Author(s): Sk Das; Awm Ng; Bjc Perera
Linked Author(s):
Keywords: Baseflow separation techniques; Baseflow conditions and runoff events; Non-point source pollution; SWAT; LOADEST
Abstract: Understanding how sediment and nutrient concentrations vary with landuse is critical to study the current and future impacts of landuse changes on water quality. Complex mechanistic models are used to quantify generation and movement of pollutants both in baseflow and surface runoff processes. However, these models require extensive data inputs, high computational power and expertise. Alternatively, specific flow regimes (i. e. baseflow conditions and runoff events) can be used to provide an insight into pollutant sources and to study their relationships with landuses, where limited time (for complex analysis) and limited data are available. This paper presents a simple procedure on how to use specific flow regimes to correlate pollutant load with landuse activities. The procedure uses a simple baseflow separation technique and a data-based model LOADEST with other readily available data and tools. Two simple baseflow separation software Base Flow Index (BFI) and Baseflow Filter Program (BFP) were evaluated comparing them with a complex model SWAT to investigate the efficiency of a simple baseflow separation technique. The baseflow separation technique was then used to define each day of the study period with respect to a specific flow regime using the following conditions; (i) baseflow conditions were defined as those days when baseflow contributes ≥70% of daily streamflow, and (ii) runoff events were defined as those days when baseflow contributes <70% of daily streamflow. Accordingly water quality grab sample data (or estimated load) on a baseflow condition day was tagged as baseflow condition data (or load). Similarly water quality grab sample data (or estimated load) on a runoff event day was tagged as runoff event data (or load). The regression model LOADEST was used to estimate pollutant loads from water quality grab samples. The methodology was illustrated using 12 stations in the Yarra River catchment of Melbourne, Australia for the period of 1998-2009. Furthermore, two specific tasks were performed: (1) to establish whether there are any statistically significant differences between concentrations in water quality grab samples of Total nitrogen (TN), Total phosphorus (TP) and Total suspended solids (TSS) representing specific flow regimes, and (2) to identify the major sources of pollutant loads. The simple baseflow separation techniques compared well with the SWAT simulated baseflow. The coefficient of determination (R2) between BFP generated and SWAT simulated baseflow were 0. 73, 0. 86 and 0. 91 for daily, monthly and annual baseflow respectively. Similarly, R2 between BFI generated and SWAT simulated baseflow were 0. 57, 0. 78 and 0. 90 for daily, monthly and annual baseflow respectively. Based on this comparison, the BFP software was then used for the two specific tasks outlined earlier. The results showed that at the selected water quality stations, the pollutant concentrations were significantly (p≤0. 01) greater in runoff events than in baseflow conditions. Overall, the water quality and pollutant concentrations were influenced by flow regimes and landuses. Major pollutant loads were generated from agricultural and urban runoff i. e. non-point sources. This suggests that Yarra River catchment managers should focus on the management of potential non-point sources to improve river water quality.
Year: 2013