Published online by Cambridge University Press: 20 January 2017
Many proxy indicators of past climate integrate the effects of temperature and precipitation, and not all indicators integrate these effects in the same way. This poses a problem in the reconstruction of past climate. A study of the response of tree growth around Lake Superior and the water supplies to that lake to summer climate shows differences and similarities in response. Tree growth is positively correlated with precipitation—although the response is reduced when conditions are very wet—and negatively correlated with temperature. In contrast, the water supplied to the lake are positively correlated with both precipitation and temperature; the correlation with temperature is, however, not very strong. Multiple regression models developed to assess the relative magnitudes of the similarities and differences show that the similarity in response to precipitation accounts for about 30% of the variance in water supplies, the nonlinear response of trees—represented by a log transformation—accounts for an additional 15%, and the difference in response to temperature is equivalent to about 30% of the variance in water supplies.