A framework for establishing a distributed, daily-time-step water resources model based on a pre-existing spatially lumped monthly representation: A case study of the Grootdraai Catchment, South Africa

Abstract

Southern Africa Proprietary monthly lumped models are pragmatic tools for water resources management and planning. However, their low spatial granularity and limited transparency pose significant obstacles to effective water quantity and quality management, particularly in developing countries. This study aimed to develop a distributed, daily-time-step model for a developing country catchment based on an existing black-box, lumped, monthly representation. An Open-Source, Python Water Resources (Pywr) model of the Grootdraai Catchment, South Africa was developed based on a pre-existing monthly Water Resources Yield Model (WRYM) representation. Nodes in the Pywr model were established at a finer spatial scale, and return flows were represented individually. Lumped, monthly natural inflows of the WRYM model were disaggregated to daily using an existing method based on daily rainfall. Abstractions and return flows in the WRYM were disaggregated evenly among the days in the month for input into the Pywr representation. Comparisons of monthly simulated WRYM reservoir storage and river flow with the daily simulations by Pywr exhibited a high level of agreement. The proposed framework can considerably reduce the time and resources required to develop spatially distributed models by leveraging existing resources and can guide the cost-effective and rapid transition from monthly-lumped to daily-distributed water resources models. • Many monthly lumped catchment models exist with major limitations for management. • We developed a framework to generate daily-distributed models using a case study. • The correctness of the daily-distributed model was validated. • Can guide the transition to daily-distributed models using existing lumped models.