To meet action 16 of the NSW plan for nature, the Department of Climate Change, Energy, the Environment and Water (DCCEEW) has piloted a new method to monitor woody vegetation regrowth in New South Wales.
This page presents results from 3 pilot regions, covering 3% of the state. Areas of detected regrowth and clearing were compared against a biophysical indicator used as a surrogate for biodiversity value. Differences between the 3 regions show that clearing and regrowth should not simply be combined into a single ‘net’ woody change value, as proposed in the review of the Local Land Services Act.
Detecting regrowth is harder than detecting clearing. Clearing causes a sudden, obvious change in vegetation cover; regrowth happens gradually and can take years to become visible.
Why regrowing woody vegetation is important
Woody vegetation in New South Wales is a dynamic part of the landscape that has changed constantly since European settlement altered land use patterns. The Statewide Land and Tree Study (SLATS) has reported on cleared or harvested woody vegetation, including plantations, since 2006. Identifying when woody vegetation regrows on cleared land is part of a complex and evolving relationship between the landscape and its inhabitants.
Monitoring regrowth provides a full picture of both woody vegetation losses and gains, for carbon accounting and habitat availability. Under action 16 of the NSW plan for nature, the NSW Government committed to ‘ongoing native vegetation monitoring, including considering establishing a legislative requirement, and enhancing the Statewide Landcover and Tree Study program to capture both vegetation losses and gains, including revegetation and regrowth, reduce reporting time lags and enable accurate reporting of net changes in vegetation cover over time’.
How we detect woody regrowth
DCCEEW scientists in collaboration with the Joint Remote Sensing Research Program developed a new method to detect woody regrowth using Sentinel-2 satellite imagery collected over 3 years. By analysing a time series of images rather than a single snapshot, the method can identify the small, gradual increases in tree and shrub cover that occur as vegetation regrows.
Regrowth is identified based on changes in foliage projective cover (FPC), which is a measure of how much sky is obscured by trees and branches. Methods to extract precise FPC values from satellite data have been developed and ground-truthed using LiDAR data. This provides a sensitive measure of vegetation cover and helps detect subtle increases in vegetation over time.
Looking at FPC change over 3 years accounts for seasonal differences, cloud and haze. Areas where FPC rises over time are reviewed by experts for visual evidence of regrowth.
Technical details of the method and the FPC product are being prepared for publication.
Woody regrowth pilot
The pilot tested the method across 3 NSW regions (tiles):
- Brewarrina
- Bathurst
- Dubbo.
Together, these capture different vegetation types, climatic conditions and landscapes. Each tile was analysed over a different 3-year period, allowing the team to test the method against seasonal variation. The pilot focused on regrowth on agricultural land, for comparison with clearing detected in SLATS's agricultural land cover class.
To estimate ‘net changes in woody vegetation cover’, we compared areas of regrowth against reported clearing for the same tiles and time periods. To gauge the potential ‘habitat value’ of the vegetation involved, we identified an index that approximates vegetation maturity using FPC values.
The data for the first time shows, from direct, systematic measurements, that regrowth varies substantially across New South Wales. These results are preliminary, and not indicative of statewide figures, since the 3 tiles cover only 3% of New South Wales.
The case study provides an opportunity for feedback from the wider community to enable the methods to be further developed. If you would like to provide feedback, please contact [email protected].
Early results show strong regional variation
The results of the pilot show large differences in the areas of clearing and regrowth between tiles – see the figure below.
There is substantial variation between the 3 tiles: Bathurst and Brewarrina had considerably more clearing than regrowth, while Dubbo had 1,405 hectares of regrowth against 1,031 hectares of clearing, 36% more regrowth than clearing when measuring only based on area.
The figure below shows FPC values for areas before clearing and after regrowth across the 3 tiles, giving a sense of the vegetation's characteristics in each case. Each box plot splits the FPC measurements into quarters: the box shows the middle 50% of values, the ‘whiskers’ show the top and bottom 25%, and the line inside the box marks the median.
Cleared agricultural areas generally show higher, more variable FPC values than regrowth areas. This likely reflects clearing across a mix of vegetation ages and densities, including mature vegetation with higher FPC. By contrast, regrowth is limited to vegetation up to 3 years old, giving a narrower and more consistent range. This pattern holds across all 3 tiles, though the degree of overlap varies. This is consistent with intuitive observations that young regrowing woody vegetation is generally even-aged saplings, often of the same species, whereas older mature vegetation in agricultural settings is likely to include trees and shrubs of a wider range of ages and species.
Dubbo shows the closest match between clearing and regrowth FPC values, with substantial overlap and nearly identical medians (123 for clearing and 122 for regrowth). Since the pilot covers only 3 of New South Wales's 109 analysis tiles across 3 time periods, these results cannot yet be compared with patterns across the state.
Further work required
This analysis shows how complex it is to compare cleared and regrowing vegetation. The small sample means no consistent overarching trend can yet be identified. Broader analysis is needed across more of the state, including tiles with different vegetation types, rainfall, and land uses.
Drawing a meaningful ‘net vegetation change’ figure from these results is difficult for several reasons, including whether regrowth is truly ‘like for like’ with what was cleared, in structure, species diversity, or vegetation type. These factors are assumed to relate to the FPC changes described above, but further research is needed to map that relationship explicitly.