More accurate vegetation fuels maps

Many of you will recall the work Badola and Panda have been doing at UAF to test development of low-cost, detailed vegetation maps for boreal Alaska using widely available Sentinel-2 satellite data. Fire managers challenged Anushree Badola (for her PhD work) to make fuels maps that were more accurate than LANDFIRE and better separated black spruce from mixed and deciduous forest types. That work (and her doctorate) was completed last year, and indeed, for the interior Alaska testing site, they appear to be twice as accurate as LANDFIRE at a 30m resolution and do a good job of distinguishing forest types. But did you know that she has made those maps (which cover most of Alaska with the exception of the western end of the Seward Peninsula and the North Slope) available for free download? All is explained succinctly in her STORYMAP and there is also a 2-minute video refresher by Anushree about the project, which was sponsored by National Science Foundation grants and the State of Alaska, with lots of input from the fire management community. Here’s a screenshot from the Storymap as a teaser!

Screenshot of Badola & Panda storymap
Panda & Badola (UAF) obtained 65% classification accuracy for the new vegetation map compared to 32% accuracy for LANDFIRE in a small test area. https://storymaps.arcgis.com/stories/41df03ea574444c98280a49351cf512a

Leave a comment