Evaluating accuracy of two types of early land survey records in the northeastern United States

The accuracy of early land survey records often used ro reconstruct colonial forest composition was tested for three regions of the northeastern United States: central Massachusetts, the southern part of Berkshire County in eastern Massachusetts and adjacent towns in Connecticut, and northeastern Pennsylvania. Simple linear regressions of early proprietor's survey data on average growing degree days revealed significant relationships for oak (Quercus spp.), beech (Fagus grandifolia), and hemlock (Tusga canadensis) with consistent slopes for all three regions. The patterns fit well with colonial forest composition as inferred from regional pollen studies and with climatic preferences of the taxa, suggesting that these proprietor's data provide an accurate representation of forest composition. The same analyses were done with a subset of five towns where two sets of data were available: proprietor's data and road survey data. There were significant differences between the two data sets, with the road surveys not related to climatic gradients. We suggest that different types of survey records represent different aspects of forest composition, so that data from specialized surveys such as those for roads should be used only in spatially specific studies where the location of the surveys within a town can be taken into consideration.