Composite Topography-Bathymetry Digital Elevation Model for Queensland Estuaries

This dataset provides a high-resolution (5 m) composite topography-bathymetry Digital Elevation Model (DEM) for all estuaries along the Queensland coast from Cape York to Gold Coast. The DEM was specifically developed to support hydrodynamic modelling of tidal inundation in Great Barrier Reef (GBR) coastal wetlands and estuaries, and served as the foundational elevation dataset for the SCHISM (Semi-implicit Cross-scale Hydroscience Integrated System Model) hydrodynamic model simulations of Highest Astronomical Tide (HAT) published at https://data.csiro.au/collection/csiro:66491v4.\n\nThe composite DEM integrates multiple authoritative elevation data sources across the land-water interface, including terrestrial LiDAR, bathymetric surveys, satellite-derived intertidal topography, and modelled tidal creek bathymetry. A key innovation is the use of satellite-derived water observations and allometric relationships to estimate tidal creek depths based on local creek width, enabling representation of tidal conveyance into coastal floodplains that are not captured by traditional survey methods.\n\nThe DEM is referenced to the Australian Vertical Working Surface (AVWS, EPSG:9458) quasi-geoid vertical datum and projected in Australian Albers Equal Area (EPSG:9473) coordinate reference system. The dataset covers the entire Queensland coastal zone with elevations ranging from approximately -60 m (offshore) to +15 m (terrestrial), with a mask layer indicating the data source for each pixel.\n\nSpatial Coverage: Queensland coast from Cape York to Gold Coast.\n\nTemporal Coverage: Data sources span from 2010 to 2024, with the composite DEM representing conditions circa 2022-2024.\n\nResolution: 5 m horizontal resolution, processed in 5000 m × 5000 m tiles in Australian Albers Equal Area projection (EPSG:9473). Composed into a single Cloud-Optimised GeoTIFF with ZSTD compression.\n\nVertical Datum: quasi-geoidal Australian Vertical Working Surface (AVWS, EPSG:9458). AVWS can be translated to AHD using GA published grids in PROJ. Coastal Mean Sea Level relative to AVWS has been previously published here: https://doi.org/10.25919/6672-jx11\n\nFormat: Cloud-Optimized GeoTIFFs (COG) with (1) elevation (CSIRO_TOPOBATHY_V8_5m_DEM.tif), (2) data source mask (CSIRO_TOPOBATHY_V8_5m_FLAGS.tif).\n\n\nLineage: The composite DEM integrates six authoritative data sources in hierarchical order:\n\n1. Airborne LiDAR Point Clouds (ELVIS): High-resolution terrestrial and nearshore topography from Geoscience Australia's Elevation and Depth - Foundation Spatial Data (https://elevation.fsdf.org.au/). LiDAR data are filtered to ground classification (class 2) and processed using 25th percentile statistics to capture bare-earth elevation while minimizing vegetation effects. Coverage includes most developed coastal areas and major estuaries across Queensland. Horizontal accuracy: ±0.3 m; Vertical accuracy: ±0.15 m (RMSE).\n\n2. Multibeam Bathymetry: High-resolution bathymetric surveys collected during the project, accessed via the CSIRO Data Access Portal catalogue (https://doi.org/10.25919/ffjf-0r37). These data provide detailed seafloor topography in surveyed regions. Typical accuracy: ±0.1 m horizontal, ±0.1-0.15 m vertical.\n\n3. Single-beam Bathymetry: Additional bathymetric soundings from historical hydrographic surveys and scientific campaigns undertaken during the eReefs project. Typical accuracy: ±1-5 m horizontal, ±0.2-1.0 m vertical.\n\n4. DEA Intertidal Elevation Model (ITEM): Satellite-derived intertidal zone topography from Digital Earth Australia (https://doi.org/10.26186/149403). ITEM uses multi-decadal tidal modelling and Landsat/Sentinel-2 imagery to derive elevation at the land-water interface. Horizontal resolution: 10 m; Vertical accuracy: ±0.25 m (RMSE).\n\n5. Modelled Tidal Creek Bathymetry: Estimated creek depths derived from satellite water observations and allometric power-law relationships between creek width and depth. This novel approach fills critical data gaps in narrow tidal channels essential for tidal conveyance modelling.\n\n6. GA AusBathyTopo Grids: Geoscience Australia's 30 m resolution regional depth model of the Great Barrier Reef (http://dx.doi.org/10.4225/25/5a207b36022d2) and 250 m resolution national bathymetry and topography grid (https://dx.doi.org/10.26186/150050). These datasets provide base elevation for offshore regions and areas without higher resolution data. Vertical accuracy: ±5-20 m depending on source data.\n\nThe composite DEM was generated using the topobathy-modelling Python package developed by CSIRO, implementing a seven-stage processing pipeline: (1) Tile Definition - A regular grid of 1000×1000 pixel tiles at 5 m resolution defined in Australian Albers Equal Area projection (EPSG:9473), filtered by elevation range (-60 m to +15 m) and Queensland boundaries. (2) LiDAR Processing - Point clouds accessed from AWS S3, reprojected from native MGA2020+AHD to EPSG:9473+9458, filtered to ground returns, aggregated using 25th percentile statistics, and rasterized to 5 m grid. (3) Creek Mapping - Sentinel-2 imagery loaded from Digital Earth Australia to calculate NDWI, water masks thresholded and skeletonized to extract creek centerlines, depths estimated using allometric power-law (depth = 10^(0.62×log10(width)-1.1)), filtered by distance (5-750 m) and slope (Survey>ITEM>Creek>Base), source mask generated (0=interpolated, 1=LiDAR, 2=multibeam, 3=single-beam, 4=ITEM, 5=creek, 6=base water, 7=base land), gaps filled via linear interpolation, optional Gaussian smoothing (σ=1.5) applied to water surfaces. (7) Output - Composites cropped to tile extents, converted to Cloud-Optimized GeoTIFF with 256×256 blocks and deflate compression.\n\nAll data transformed to Australian Vertical Working Surface (AVWS, EPSG:9458) using pyproj with PROJ network-enabled transformations and AUSGeoid2020 grids. Processing executed on CSIRO Earth Analytics Science & Innovation Platform.