Paper No. 9
Presentation Time: 3:45 PM
ASSESSING LANDSCAPE RESPONSE TO LAND-USE CHANGE USING SEDIMENT-CHEMICAL CHRONOLOGIES, CLUSTER ANALYSIS, AND BACKCAST MODELING
LONG, David T., Geological Sciences, Michigan State University, 288 Farm Ln, East Lansing, MI 48824, VANNIER, Ryan G., Geological Sciences, Michigan State University, 206 Natural Science Building, East Lansing, MI 48824, PIJANOWSKI, Bryan, Forestry and Natural Resources, Purdue University, Forestry and Natural Resources 715 West State Street Purdue University, West Lafayette, IN 47907 and PARSONS, Matthew J., Geological Sciences, Michigan State University, 206 Natural Science, East Lansing, MI 48824, long@msu.edu
Understanding the impact of land-use change (e.g., urbanization, deforestation) on the flows of mass and energy in watersheds has been complicated by the lack of metrics to learn from the past. Recently, an innovative Backcast Model, which is a modification of an artificial neural network and GIS-based forward casting Land Transformation Model has been developed. In this paper we compare temporal chemical trends from lake sediments in Michigan, U.S.A. to past changes in land use based on the results of the Backcast Model. The hypothesis is that in the absence of irregular and/or rapid changes the flow of mass and energy in a watershed comes into balance (equilibrium, steady state) with watershed physical (e.g., erosion), chemical (e.g., weathering), and biological processes (e.g., succession). The condition of balance might be recognized by a constant pattern in the relative concentrations of chemicals in the lake sediments over time. Sediment cores were collected and sectioned from lakes and metals analyzed via ICP-MS. Historic land-use change was modeled in a 500 meter buffer around each lake. Metals were grouped to represent proxies for pollution and watershed export and patterns within these proxy groups explored as a function of time using cluster analysis.
The following questions were addressed: 1) can periods of balance be recognized, 2) can a relationship of temporal patterns with land-use change be identified, 3) can tipping points and regime shifts be recognized and 4) are the systems returning to balance? The results are positive for all four questions. Long periods of balance could be recognized; relationships to land use change (e.g., urbanization, disappearance of agriculture, decreased forest cover) and regime shifts identified; (e.g., logging); and a possible return to balance indicated. If the systems are returning to a new state of balance it is not the same state as pre-disturbance conditions indicating 1) human influences (coupled with climate change) on the landscape maybe at such an intensity that systems cannot return to pre-disturbance states and 2) interpreting pre-disturbance states as a reference condition must be made with caution.
