How fragmented are forests in the United States? Different mapping approaches tell different stories
Forest fragmentation maps inform wildlife management, landscape-level planning, and conservation decisions. However, forest cover mapping approaches vary, and a new study shows this can have big impacts on forest fragmentation estimates across the United States.
Maps of forest fragmentation are used to determine habitat connectivity, informing management decisions from installing wildlife crossings to designating conservation areas. But estimates of forest fragmentation rely on maps of forest cover, where pixels in satellite images are classified as either forest or non-forest. Surprisingly, these mapping efforts are not always clear-cut. For example, one approach is to classify a pixel as forest if tree cover exceeds a certain portion of the pixel. Another approach defines forest in terms of the confidence one has that the image shows forest versus some other land cover.
In a new study by Forest Service scientists, different forest mapping approaches showed up to 36% difference in the total amount of forest across the lower 48 states. Lowland forests, including in coastal plains and around the Great Lakes, were excluded in some maps, instead classified as wetlands. Some maps classified tree cover in developed areas as forest, while others deemed it urban. As a result, fragmentation estimates also varied, especially in lowland, mixed-use, and recently disturbed areas. Estimates of fully connected forest area varied from 5% to 16% of total forest area, with highly fragmented forest varying from 37% to 44% of the total. Focusing on the details – when to call a pixel a forest – may be exactly what is needed to see the bigger picture of forest fragmentation.
