Assessing the response of watersheds to catastrophic (logging) and possible secular (global temperature change) perturbations using sediment-chemical chronologies

David T. Long, Matthew J. Parsons, Catherine H. Yansa, Sharon S. Yohn, Colleen E. McClean, and Ryan J. Vannier

    Geochemical profiles of sediment cores from two oligotrophic lakes (Elk and Mullett) in northern Lower Michigan were studied to examine the response and recovery of watersheds to large-scale logging that occurred between 1850 and 1920. Specific questions addressed were: can the impact of extensive clear-cutting of forests be recognized in the sediment-chemical chronologies, can states of system stability be identified prior to the logging, and are there indications that the systems are recovering and possibly returning to a stable state? To answer these questions, elements were put into four groups as proxies for watershed runoff or export (e.g., Al, Mg), pollution (e.g., Pb, Cu), redox (e.g., Fe, As), and productivity (e.g., Ca, P). It was observed that vertical patterns of all proxies were influenced by logging and the early increases in concentration of pollution proxies were due to increased watershed export, not pollution. System stability might be recognized by relatively symmetrical vertical patterns among all of the proxies or secular changes of slowly increasing or decreasing vertical concentration trajectories. Some pre-logging trajectories were punctuated by episodes of slightly elevated concentrations that appear to be related to comparatively warmer periods during the Little Ice Age. Iron and Mn enrichments caused by increased watershed runoff might be misinterpreted as paleo-redox horizons. Results are interpreted to indicate that (1) reference conditions may be better defined as the temporal trends among proxy groups and not individual concentrations, (2) simply assuming pre-1800 conditions as a reference may not be appropriate, (3) inter-proxy group comparisons are needed to help for interpretations of intraproxy group patterns, (4) the possible regime shift identified here might be expected for other ecosystems because of the intensity of human disturbances and secular changes, and (5) without consideration of a possible regime shift, recovery from logging is estimated to be on the order of 75130 a, but shorter if regime shifts are considered.