A network analysis technique using generalized least-squares regression was used to evaluate the current (1993) peak-flow gaging network that provides regional peak-flow information for North Dakota. The analysis was conducted to evaluate the current (1993) network and to determine if reactivating discontinued gaging stations and adding new gaging stations on small drainage areas would improve regional peak-flow information.
Peak flows having recurrence intervals of 15, 50, and 100 years and planning horizons of zero and 10 years for three hydrologic regions in North Dakota were used in the network analysis. Results of the network analysis indicate that the average sampling mean-square error could be reduced by about 10 percent for the 15-, 50-, and 100-year recurrence intervals by reactivating a minimum of two to five discontinued gaging stations in each hydrologic region. The reactivated discontinued gaging stations added to the current (1993) network should be located on streams having small drainage areas and steep main-channel slopes. For the 15-year recurrence interval and a 10-year planning horizon, adding a new gaging station at two new locations hi each region instead of reactivating two discontinued gaging stations in each region would reduce the average sampling mean-square error by an average of about 13 percent in each region. The new gaging stations added to the current (1993) network should be located on streams having small drainage areas and mild or steep main-channel slopes in order to obtain improved regional peak-flow information.
