Laserfiche WebLink
TABLE 3 <br />Summary statistics for eight New England hydrologic regions <br /> <br />Region Median <br />cfsm <br />Mean cfsm (range) <br />SD 95% C.I. <br />lower 95% C.I. <br />upper Sample <br />size <br />Mountain-Windward 0.63 0.62 (0.50-0.78) 0.07 0.58 0.67 11 <br />Mountain-Leeward 0.29 0.31 (0.21-0.56) 0.09 0.26 0.36 15 <br />Interior-West 0.34 0.33 (0.25-0.39) 0.07 0.25 0.41 3 <br />Interior-Central 0.17 0.21 (0.16-0.33) 0.08 0.08 0.34 4 <br />Interior-East 0.27 0.28 (0.21-0.43) 0.07 0.22 0.33 8 <br />Gulf of Maine 0.38 0.38 (0.37-0.38) 0.01 0.31 0.44 2 <br />L. 1. Sound 0.36 0.40 (0.22-0.62) 0.12 -0.11 0.90 3 <br />St. Lawrence 0.42 0.43 (0.38-0.49) 0.06 0.29 0.57 3 <br />All regions 0.34 0.38 (0.16-0.78) 0.16 0.21 0.63 49 <br />gland, westerly winds convey storm sys- <br />tems through the region during summer <br />months (NOAA 1977a, 1977b, 1977c). To <br />determine if this hypothesis was true, high <br />elevation gages (greater than 1,200 feet) <br />were sorted, using USGS topographic maps, <br />on the basis of whether each drainage was <br />predominantly on the windward (west) or <br />the leeward (east) side of mountain ranges. <br />The difference in the mean cfsm values of <br />these two groups was tested and found to <br />be significant (P < 0.05). <br />Streams with mean basin elevations less <br />than 1,200 feet were termed interior, and <br />were initially subdivided into western, <br />central, and eastern interior, dictated by <br />location relative to dominant mountain <br />ranges (as defined in Table 2). The general <br />position of these gaged drainages in rela- <br />tion to mountains was chosen as a basis for <br />the subdivisions, due to the importance of <br />topography in inducing localized climatic <br />and rainfall patterns (Chow 1964). Coastal <br />drainages are also known to have localized <br />climatic precipitation patterns (NOAA <br />1977a) and also generally low mean basin <br />elevation and stream slope; all or most of <br />these drainages were within 30 miles of <br />the coastline. Because of observed differ- <br />ences in rainfall between coastal southern <br />New England and the Gulf of Maine <br />(NOAH 1979), the coastal region was sub- <br />divided into Long Island Sound and Gulf <br />of Maine subregions. An eighth region (St. <br />Lawrence) included all basins located to <br />the north of the Appalachian Mountains, <br />at elevations below 1,200 feet, in an area <br />influenced by St. Lawrence River valley <br />climate (NOAA 1977b, 1977c). <br />The next step after eight possible hy- <br />drologic regions were identified, based on <br />physiographic basin characteristics (Table <br />3), was to test the statistical validity of each <br />region's cfsm value and to test the hypoth- <br />esis that significant statistical differences <br />exist among the regions. The mean and <br />median (as measures of central tendency) <br />were used as indices to verify that the clas- <br />sification of cfsm values into the hydro- <br />logic regions was reasonable. <br />Although most regions have small sam- <br />ple sizes, there is still good conformance <br />between mean and median values for all <br />regions. Therefore, mean regional values <br />could be safely used as a parametric statis- <br />tic for the additional analysis presented be- <br />low. <br />Sample sizes of various regions ranged <br />from 2 (Gulf of Maine) to 15 (mountain <br />leeward). The mean, median, standard de- <br />viation, and 95 percent confidence interval <br />(C.I.) for each region are presented in Table <br />3. Mean cfsm values ranged from 0.21 cfsm <br />(central interior) to 0.62 cfsm (mountain <br />windward). The mean and 95 percent C.I. <br />for each region are plotted in Figure 5. <br />The mountain windward region has ex- <br />cellent agreement among median and mean <br />values, as well as low standard deviation <br />and a small C.I. Good general conformance <br />is also evident for mountain leeward and <br />interior east regions, each of which con- <br />tained relatively large sample sizes. <br />Of the other regions, small sample size <br />and wide C.I.'s make statistical differentia- <br />tion of discharge data among those groups <br />difficult. For example, in the case of the <br />Long Island Sound gages, both the small <br />I Z 18 Rivers • Volume 1, Number 1 January 1990