Changes in land cover and wildlife habitats in two watersheds in the lower Peninsula of Michigan

Changes in land cover and changes in wildlife habitats were analyzed in two watersheds in Michigan's Lower Peninsula from the early to mid 1800's to the early 1990's. The Huron River watershed in southeastern Michigan near Detroit underwent extensive conversion from mostly forested (70%) to mostly agricultural (55%) from the early 1800's to the late 1930's. From the 1930's to the 1990's, urban areas expanded from 5% to 29% at the cost of agricultural land. Forest and nonforest areas increased during that time period as well. The Black River watershed in the north central Lower Peninsula underwent extensive clearcutting from the mid 1800's to the early 1900's. Since timber harvesting stopped, the Black returned to a mostly forested condition (73%). However, forest age and composition changed markedly. Conifer forest declined 56%, from 84,000 ha in the 1800's to 37,000 ha by the 1990's. Broadleaf forest increased 14%, as large gains in early successional aspen/white pine (5,000 ha to 40,000 ha) offset losses in northern hardwoods (56,000 ha to 16,000 ha). In both watersheds, mean patch sizes and the number of patches have decreased/increased, respectively, by one to two orders of magnitude.; Habitat changes varied for both watersheds and depended upon the species in question. Losses in the amount of potential habitat occurred primarily from the 1800's to the 1930's. From the 1930's to the 1990's, the amount of potential habitat showed minimal change in the Black River watershed and actually increased for a majority of species in the Huron River watershed. However, in both watersheds, patch sizes of potential habitat typically declined by one to 2 orders of magnitude from the 1800's to the 1930's and remained the same for most species from the 1930's to the 1990's.; The feasibility of modeling future land cover change based on observed patterns of land cover change was investigated. Overall models performed well at predicting anthropogenic changes related to regular features on the landscape such as roads. The models performed poorly at predicting natural changes such as succession. Additional information would be needed to increase the ability of the models to predict future land cover change.; Despite extensive habitat changes, 90% of species still occur in both watersheds. How many and which species will continue to persist in these modified landscapes will require further research. In particular, research should focus on understanding species-habitat relationships at landscape scales (typically 10's to 100's of kilometers) and where land cover data are limited to broad categories. This information, when combined with more detailed studies of wildife-habitat relationships, will provide important insights into species abilities to persist in highly-modified landscapes.