From the western 40 percent of the Laboratory area, most of the ground water moves underground toward the Carmans River, reaching it somewhere between the railroad crossing at Yaphank and Route 27.
The three principal rivers in the Upton area, the Carmans, the Forge, and the Peconic, are fed a]most entirely by ground water; that is, they receive very little direct overland runoff [except during very heavy rainfalls -- RT]. For this reason, and because the apparent topographic drainage areas of these streams do not correspond with the ground-water contributing areas, the flow of these streams bears little relation to the area apparently drained by their valleys. Instead it is determined by the configuration of the water table. The differences between apparent topographic drainage areas and the actual ground-water drainage areas are shown in plate 8.
Plate 5 shows profiles of Carmans River from Route 25 to Bellport Bay. Plate 6 shows profiles of channel bottom, stream level, and water table for the Peconic River. Figure 37 shows topographic drainage areas for streams in the Upton area. Plate 7 shows water-table contours, ground-water flow lines, and areas contributing ground-water flow to selected gaging stations. The maps on plate 8 show a comparison of topographic and ground-water drainage area for the Carmans, Forge, and Peconic Rivers.
The valleys of both the Carmans and Forge Rivers may be divided into upper, middle, and lower sections. In the upper sections the streambeds are normally dry because they are above the water table; they carry water only on rare occasions when heavy rain has fallen on soaked or frozen ground. All other precipitation not lost by evapotranspiration soaks into the ground to join the water table. The ground-water flow in these areas may be in quite a different direction from the slopes of the streambeds, as in the upper valley of the Carmans River where the ground water flows north although the valley slopes to the south. Because the water table stands high above sea level in these areas, some of the ground water moves downward to recharge the deep aquifers in the Magothy (?) and the Lloyd.
In the middle and lower sections of the Carmans and Forge Rivers and in the lower half of the Peconic, the flow is perennial and is fed by ground water that moves in laterally and also upward from the lower part of the upper Pleistocene aquifer. The relation of ground-water flow to streamflow was used to calculate the transmissibility of the upper Pleistocene aquifer adjacent to the middle section of the Carmans River.
In the lower sections of the three rivers, streamflow is sluggish and is periodically slowed or even reversed by tidal fluctuations in the bays. In these areas the ground-water inflow from the water-table aquifer is augmented by water moving upward form the deeper artesian aquifers. Because of these complexities, quantitative streamflow measurements in the lower reaches of the streams are difficult to make and difficult to interpret and were attempted only for the Carmans River. In the following sections, the Carmans, Forge, and Peconic Rivers are described with the aid of maps (fig. 37 and plls. 7-9), profiles (plls. 5,6) and tables of streamflow measurements.
The topographic drainage area of the Carmans River is about 100 square miles, but this figure is deceptive because only about half this area contributes water to the river. Some 34 square miles of the apparent drainage area lies north of the ground-water divide, and precipitation falling in this area infiltrates to the water table and then flows north to Long Island Sound. South of the ground-water divide on the west, there is a second area of some 15 square miles that contributes ground-water flow, not to the Carmans River, but to several small south-flowing streams which lie to the west of the river. If the terrain of Long Island were not so very permeable, there would be surface runoff from these areas to the Carmans River, but under the existing circumstances they are drained entirely by ground-water flow. The total area contributing to the flow of the Carmans River is, in fact, only 48 square miles, about one half its apparent drainage area.
Pereimial flow of the Carmans begins at Artist Lake, where the river is crossed by Route 25, 3 miles west of the north entrance to the Laboratory, and ends 8 miles to the south at Bellport Bay. The distance, some 12 miles in length as the stream flows, will be divided into three main segments for the purposes of this discussion (plls. 5, 7).
The first section, roughly that part of the river valley which lies north of and which traverses the Ronkonkoma moraine, extends from an indefinite point north of Artist Lake to Bartlett Road. After a series of wet years, resulting in a high water table, the flow of the river probably begins in a small lake just north of Artist Lake, but after several dry years the flow in late fall, when the water table is low, probably begins near Bartlett Road, more than a mile and a half to the south. Five and a half miles south of Artist Lake at Yaphank, where the river is crossed by the Greenport Division of the Long Island Rail Road, a stream-gaging station has been operated by the U.S. Geological Survey since July 1942. Monthly and yearly records of streamflow for this station from 1942 to 1953 are listed in table 11, and more recent records are available in the publications of the U.S. Geological Survey. The flow of the river at this point is in some measure controlled by the two small artificial ponds upstream. The ground-water contributing area upstream from the gaging station is 21.5 square miles, and the average flow for the period of record at the station is 21.76 second-feet, or about 1 second-foot per square mile, which represents an average annual recharge of 13.5 inches. For other points on the stream there are only scattered measurements (table 12).
The second section of the stream lies between Bartlett Road and the dam just above Route 27, 6.2 miles downstream, on the outwash plain south of the Ronkonkoma moraine, which it has slightly dissected. The streamflow at Route 27, just below the dam, was measured on July 29,1952, at a time when the flow was probably close to average. At this point the stage and flow of the stream are somewhat affected by the tides in Bellport Bay, and corrections for this and other factors were required. From 9:13 a.m. to 10:34 a.m., when the stage was falling because of the falling tide in the bay, the discharge was 52.8 cfs. From 12:54 p.m. until 2:13 p.m., when the stage was rising, the discharge was 36.5 cfs (fig. 38). After corrections for changes in pond and bank storage, the average discharge during this period of normal streamflow was computed to be 47.8 cfs. Partial measurements on one or two other occasions tended to confirm this figure.
The apparent gain in streamflow between the gaging station at Yaphank and the bridge on Route 27 is, therefore, the difference between 47.8 and 21.76 cfs, or about 26 cfs, but two small corrections must be made. About 6.1 cfs was being diverted through the Carmans River Duck Farm upstream from the highway bridge, and the rising change in stage of several small ponds in the Suffolk County Game Preserve represented the holding back of about 0.55 cfs, a total ungaged flow of about 6.65 cfs. The corrected flow at the bridge, therefore, is about 54.5 cfs, and the gain in flow in the 2% miles of river upstream is between 32 and 34 cfs.
The third section of the river, from the highway bridge to the mouth of the river at Sandy Point, a distance of about 3.15 miles, crosses the outwash plain south of the Ronkonkoma moraine. In this section, however, the river is a tidal estuary; it has been aggrading its bottom and its small flood plain. Tidal fluctuations in Bellport Bay and Great South Bay are the main cause of variations in stage of this part of the river which, at the bridge on Route 27, varies from a maximum of 3.82 feet above sea level, to a minimum of 0.34 foot above sea level. The average daily range at this point is 0.83 foot.
The discharge of the stream at its mouth could not be measured, but the flow at this point has been estimated to be 72 cfs.
RELATION OF STREAMFLOW TO GROUND-WATER CONTRIBUTING AREA
The average discharge of the Carmans River at the gaging station in Yaphank for the period of record through September 1953 is 21.8 cfs, and the area contributing ground water to the stream, as determined from the water-table map, is 21.5 square miles. The average runoff thus is the equivalent of 13.5 inches of water. During these years the rainfall averaged 43.5 inches; because about 22 inches was lost by evapotranspiration, the recharge to the water table must have averaged 21.5 inches. The difference between this recharge and the 13.5 inches of streamflow, or about 8 inches, probably represents recharge to the deeper aquifers, the Magothy(?) Formation and the Lloyd Sand Member of the Raritan Formation.
The measured discharge of the river at the bridge at Route 27, at a time of probable near average flow, was between 54 and 55 cfs. The water-table map shows a contributing area of 36.5 square miles, which represents an average annual runoff of about 20 inches, or 6.5 inches more than at Yaphank, and an amount only slightly less than the average annual recharge. The increase in flow in the 2.75 miles of stream between Yaphank and the bridge is 32 cfs and the contributing area is about 15 square miles; runoff for this area is therefore about 30 inches, or somewhat more than the recharge. The excess over the recharge is due to upward leakage from the aquifers below the Gardiners Clay, but, because the average flow for the total area of some 36.5 square miles contributing to the flow at the highway bridge is only about 20 inches, it is apparent that not all the deeper recharge has come back to the water-table aquifer at this point. There must be considerable additional upward leakage into the area south of Route 27, and the estimated discharge of 72 cfs for the mouth of the Carmans River at Sandy Point is based on the estimate that the 48 square miles of area furnishing this flow contributes an average 22 inches of runoff.
RELATION OF STREAMFLOW TO GROUND-WATER LEVELS
Because the streams are almost entirely supplied by ground water discharge, a close correlation exists between the height and slope of
the water table and streamflow (pl. 7). The relation is, however, not always simple or direct, and the seasonal high or low in a particular observation well may come earlier than, at the same time as, or later than the seasonal high or low discharge at a point on a neighboring stream. The more important factors influencing this relation are (1) the depth to the water table at the well site, (2) the position of the well in the pattern of ground-water flow, (3) the distance of the well from the stream, (4) the hydrologic characteristics of the aquifer in question, (5) the hydraulic gradient, and (6) the variations in the amount of water in storage in the aquifer.
Well S3533 (pl. 1) is about 1 mile east of the nearest point on the Carmans River, and about 3.2 miles N. 16° W. of a point on the river at Bartlett Road, on the ground-water flow line passing through the well. By trial and error a good correlation was found between the water-level stage in this observation well and the flow of the river at the gaging station in Yaphank, although the high and low stages in the well lag about 6 weeks behind the corresponding high and low flow in the river. This relation is shown in figure 39, and is used in figure 40 to calculate streamflow from the well hydrograph.
The derived value for streamflow is within a few percent of the gaged value, except for periods such as August 1947 and November 1952 when there was long, intense rainfall. A somewhat similar empirical relation can be worked out between streamflow and the water level in almost any nearby observation well, whether it is in the area contributing ground-water flow to the stream or not, because both are strongly influenced or even controlled by the cumulative recharge to the water table. For example, the water level in well S3532, about 3.5 miles northeast of Artist Lake and 0.5 mile north of the groundwater divide, can be closely correlated to the flow of the Carmans River, although in this case the stage of the well lags about 2.5 months behind the streamflow.