Using distributive surface water and groundwater modelling techniques to quantify groundwater recharge and baseflow for the Verlorenvlei estuarine system, west coast, South Africa.

Watson, Andrew. P. (2018-12)

Thesis (PhD)--Stellenbosch University, 2019.

Thesis

ENGLISH ABSTRACT: For effective management of groundwater resources, recharge rates and baseflow volumes need to be quantified to determine sustainable abstraction regimes and to quantify the ecological reserve, the amount of water needed to maintain the natural environment. While a variety of methods have been used to estimate groundwater recharge, estimates vary due to method, temporal and spatial resolution used. In rainfall/runoff modelling where potential recharge is determined by calculating the amount of water that percolates through the unsaturated zone, aquifer components are usually lumped resulting in over or under estimation of recharge. In contrast, groundwater models which include distributive aquifer components are commonly setup to lump climate and surface variables, thereby neglecting seasonal and climatic variability. In this study, a combined rainfall/runoff and groundwater modelling approach was used to determine the net recharge and baseflow in the RAMSAR-listed Verlorenvlei sub-catchment on the west coast of South Africa. This sub-catchment is an important biodiversity hotspot but is also an important agricultural region, hence there is competition for water resources. To understand the water dynamics within the catchment a four-phase approach was taken to determine the baseflow and ecological reserve requirements. This involved firstly, determining the limits of the sub-catchment boundary. Although the Verlorenvlei lake is supported by the Verlorenvlei sub-catchment which is itself fed by four main tributaries (the Hol, Krom Antonies, Kruismans, and Bergvallei), previous research has indicated that only one of these tributaries (the Krom Antonies) played an important role in the delivery of fresh water to the lake system. Initially the catchment boundary was thus modelled on the Krom Antonies tributary, although the understanding gained by the delineation was applied to the entire sub-catchment. To include spatial and temporal variability in groundwater recharge estimates, a rainfall runoff model was used to determine potential recharge using regionalised climate and assumptions regarding aquifer hydraulic conductivity. The potential recharge estimates within the sub-catchment exceeded previous studies (30 % higher), with the daily timestep nature of the J2000 model (Krause, 2001) assumed to account for this difference. To determine whether aquifer hydraulic conductivity could impact groundwater recharge rates, a groundwater model (MODFLOW) was constructed for the main assumed freshwater source of the Verlorenvlei, the Krom Antonies. The groundwater model included distributive aquifer hydraulic conductivity, although the input recharge was lumped which reduced climate seasonality and daily variability. The resultant output from the groundwater model was net recharge (0.3-11.4 % of rainfall) and average baseflow (14, 000 - 19, 000 m3.d-1), with the model suggesting that the baseflow from the Krom Antonies was not enough to meet evaporation demands (90, 000 m3.d- 1) and that there must be another much larger source. To incorporate daily climatic fluctuations and seasonality in baseflow estimates, the groundwater components of the J2000 were distributed using the net recharge and aquifer hydraulic conductivity from the Krom Antonies. By distributing the groundwater components within the J2000 model, the proportion of interflow to recharge was improved allowing for comprehensive estimates of runoff and baseflow from each tributary. While the model was calibrated using streamflow measurements from the gauging structure on the Kruismans, the measurements were particularly hindered by the DT limit (discharge table) of the station (3.675 m3.s-1), which resulted in reduced confidence in modelling high flow events. To incorporate the limited resolution of the station as well as limited length, an Empirical Mode Decomposition (EMD) was applied to the runoff data and water levels measured at the sub-catchment outlet. The results of the model adaption was that the Krom Antonies was not in fact the main freshwater source, with the Bergvallei supplying the majority of groundwater (49 %) as well as a large contributor of streamflow (29%). While the Hol was initially believed to be a minor contributor, the tributary had the largest ratio of baseflow (0.56), which acted to reduce its flow variability. While the Krom Antonies and Bergvallei is comprised of highly conductive sandstones and quaternary sediments, the Hol which is mainly comprised of shales, resulted in a larger groundwater flow attenuation which reduces its susceptibility to drought and climatic variability. The results of this study highlighted that on average the streamflow (20, 500 m3.d-1) from the feeding tributaries was not able to meet the evaporative demand of the Verlorenvlei and that the lake was mainly supplied by low occurrence high flow events. With the Verlorenvlei under threat due to continued agricultural expansion, it is likely that the lake will dry up more frequently in the future, especially if flows are hampered during wet cycles, when ecosystems regenerate.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/105237
This item appears in the following collections: