National Land Cover Account methodology

The National Land Cover Account is produced by the Australian Bureau of Statistics (ABS) and released in alignment with the Commonwealth government’s Common national approach to environmental-economic accounting in Australia. Experimental estimates published in the Land Cover Account have been developed in collaboration with the Department of Agriculture, Water and the Environment (DAWE) and Geoscience Australia (GA), including GA’s internal research area:  Digital Earth Australia (DEA).

This account is part of a suite of environmental-economic accounts produced by the ABS based on the United Nations System of Environmental-Economic Accounting (SEEA). The SEEA framework extends the boundaries of the System of National Accounts (SNA) framework to include environmental resources, which occur outside economic production boundaries measured by the SNA.

The SEEA Central Framework (SEEA CF) uses a systematic approach to organise environmental and economic information. It covers, as completely as possible, the stocks and flows relevant for analysis of environmental and economic issues. This framework applies accounting concepts, structures, rules and principles of the SNA. Environmental-economic accounts deliver important extensions to SNA accounts. These accounts may include physical supply and use tables, functional accounts (such as environmental expenditure accounts), and asset accounts for natural resources.

For this account the physical asset, change matrix and stock position tables align with the SEEA CF where possible. Where data is unavailable to complete the tables, 'na' has been used to maintain the SEEA account framework. 

• Stock position tables
• Physical asset account tables
• Change matrix tables

• Statistical area level 2 datasets

The SEEA CF definition of land is:

Land is a unique environmental asset that delineates the space in which economic activities and environmental processes take place and within which environmental assets and economic assets are located.

The SEEA CF provides guidance for standardised asset account tables (physical and monetary) and change matrix tables.

The change matrix tables show how characteristics of land transitioned between two time points. The rows of the matrix start with an opening stock position for the first time point and finish with a closing stock position for the second time point. The data in each column represents the net change between the two intersecting categories, that is the movements in and out of these combinations of classes. A positive change represents an overall increase in the land class and a negative change represents an overall decrease in the land class.

In the table below we can see that trees had 80 units in time period 1 (the opening stock) and 110 units in time period 2 (the closing stock). The values in the rows give the movement to the class represented in the column. For instance, we can see that 10 units from grass and 20 units from bare earth transitioned to trees. By summing these values, we get the net change from the opening stock to the closing stock. In this case 30 units transitioned to trees from other classes. Looking at bare earth there was an opening stock of 60 at time period 1 and a closing stock of 45 at time period 2. From the rows we can see that overall 20 units of bare earth in time period 1 transitioned to trees in time period 2 and that 5 units transitioned from grass to bare earth between time period 1 and time period 2. The net change indicates that overall 15 units transitioned out of bare earth to other classes.

The National Land Account data standards and statistical geography was developed to provide consistency in data development approaches and facilitate alignment of input datasets to produce account ready datasets. These steps spatially enable the compilation, analysis and interpretation of the National Land Cover Account. The data standards and statistical geography sections outline dataset specifications and linkages to the geographic boundary for the National Land Account and sub-national reporting areas.

The National Land Account data specifications are as follows:

• Format: raster
• Cell resolution: 250m (also referred to as the Basic Spatial Unit, BSU)
• Coordinate system: GDA94 datum with Australia Albers EPSG3577 projection

                   aulx: -2189542.25149

                   auly: -1047686.305317

                   alrx: 2468707.74851

                   alry: -4964936.305317

Attribution or resampling rule: categorical data by mode numerical data by bilinear rule, nearest neighbour for reprojection and sub setting

Boundary cell rule: included if the raster cell centroid is within the boundary, excluded if the centroid is not within the boundary.

The statistical geography used by the National Land Cover Account is based on the Australian Statistical Geographic Standard (ASGS, 2016 version) – a geographic classification of Australia into a hierarchy of statistical areas. The 2016 ASGS Statistical Area 2 (SA2) boundaries were converted to a 250m raster following the National Land Account data standards. Resampling from the original dataset resolutions to match the coarser resolution of the National Land Account data specification will impact SA2 area totals compared to the area published in the ASGS, 2016 version. 

There are over 2200 SA2 areas defined in the 2016 ASGS. However, SA2s that cover Christmas Island, Cocos (Keeling) Islands, Lord Howe Island and Norfolk Island have been excluded from the SA2 raster and are not included in the National Land Cover Account. These SA2s are not within the geographic coverage of the DEA land cover spatial product. SA2s are designed to reflect functional areas that represent a community that interacts together socially and economically. They consider suburb and locality boundaries to improve the geographic coding of data to these areas, and in major urban areas, SA2s often reflect one or more related suburbs. The SA2 is the smallest area for the release of many ABS statistics.

As the ASGS is a hierarchical system SA2 level data can be aggregated to broader geographic regions within the ABS geographic structures, such as states and territories. Accordingly, national, state and territory figures in the National Land Cover Account have been compiled by aggregating the data from SA2s included in the raster product. For the purposes of aggregating to states the Jervis Bay SA2 is included in New South Wales figures.

Digital Earth Australia (DEA) is a program area within GA, Australia’s public sector geoscience organisation.

Land cover physical asset tables are based on the DEA land cover spatial products. The DEA land cover spatial products provide consistent, continental, annual land cover classifications for Australia. The DEA platform uses spatial data and images recorded by satellites orbiting our planet to detect physical changes across Australia. DEA land cover is released as annual products where each product is for an observation period from April in one year to March in the following year. For the purposes of the National Land Account, the year 1988 is used to describe values from the product that covers observations from April 1988 to March 1989 and the year 2020 is used to describe values from the product that covers observations from April 2020 to March 2021. Each year’s product is released on a 25m resolution grid with each grid cell assigned a land cover classification that best describes the land cover of that cell over the 12-month observation period. The classification is based on the Food and Agriculture Organisation (FAO) Land Cover Classification System (LCCS, version 2), a globally accepted land cover classification standard. The DEA land cover products were made account-ready by resampling to a 250m resolution consistent with the National Land Account geographic data standards. No edits are made to values in the source DEA data beyond aggregating to ABS Land Cover classes in the compilation process.

SA2 level stock positions for a year are obtained by combining the DEA land cover product (raster) for a given year and the rasterised version of the 2016 ASGS SA2s. The two rasters are cross tabulated, and this gives the count of Basic Spatial Units (BSU) that belong to each combination of SA2 and LCCS class.  The counts are multiplied by 6.25 to give the area covered in HA. LCCS classes are concorded to the ABS Land Cover Classifications. The SA2 stock position for an ABS land cover class is the sum of area belonging to the class in the SA2.

The State/Territory stock positions are created by aggregating the SA2 stock positions according to the state/territory the SA2 falls within. The national stock position is created by aggregating all SA2 stock positions.

As there is only a single land cover observation for each BSU within the DEA Land Cover product, the yearly stock position reflects the land cover classification that best describes the land cover of the aggregated BSUs over the April to March observation period rather than the cover at any given time.

The physical account and change matrices are built from the counts of BSUs belonging to a combination of LCCS classes in opening and closing years within each SA2. For each account period three rasters are combined. The DEA land cover raster for the opening year, the DEA land cover raster for the closing year and the rasterised 2016 ASGS SA2s. The three rasters are cross tabulated returning the count of BSUs identified as a unique combination of the LCCS classification in the opening year, LCCS classification in the closing year, and SA2 classification. These counts are multiplied by 6.25 to give the area in HA. LCCS classes in the opening and closing year are concorded to the ABS Land Cover Classification and aggregated resulting in the count of BSUs belonging to a unique combination of ABS Land Cover class in the opening and closing years within each SA2.  These counts are also aggregated to the State/territory level and national level.

The physical account for a given region (i.e. an SA2, State/territory, or national) is compiled according to the following steps:

• Summing the area covered by each ABS Land Cover class in the opening year to obtain the opening stock
• Summing the area covered by each ABS Land Cover class in the closing year to obtain the closing stock
• Identifying class combinations where the ABS Land Cover class in the opening year is different to the ABS Land Cover class in the closing year
• Identifying class combinations where there has been a change in ABS Land Cover class and one of the ABS Land Cover classes is either ‘Artificial Surfaces’ or ‘Cultivated Terrestrial: Herbaceous’ and categorising these as a ‘managed change’. All other combinations are categorised as ‘other change’.
• Summing the area covered by each ABS Land Cover class in the closing year where the observation has been identified as featuring a change in ABS Land Cover classification and the observation has been identified as a ‘managed change’ to obtain ‘Other expansion to stock’ for each ABS Land Cover class.
• Summing the area covered by each ABS Land Cover class in the closing year where the observation has been identified as featuring a change in ABS Land Cover classification and the observation has been identified as an ‘other change’ to obtain ‘Other additions to stock’ for each ABS Land Cover class.
• Summing the managed additions to stock and other additions to stock for each ABS Land Cover class to obtain the total additions for each ABS Land Cover class.
• Summing the area covered by each ABS Land Cover class in the opening year where the observation has been identified as featuring a change in ABS Land Cover classification and the observation has been identified as a ‘managed change’ to obtain ‘Managed regression to stock’ for each ABS Land Cover class.
• Summing the area covered by each ABS Land Cover class in the opening year where the observation has been identified as featuring a change in ABS Land Cover classification and the observation has been identified as an ‘other change’ to obtain ‘Other regression to stock’ for each ABS Land Cover class.
• Summing the managed reductions to stock and other reductions to stock for each ABS Land Cover class to obtain the total reductions for each ABS Land Cover class.
• Summing the area covered by each ABS Land Cover class in the closing year to obtaining the closing stock.

The net change matrix is compiled for each region according to the following steps:

Creating a matrix for a region where rows represent the opening ABS Land Cover classification and columns represent the closing ABS Land Cover class (Matrix A). Rows and columns are both ordered by ABS Land Cover class.

• Create a second matrix by transposing Matrix A so that rows now represent the closing class and columns the opening ABS land cover class (Matrix B).

• Calculating the net change in land cover area for each combination of classes by subtracting Matrix B from Matrix A. 

• Calculating the net change into a class by summing each column in the net change matrix created by subtracting Matrix B from Matrix A

• Calculating the opening stock of each land cover class by summing each row in the matrix that represents opening class by row (Matrix A)

• Calculating the closing stock of each land cover class by summing each column in the matrix that represents closing class by column (Matrix A)

The table structure for the land accounts is that outlined in the SEEA CF.

The land cover classification used by DEA is based on FAO LCCS version 2. This is a hierarchical classification that created land cover classification types at 4 levels. For the purposes of the National Land Cover account these were simplified to align more closely with SEEA recommended land cover classes.  The table below shows the Level 4 LCCS classifications used in the DEA land cover products, the corresponding ABS land cover classifications, and SEEA land cover classifications.