Australian total, woody and grass foliage cover (version 3)

These data describe the fraction of ground covered by green (live) foliage across Australia, including separate estimates of the cover of woody vegetation and grasses. Total foliage cover is derived from MODIS 250 m, 16-day NDVI imagery (MOD13Q1, Collection 6.1), starting in 2001. Woody and grass foliage cover estimates are derived from total cover using version 3 of the persistent-recurrent splitting algorithm of Donohue and Renzullo (in review). Version 3 of this algorithm has reduced errors due to background soil colour effects; it has improved the detection of dense grasslands and pastures; and it better captures the effects of major disturbances (such as fires). For woody vegetation types (including very sparsely wooded types), estimates of total, woody and grass cover are between 86% and 95% accurate. For treeless vegetation types, total and grass cover accuracies vary between 77% and 92%, 66% and 96%, respectively, with cover being increasingly underestimated as grassland density increases.\n \nTwo copies of these data are accessible. One copy can be downloaded from this DAP site. This copy is updated only annually. The second copy is hosted on a separate web server, accessible via a https link (See 'Services' tab), and is updated quarterly.\n \nData are 16-day, 0.0025 degree resolution Geotiff files spanning continental Australia. Units are fractions (0.0-1.0). No data is -9. See the associated metadata and file naming documents.\n\nLineage: 1) 16-day NDVI imagery was converted to total foliage cover using the linear transformation of Carlson and Ripley (1997) with minimum NDVI for the arid region parameterised as the lowest observed NDVI in the MODIS record. Outside this, minimum NDVI was set to 0.2.\n2) Total foliage cover was separated into woody cover and grass cover using the persistent-recurrent splitting method, which identified woody cover as a curve fitted to the annual minimum total cover time-series. Grass cover was the difference between total and woody cover.\n3) Corrections were made to better capture the effects of major disturbances and to better capture perennially green pastures (which are otherwise interpreted as woody cover).\n4) Validation was performed using 4245 field-observed grass and woody cover measurements sourced from the SLATS Star Transects dataset and several Ozflux datasets.\nCarlson TN, Ripley DA (1997) On the relation between NDVI, fractional vegetation cover, and leaf area index. Remote Sensing of Environment 62, 241-252.\nDepartment of Environment and Science, Queensland Government (2022): SLATS Star Transects - Australian field sites. Version 1.0. Terrestrial Ecosystem Research Network. https://portal.tern.org.au/metadata/TERN/24a40c29-0d7c-4fe8-bdde-9c4ea495bfb8\n