Water resources management decisions are made based on information from predictive models that are capable of simulating the behaviour of hydrological systems. More of these models are in use today and it is becoming increasingly difficult to choose which model to use for particular space and time scales as well as climate. In addition, as a result of climate change there is an increase in the degree of randomness in hydrological systems leading to reduced predictability of these systems and thus different models are prone to perform differently under varying conditions. In this study, meta-analysis was conducted involving seven commonly applied models in hydrological assessment to try and establish patterns that these models exhibit under varying circumstances. This involved looking at the homogeneity of the studies at the various space and time steps. In addition to meta-analysis, a second stage of analysis looking at the variation in performance of the models with catchment characteristics such as climate, mean altitude and catchment size was assessed. The utility of the Nash and Sutcliffe (NSE) model efficiency criteria reported for the case studies was used in categorization of the studies. Performance of Soil and Water Assessment Tool (SWAT), Hydrologiska Byrans Vattenbalansavdelning (HBV), Variable Infiltration Capacity (VIC), and Precipitation-Runoff Modelling System (PRMS) improves with increasing catchment size, while Hydrological Simulation Program-Fortran (HSPF), Systeme Hydrologique Europeen (MIKE-SHE) and Hydrologic Engineering Center-Hydrologic Modelling System (HEC-HMS) performance is insensitive to catchment size. HSPF, MIKE-SHE and PRMS improves in performance at higher altitudes, while HEC-HMS shows declining performance with increase in altitude. Whereas, the performance of SWAT and HSPF declines at lower latitudes, that of MIKE-SHE and PRMS improves at lower latitudes. Results from such review studies are important in aiding decision making regarding model selection, especially in a situation where one is faced with a declining hydrological measurement network in the face of climate change.
Keywords: Hydrological models, heterogeneity, confidence interval, altitude, latitude.