Modelling the impact of landscape connectivity on catchment water balance and groundwater response

Abstract

Listen

Translate article

Terrain analysis based on digital elevation models is being routinely used in hydrological modelling. However, landscape connectivity enabling the routing of flow and nutrients from upslope landscape units to adjacent downslope landscape units within a sub watershed is not commonly incorporated into catchment scale models. This paper describes a process of generating connected landscape units within sub watersheds based on a threshold area and evaluates the impact of landscape connectivity on catchment water balance and groundwater response. Digital elevation models (DEMs) are commonly used for the automatic delineation of flow paths, sub watersheds and flow networks for hydrologic modelling. The capacity to define a flow path network describing how flow is routed to a drainage feature is fundamental to distributed hydrologic models. There is a variety of approaches for delineating flow networks using different flow direction algorithms, for example drainage to a single neighbouring cell or the partitioning of flow between multiple neighbouring cells. The resultant flow network underpins the watershed delineation based on upstream drainage area. This paper comments on each approach for representing the networks of rivers and streams and describes an approach using the watershed delineation to define landscape units using elevation intersects and flow paths connecting boundary nodes to drainage lines. This approach results in the generation of connected landscape units of variable size. Catchment scale models typically adopt an aggregated or lumped spatial unit within which land use, soil and climate are proportionally assigned with limited spatial reference. This paper reports on the integration of connected topographic landscape units into a catchment modelling framework (Catchment Analysis Tool, CAT) with application to the Loddon catchment in central Victoria. Comparative results derived using the CAT under historical climate conditions show a 12% improvement in streamflow prediction compared to observed when accounting for landscape connectivity relative to the lumped approach. Comparative results also show significant variation in the recharge patterning and up to 30% variation in recharge rates depending on landscape position as estimated when using landscape connectivity relative to the lumped approach. Incorporating landscape connectivity into a catchment modelling framework is shown to improve the predictive capacity of catchment models to estimate streamflow and groundwater recharge with associated improvement in water resource evaluation and flow and transport modelling