FLOOD08374 - Laserfiche WebLink

Home Browse Search

-9- the hydrologic regime, but detection of such changes would require de- tailed, closely-spaced meteorological data which do not exist. There- fore, climatic trends cannot be incorporated as a predictive factor in a model relating urbanization and changes in flood expectancy. Moreover, inclusion would have little usefulness since future climatic fluctuations cannot be estimated reliably. The study area is located in southeastern New England as shown in Figure 2. The terrain of this region is distinctly glacial in origin. The western part is characterized by gently rolling, till-mantled hills with intervening ponds, swamps and stream valleys underlain by glacial sand and gravel. The eastern portions are generally flatter and consist of extensive outwash plains, wetland areas and scattered hills of till, bedrock or ice-contact glacial deposits. Owing to the humid climate, streams tend to be permanent and flow in deep, well-established channels. Twenty-six gaging sites were originally selected from all available sites in the region on the basis of the length of hydrologic records. Flood frequency relations were appraised for these watersheds and informally related to the historical pattern of urban growth. Eighteen of the ori- ginal 26 watersheds were then identified as being suitable for model development on the basis of more detailed hydrologic information. The remaining eight were disqualified for several reasons including excessive regulation of streamflow, indications of major diversions during peak discharge, a drainage area larger than 500 square miles which was judged to be impractical and might produce a non-homogeneous sample, and finally redundancy of hydrologic information in the case of one watershed being a subdrainage of another. Since two of these criteria are a matter of degree, it is clear that the study watersheds are only relatively natural