Modelling of Lake Erie Water Levels (1870-1970)

During my university days in Civil Engineering, I was interested in hydrology and modelling. I guess this explains my current interest in the debate and science of climate change. I dug out a paper I wrote in 1973 to analyse the the "extraordinary" high-water levels of Lake Erie. My home was near Lake Erie and it was abundantly clear, because of the severe flooding along the lakeshore, that the "environment" around the late was changing. We did not blame anthropogenic global warming (AGW), of course, but as I was required to a project for my Hydrology course (CE 499P) I decided to attempt to model the historical record with the idea being it would be a basis for predictions about the future.

I'm not sure that the idea was sound; but what in retrospect seems to be sound upon re-reading the paper 37 years later:

• We gathered actual environmental data and reviewed the data for relevance
• We published the raw data
• We modeled the time series using current technology (auto correlation, frequency analysis, statistics, etc.)
• It is clear that the environment changes. That's what Nature does.
• We worked very hard to validate the model parameters by confirming that we could model "history". Our view was that unless the model "predicted" historical data, then we had no basis to accept that the model was valid.

We confirmed that we could build a model which accurately "predicts" historical data. I'm not sure now, with the passage of time, that I would feel comfortable using this model for future prediction, however. If I get time and energy I may attack the data with tools now available, e.g. Monte Carlo simulation, more hydrologic/climatic science in the model, etc.

However, I offer the original paper to anyone (especially climate change modellers) who wishes to learn from the approach or use the data.