Improving gully density maps for modelling water quality within Great Barrier Reef Catchments

Abstract

Fine-sediment has been shown to have a detrimental impact on water quality across the Great Barrier Reef (GBR). Gully erosion is one of the dominant sources of fine-sediment loads to the GBR, in particular from the Burdekin and Fitzroy basins. Significant funding has been allocated across the GBR catchments to reduce sediment erosion from gullies. Modelling the relative fine-sediment contribution from gully erosion, and assessing the potential water quality improvement due to investment in gully remediation projects within GBR catchments, is reliant on accurate maps of gully density. Hence, techniques that improve mapping of gully density are essential for improving the quality of inputs into catchment models, and will ultimately lead to improved modelled load estimates and better representation of the impact of remediation programs on sediment loads that are delivered to the GBR. Previous attempts to map gully density within the GBR catchments have been conducted by either intensively mapping gully erosion for relatively small isolated areas where gullies are prominent, or by defining the extent of gully erosion at a number of sample sites and then using predictive models to estimate gully density across much larger areas. Due to scale limitations, low accuracy or limited geographic extent, both these approaches have produced maps with limited usefulness for modelling water quality improvements. Consequently there is a need for a methodology that can improve the confidence in gully density maps over broad areas, in a timely fashion, and at a spatial scale that enables the modelling of water quality improvements due to on-ground investments, and allows prioritising of remediation strategies in the GBR. This paper outlines a repeatable process that allows an operator to map the presence or absence of gully erosion within a grid cell, using custom-built geographic information system (GIS) tools, aerial photography and uniform grids. Initially, a number of catchments in the Fitzroy and Burdekin basins were mapped using this approach to improve baseline model inputs for gully erosion. Over the past three years 125 000 km2 have been mapped using this grid-based presence mapping (GBPM) approach. Stage two of the work combined the grid-based mapping with a range of landscape attributes such as slope, distance-to-stream and soil erodibility to produce a predictive model that has the ability to generate gully density maps for all GBR catchments. Finally, the maps from both processes were compared against ground-based observations and previously published estimates of gully density, to determine if these approaches delivered improved inputs for catchment models. Comparison of grid-based gully erosion mapping against ground-based observations show that the grid-based mapping improves the accuracy of the maps compared with previous mapping approaches. The grid-based mapping method developed in this project provides an effective way of capturing gully density data across broad areas. Combining grid-based mapping with predictive modelling enables the acceleration of gully-density mapping.