Australian habitat image collection

Full description

These images of terrestrial biodiversity habitats across Australia relate to a project that aimed to construct and test a method for habitat condition data capture across Australia using expert elicitation. The images in this collection were assembled from various sources to represent habitats from a wide variety of vegetation types and climates in a variety of different condition states (from ‘good’ to ‘poor’). These images represent a continent-wide library suitable for various purposes, including training and validation of model-based approaches to habitat condition assessment.
Lineage: Background
In order to enable the cross-calibration of habitat condition assessments between experts, a set of reference vegetation images was collated. The reference image criteria were developed from the range of unique structural groupings/geographic combinations which were generated by intersecting the Australian National Vegetation Information System (NVIS) Major Vegetation Groups (MVGs) and Hutchinson agro-climatic zones (Hutchinson et al. 2005). Each image was required to clearly show a landscape perspective while enabling elements within that landscape perspective to be discernible. As no single such reference vegetation image library existed, a specialised library was compiled for the project from disparate sources.

The first stage of compiling the image library was to audit existing vegetation image libraries identified repositories in State and Federal agencies, and in research, conservation and citizen science projects. Of the image libraries identified, only a selection included publicly accessible images. Also identified were a number of extensive privately owned libraries with permission granted for project use. Those image collections which included reliable spatial reference, open access and condition range (poor, moderate and good) were prioritised to include as principal image sources. The AusPlots images, including 590 images with a comprehensive spatial range, were selected as the primary image source, and supplemented by the private libraries and some publicly available images, resulting in a central project library of 777 images.

Image processing methods
All images needed location information to be aligned to MVGs and Hutchinson zones. AusPlot images possessed spatial coordinates. Images from private collections had either coordinates for their location or a description. Images with descriptions were located in Google Earth and given coordinates by cross referencing their descriptions with the MVG assigned by their authors. Some private images were assigned to more than one MVG by the image author. The first MVG was considered the most important in most cases.

All images that were assigned a MVG group and Hutchinson class were visually inspected and graded as suitable or not suitable for image assessments. Suitability was based on good colour and exposure, a clear and broad view of the landscape. All suitable images were rated based on their condition as either good, moderate or poor ecological condition, by the photo authors.

Data products
The attributes of each image in the final collection of images were formatted and prepared into the following files:
ImageAttributes.csv – Key attributes of each image collated.
Descriptor_ImageAttributes.csv – A file describing the fields used in the ImageAttributes.csv data file.
ImageFiles – A folder holding all the habitat images collated.

References
Hutchinson, M. F., McIntyre, S. , Hobbs, R. J., Stein, J. L., Garnett, S. and Kinloch, J. (2005), Integrating a global agro-climatic classification with bioregional boundaries in Australia. Global Ecology and Biogeography, 14: 197-212. doi:10.1111/j.1466-822X.2005.00154.x