A physically based distributed karst hydrological model (QMG model-V1.0) for flood simulations

Abstract. Karst trough and valley landforms are prone to flooding, primarily because of the unique hydrogeological features karst landforms, which conducive spread rapid runoff. Hydrological models that represent complicated hydrological processes in regions effective for predicting but their application has been hampered by complex model structures associated parameter set, especially distributed models, require large amounts data. Distributed flooding highly dependent on modelling, boundary settings extensive data processing, consumes both time computational power. Proposed here is a physically based known as QMG (Qingmuguan) model. The structural design this relatively simple, it generally divided into surface underground double-layered structures. parameters functions each layer have clear physical meanings, fewer included than current models. This allows areas be modelled with only small amount necessary Eighteen flood across river Qingmuguan simulated model, values agree well observations: average Nash–Sutcliffe coefficient water balance 0.92, while relative flow process error 10 % peak 11 %. A sensitivity analysis shows infiltration coefficient, permeability rock porosity most attention calibration optimization. improved predictability enabled proposed promotes better mechanistic depiction runoff generation confluence valleys.