Pairing paleolimnological inference models with mechanistic water column models enhances assessment of lake water quality

Reconciliation of paleolimnology inference models with hindcasts from mechanistic water column models aided the reconstruction of past relationships for total phosphorus/dissolved oxygen and acid neutralizing capacity/pH in highly polluted lake ecosystems in New York State. Pre-disturbance Onondaga Lake, Syracuse, NY, was shown to have experienced seasonal hypolimnetic anoxia even under oligotrophic (<10 ug l⁻¹) phosphorus levels. In the Adirondack Mountains of New York State the paired modeling confirmed that, while many lakes have the potential to eventually recover from acidification by atmospheric deposition, approximately 30% likely experienced naturally acidic conditions (pH < 6) prior to increases in industrial emissions. Comparison between the model results illuminated areas of individual model inadequacy, improved understanding of lake ecology, and increased confidence in the ability of predictive water column models to accurately develop restoration scenarios representing improved conditions. The work presented here is the first such comparison modeling for total phosphorus, dissolved oxygen, and acid neutralizing capacity. The technique remains to be more widely applied geographically and extended to less heavily stressed lake systems. Because a fossil inference and mechanistic hindcast should independently lead to similar results, comparison modeling is a potentially powerful tool for examining past interactions between ecosystem structure and ecosystem functioning.