Datasets & Software
An open-access global training set of coastal ecosystems to support global analyses of distribution and change of the global coastal zone. Built in collaboration with the Global Mangrove Watch and Allen Coral Atlas teams. Learn more about coastTrain here.
Global Tidal Marsh Dataset
Our collaborative team led by Dr Tom Worthington developed the first globally consistent tidal marsh distribution map. The map depicts the estimated distribution for tidal marshes globally for the year 2020 at 10-m resolution. Read the preprint, which describes the methods and results of this analysis, here and access the data via web-app here.
Australia Saltmarsh Map
Global Ecosystem Typology
Led by Professor David Keith, the IUCN Global Ecosystem Typology was developed to create a comprehensive and consistent framework that supports efforts for assessing and managing the world’s ecosystems and their services. Visit www.global-ecosystems.org.
Myanmar ecosystem assessment
To support the Threatened Ecosystems of Myanmar project, a national scale Red List of Ecosystems assessment, we developed classification of Landsat and Sentinel data to map the Biomes, Realms, and Ecosystems of Myanmar. Download the raster and vector files of Myanmar's terrestrial ecosystem distributions here. Read our paper on the status of Myanmar Ecosystems in Biological Conservation and visit https://myanmar-ecosystems.org for further information.
Earthtones: R package
Earthtones is an R package that downloads a satellite image for a place on earth, translates the image into a perceptually uniform color space, and returns a color palette for use in figures. Developed with a simple idea over lunch at UNSW with Will Cornwell and Mitch Lyons. Earthones is here.
Global aquaculture (experimental)
We host 20TB of analysis ready coastal covariate layers suitable for quickly developing and deploying novel remote sensing analyses at the global scale. Our experimental global aquaculture map has been run only once globally using our in-house coastal training sets, and is likely to benefit from substantial further model development.