IDENTIFYING THE INFLUENCE OF URBANIZATION ON LANDSCAPES THROUGH SEDIMENT-CHEMICAL CHRONOLOGIES AND BACKCAST MODELING

Understanding the impact of urbanization on the flows of mass and energy in watersheds is complicated by the lack of metrics to learn from the past. Here we compare temporal elemental concentration trends in inland lake sediments to past changes in land use based on results from a Backcasting Land-Use Model which is a modification of an artificial neural network and GIS-based forward casting Land Transformation Model. The hypothesis is that urbanization will disrupt the flow of mass across a watershed, but as rates of change lessens, the flow will come into balance (equilibrium/steady state) with watershed physical, chemical, and biological processes, but at a different state from that of historical.

Sediment cores (Michigan, USA), were sectioned immediately after collection. Elements were extracted by microwave digestion and analyzed via ICP-MS. Land-use change was modelled since 1900 in a 500 meter buffer around each lake. Elements examined included proxies for pollution (e.g. Pb), in-lake processes (e.g., Ca), and terrestrial export (e.g. Al). Elemental patterns (i.e., relative associations of elements) for each time slice down core were revealed through cluster analysis.

Land use in 1900 followed the order forest>agricultural>wetland >= grassland>> urban. In recent time, urban>forest>wetland>=grassland>>agriculture. Key time periods (regime shifts?) in order of occurrence recognized by changes in elemental patterns are 1) the onset of urbanization, 2) urban becoming more dominant than agriculture, 3) urban becoming more dominant than forest. These changes were recognized in patterns of both the pollution and terrestrial proxies. The most pronounced pattern change occurs when urban becomes more dominant than agriculture. Recent patterns show an indication of a return to system balance (recognized by symmetrical temporal patterns among and within proxy groups) albeit at a different state from 1900. These results are interpreted to support the hypothesis, but more work is needed to better understand the feedbacks between land-use change and the flows of mass and energy in watersheds. The results do show the potential of combining backcast land-use modelling with temporal environmental records to understand the impact of urbanization on landscapes and assessment of system recovery.