated by good agreement between observed and simulated flow duration curves, there are many years where the annual volumes are over-or underestimated. The long records of observed flow at downstream stations are successfully used to bias correct the rainfall inputs to the upstream sub-basins using an analysis of their individual contributions to downstream flow and their annual rainfall-runoff response ratios. The results show improved simulations for the relatively shorter observation periods at the upstream gauging stations.

Assessing development and climate variability impacts on water resources in the Zambezi River basin: Initial model calibration, uncertainty issues and performance

“…This appears to be particularly true in the northern parts of the Lake Malawi sub-system, as well as over the Angolan parts of the basin. Previous experience with the CRU data in the Okavango River basin ( Hughes et al, 2011b ; Hughes and Gray, 2017 ) suggest that the rainfall data inputs to the western parts of the basin in Angola and parts of Zambia are particularly problematic, especially over the last few decades due to a reduction in the number of real rainfall observations. Fig.…”

Flood Inundation Mapping of the Sparsely Gauged Large-Scale Brahmaputra Basin Using Remote Sensing Products

“…The TRMM rainfall was not corrected for the other basins for which no gauge rainfall was available. Hughes and Gray [51] presented an interesting approach of correcting TRMM rainfall using hydrological information at the outlet of the basin where no gauge rainfall is available. However, such an approach requires further investigation as Brahmaputra is a much larger catchment compared to the Okavango River basin considered by the authors.…”