Channel-routing models were used to route hypothetical releases from reservoirs in the upper Tioga River basin, Pennsylvania. These releases were routed northward down the Tioga River to Lindley, Erwins, and Corning, New York: combined with flows routed down the Cohocton River from Campbell to Corning, New York; and then routed southeastward down the Chemung River from Corning to Chemung, New York. The models used to route the flows of Cohocton and Chemung Rivers accounted for bank-storage discharge and streamflow depletion by well pumpage. In general, 17 water years of concurrent streamflow data were available for model calibration and verification.
Three hypothetical reservoir releases were made from the reservoirs and routed to Wilkes-Barre, Pennsylvania, using the models developed in this study and models developed downstream to Wilkes-Barre in a previous study. A hypothetical makeup water requirement of 65 cubic feet per second was assumed. Two historical low-flow periods were investigated. The first hypothetical release investigated was a constant 100 cubic feet per second, and the second release was a constant 70 cubic feet per second. The third scheme was a hypothetical release of 100 cubic feet per second for three days followed by a constant 70 cubic feet per second for the duration of the period considered. Constant 100 cubic feet per second releases arrived downstream more quickly than constant 70 cubic feet per second releases for both test periods, but delivered more water than required to satisfy the assumed makeup requirement. The third release scheme was generally the most efficient of the three schemes tested.
Although inherent modeling errors exist in all the simulated data, the accuracy of the estimated routed reservoir releases at the downstream sites is considered good.
