Challenges in Quantifying Losses in a Partly Urbanised Catchment: A South Australian Case Study

Abstract

Quantifying hydrological losses in a catchment is crucial for developing an effective flood forecasting system and estimating design floods. This can be a complicated and challenging task when the catchment is urbanised as the interaction of pervious and impervious (both directly connected and indirectly connected) areas makes responses to rainfall hard to predict. This paper presents the challenges faced in estimating initial losses (IL) and proportional losses (PL) of the partly urbanised Brownhill Creek catchment in South Australia. The loss components were calculated for 57 runoff generating rainfall events using the non-parametric IL-PL method and parametric method based on two runoff routing models, Runoff Routing Burroughs (RORB) and Rainfall-Runoff Routing (RRR). The analysis showed that the RORB model provided the most representative median IL and PL for the rural portion of the study area as 9 mm and 0.81, respectively. However, none of the methods can provide a reliable loss value for the urban portion because there is no runoff contribution from unconnected areas for each event. However, the estimated non-parametric IL of 1.37 mm can be considered as IL of EIA of the urban portion. Several challenges were identified in the loss estimation process, mainly when selecting appropriate storm events, collecting data with the available temporal resolution, extracting baseflow, and determining the main-stream transmission losses, which reduced the urban flow by 5.7%. The effect of hydrograph shape in non-parametric loss estimation and how combined runoff from the effective impervious area and unconnected (combined indirectly connected impervious and pervious) areas affects the loss estimation process using the RORB and RRR models are further discussed. We also demonstrate the importance of identifying the catchment specific conditions appropriately when quantifying baseflow and runoff of selected events for loss estimation.