Spatial BioCondition

Written By Catherine James

Background

Scientists from the Department of Environment and Science are developing, in collaboration with Dr Leonardo Hardtke from the Joint Remote Sensing Research Program, a spatial modelling framework to map state-wide vegetation condition for biodiversity for the Queensland Government.

The Spatial BioCondition framework aligns with the Queensland Government’s Regional Ecosystem mapping and BioCondition frameworks, and aims to complement the Statewide Landcover and Trees Study.  The BioCondition framework provides a site-based metric to reflect the capacity of an ecosystem to maintain biodiversity values at a local or property scale. The metric is a scale between 0 and 100, with higher scores representing increasing functional condition for biodiversity.  The aim of Spatial BioCondition is to extend this metric beyond the site.

Site-based assessments are time intensive and not feasible when representation across the State is needed to support biodiversity planning and policy.  However, with advancements in satellite technology, greater computer processing ability and suitable training data sets, assessing the condition of terrestrial vegetation remotely at a continental scale is now possible.

How

Research and development into the Spatial BioCondition Framework started in 2019 and the first machine-learning model was run using both Sentinel and Landsat data. Later iterations of the model now only require Sentinel data, as it is available State-wide and easily repeatable. The model measures the relative difference between any location, and its reference or ‘best-on-offer’ state, relative to a specific Regional Ecosystem.

The code Leo has created to run the Spatial BioCondition model relies on Queensland’s Regional Ecosystem mapping and lots of field data from site based BioCondition assessments.  As the code and product conception have developed and evolved since 2019, independent field assessments have been completed to validate the accuracy and validity of the output of each model iteration, with positive results.

The ongoing enhancements, such as incorporation of fractional cover derivations like temporal metrics for foliage projective cover and growing season parameters from the NDVI, have enabled the original thirty model input datasets to be refined to twelve.

Applications

Although the full extent of possible applications is yet to be recognised, the following are predicted:

  • Support assessment methods for biodiversity and vegetation condition for Natural Capital Accounting

  • Planning nature conservation reserves by conservation managers

  • Support equivalency of development and offset sites for Biodiversity offsets

  • Co-benefits for carbon accounting

  • Monitoring the state of native vegetation for State of the Environment reporting.

Products

The Spatial BioCondition product will be a 30m raster available across Queensland. This is hosted on the Qld Government high performance computing and spatial data infrastructure.

Where to next

  • Finalise the testing - user groups to trial the product, ensuring accuracy, reliability, and functionality

  • Extend to multi-temporal measurements of BioCondition to try monitor trends over consecutive years.

Acknowledgements

  • The University of Queensland

  • Queensland Herbarium and Biodiversity Science, Department of Environment and Science, Queensland Government

  • Earth Observation and Social Sciences, Department of Environment and Science, Queensland Government

To find out more

Catherine James
Previous

Woody Regrowth
Next

Lidar