DIRECTLY CONNECTED IMPERVIOUNESS MAPPING FOR MELBOURNE WATER REGION
A growing body of research points towards stormwater as a key disturbing factor in urban streams. This impact is primarily caused by impervious surfaces that are directly connected to streams through drainage systems. For accurate estimation of directly connected imperviousness (DCI), there is a need to map impervious surfaces through a cost-effective and reliable approach and then estimate their connectedness to streams.
Recently marketed, high resolution aerial photos with infra-red (IR) band data provides means for automation of impervious mapping for large areas. Grid-cell based elevation modelling enables derivation of flow parameters to these impervious surfaces. Impervious mapping has been carried out for most of Melbourne Water’s region. This project involved the use of fast look aerial photos with an IR band of 35cm resolution and an elevation model developed using contours. A Normalised Vegetation Index (NDVI) was used to automatically map impervious surfaces using aerial photos followed by a manual verification process. The piped drainage network was used to condition the elevation model and to derive sub-catchments for every piped network where it enters a stream. Overland flow distances from impervious polygons to the nearest pipe and/or stream were determined. An exponential decay function was used to determine the degree of connectedness of an impervious polygon to the stream. Total imperviousness (TI) and DCI were aggregated for each sub-catchment and cumulated down the catchment.
The following GIS data layers were developed as part of this project.
Impervious_areas_mosaic_with_dci_distances.tab: This data set includes the base polygons which define impervious areas across the study area. Each polygon is tagged as either within a parcel i.e. a private allotment (roofs and paving) or outside a parcel and hence a road area. Each polygon contains flow distance measures from the polygon to the nearest pipe, drain and stream which form the basis for the DCI calculation.
All_Pour_Points.tab: This data set constitutes a pour point layer for the region. Through the deployment of catchment modelling techniques and other spatial tools these points were identified at natural waterway confluences and at the intersection of pipe networks and waterways.
All_Catchments_DCI.tab: This data set constitutes a seamless layer of sub-catchments for the region derived through the deployment of catchment modelling techniques using a digital elevation model (DEM). It contains TI and DCI parameters for both individual sub-catchments and for the entire upstream catchment.
DCI_River_Network.tab: This is the prime data set. It contains cumulative DCI (i.e. the proportion of impervious surfaces for the entire upstream catchment estimated to be directly connected to the stormwater drainage network using various methods) in a combined waterways network dataset. A unique SUB_CAT_ID_TXT field connects these data to the All_Catchments_DCI.tab dataset. This waterway network is an assembly of Melbourne Water drainage data ie DR_Natural Waterways, DR_Channels and DR_Underground Pipes.
Macroinvertebrate assemblage composition in streams of the Melbourne region are strongly explained by DCI, either calculated using an exponential decay function to weight distance of impervious surfaces to stormwater drains (half-decay distance of 0-9.4 m), or by including only those impervious surfaces within 10-40m of stormwater drains (Walsh and Kunapo 2009). Using the former method, streams with DCI values of > ~2% are invariably in degraded ecological condition. Disconnection of impervious surfaces throughout catchments to < ~2% is therefore required for achievement of good ecological condition. To assess and achieve disconnection design standards for stormwater runoff and quality are being developed as part of the Little Stringybark Creek pilot project (www.urbanstreams.unimelb.edu.au).
Example of DCI modelling.