Assessment of the economic feasibility of blue carbon projects in Australia

Brief description

This data collection presents the results of a continental scale assessment of the economic feasibility of blue carbon projects in Australia. A land use trade-offs approach is employed to estimate the economic profitability of current agriculture and blue carbon projects at locations identified as having the potential to support either mangrove, saltmarsh or salt flat natural systems. Where blue carbon economic returns are greater than estimated current agricultural economic returns then a project is considered economically feasible.
The specific projects in scope are those that re-introduce tidal flow to coastal ecosystems and lead to net abatement of carbon emissions, that is, projects that are relevant to the Tidal Restoration of Blue Carbon Ecosystems method of the Australian Government Clean Energy Regulator.
At each identified location estimates of abatement were calculated using the BlueCAM model. Carbon abatement was calculated using the BlueCAM method (Lovelock et al. 2022) as described in the BlueCAM technical report (Lovelock et al. 2021), the carbon credits determination (Australian Government Clean Energy Regulator 2022) and the BlueCAM spreadsheet tool (Australian Government Clean Energy Regulator 2021). Our implementation was simplified according to our objective of estimating abatement per se rather than formally calculating carbon credits. For example, abatement was calculated on a one-year time step rather than reporting periods and project-level emissions (such as fuel use) were not included. Total abatement is calculated over 25 years following tidal introduction.
The net present value of blue carbon projects was estimated for combinations of carbon price, costs, and discount rates. Scenario settings:
•\tcp -> carbon price [$35, $50, $65, $80]
•\tec -> establishment costs per hectare [$1000, $2000, $4000, $6000, $8000 and $10,000]
•\tac -> 5-yearly costs per hectare [$100, $200, $300]
•\tdr -> discount rate for net present value [4%, 7%]
Modelled economically feasible abatement and area for carbon estimation areas under four carbon prices are presented aggregated by Statistical Area 2 (2011) and Local Government Areas (2023), which were clipped to 100km from the coast, by primary sediment compartments (McPherson et al., 2015) and by state/territory. Results are grouped by combinations of scenario settings and includes modelled maximum potential abatement and area.

Lineage: The approach we take relies on inferences derived from modelling inundation at Highest Astronomical Tide and comparing the inundation with current land uses. If land use within an area where inundation is predicted is classified as agricultural land, we infer that there might be a barrier to tide. Stronger inferences would be facilitated by nationally consistent maps of the locations of barriers to tide, but these do not currently exist for Australia.
The process of generating estimates of abatement followed a sequence of steps. First, a digital elevation model was constructed. Inundation over this was then predicted from modelled predictions of Highest Astronomical Tide (HAT). The land uses classes within areas that the model predicted would be inundated were then identified, and transitions to alternate natural systems were predicted based on elevation. The approach relies on national datasets, which might not be accurate in all places. Data are therefore not appropriate for uses that encompass small spatial extents (e.g. tens of hectares project). Data are instead provided in aggregated spatial units that give an approximation of the potential abatement that might be possible.
To create consistent digital elevation (topography and bathymetry) data for Australia we analysed each jurisdiction (i.e. state and territory) separately but with the same method. After a few trials we determined that a 10 m × 10 m grid was an appropriate resolution. For ease of computation, the entire country was tiled into a 100 km × 100 km grid as defined by Digital Earth Australia (DEA). We used the Australia Albers Projection (AusAlb) with the GDA2020 datum and AVWS vertical datum.
A baseline dataset was created using a CSIRO modified version of Geoscience Australia 250 m × 250 m bathymetric data and the Multi-Error-Removed Improved-Terrain DEM 30 m × 30 m terrestrial data.
Datasets from national, state and regional agencies were acquired, aiming for open and public data as primary sources. Each dataset was assessed for quality and appropriateness. Elevation data were then projected into the AusAlb projection, tiled into the DEA grid and then gridded into 10 m × 10 m cells. Once all data for a jurisdiction was processed into the same coordinate system, spatial resolution and ranking, the data were sequentially combined onto the baseline to create the best data for each cell.
To model HAT, the Australian Hydrographic Office (AHO) provided to CSIRO a tidal plane dataset that is derived from measured data from all historically available water level gauges. Accuracy of the AHO HAT tidal plane was assessed against the recently released Global Extreme Sea Level Analysis v3 (GESLA3) which includes a public release license for 125 tidal gauge data from the Australian Bureau of Meteorology (Haigh et al., 2023). The agreement between the AHO HAT and the HAT calculated from GESLA3 is very good, which is to be expected as the same tidal gauge data is used in both. The AHO HAT tidal plane was utilised in this study. The HAT model does not account for the effects of estuaries attenuating or amplifying tide.
The baseline land subtype was estimated from land use mapping using an approach that combined land use, topographic feature and forest cover attributes from land use datasets. Land use was obtained from the Catchment scale land use of Australia 2020 map (CLUM, resolution 50 m, ABARES 2021). Topographic feature and forest cover were obtained from the Land use of Australia 2015–16 map (National Land Use Map, NLUM, resolution 250 m, ABARES 2022). More specifically, topographic feature and forest cover were obtained from the input layers that accompany the NLUM map. The land use classes within areas that the model predicted would be inundated were then identified, and transitions to alternate natural systems were predicted based on elevation.
See Vanderklift et al. (2024) for further details.