Remote sensing of grass condition

Grass fuel condition is an important metric for grass and crop fire behaviour potential. Key decision tools such as the Australian Fire Danger Rating System (AFDRS) and the Spark wildfire simulation toolkit rely on data-driven knowledge of grass fuel conditions over large areas of the Australian land mass.

Grass condition is used for predicting the spread of active fire which directly informs community warning messages and incident management tactics. At present, grass fuel condition is classified as: i) natural/unharvested, ii) grazed/harvested and iii) eaten out/harvested & baled. With three classes only, this classification provides a limitation to any future improvements to grass fire models. Also, grass fuel condition is commonly estimated and extrapolated from low-quality data, there are no standardised and automated ways to detect and account for harvest or other changes to grass fuel condition in pasture and crops, and there has been limited engagement between jurisdictions to address this problem. Inaccurate grass fuel conditions lead to misrepresentation of the fire behaviour index and hence fire danger ratings and the issue of appropriate warnings. This impacts the public, agricultural business, prescribed burning activities and other land management practices.

This project aims to develop a national satellite-based methodology for detecting, attributing and validating changes in grass fuel conditions (in near real-time) across a range of soil types in crop and pasture fuel types to increase the realism of fire danger ratings and warnings for fires in such grassy fuel types.