Investigation of the potential use of the root-zone storage concept in hydrological modelling under South African conditions.

Abstract

Hydrological models are currently an accepted method used in determining the impacts of Streamflow reduction activities (SFRA) in South Africa. However, the limited availability of soils and rooting depth data create high uncertainty within hydrological modelling exercises. Following poor simulations of streamflow, evaporation and soil water by the ACRU model at Two Streams and Cathedral Peak Catchment VI, the root- zone storage capacity was calculated for both catchments using three internationally published over the period 2007 to 2013 and 2014 to 2018, respectively. The input and calibration data used in the running of the ACRU model was undertaken using observed data commonly available for research catchments in South Africa. Additional data that was available for these specific catchments (observed evaporation and soil water at Two Streams and evaporation at Cathedral Peak Catchment VI) were used in the validation of results. The three methods produced similar mean root- zone storage capacities in both catchments but the Nijzink and DiCaSM methods produced the deepest root-zone storage capacity in the summer months. The results of the Nijzink method were the most variable and DiCaSM the least variable in both catchments. The Nijzink method was most sensitive to the actual evaporation in both summer and winter and sensitive to the precipitation in summer. The Wang method most sensitive to precipitation in summer. The DiCaSM method was found to not be sensitive to the rainfall in either season but highly sensitive to the actual evaporation year-round. The root-zone storage concept better reproduced the observed soil water throughout the soil profile at the Two Streams catchment than the ACRU model. The validation of the root- zone storage capacity against observed soil water illustrated that the root zone storage capacity reflects climate conditions rather than the soil depth and is independent of vegetation, soils and rooting characteristics. This study found that traditional methods of estimating the actual evaporation does not always capture the variability in timing and magnitude of evaporation. The most significant finding is that simple climate driven water balance routine could provide a better representation of soil water than a complex, layered model under South African conditions. The root-zone storage capacity could be a valuable tool in the improvement of hydrological modelling and fundamental in improving the precision of SFRA assessments in South Africa.