Stream-restoration projects utilizing naturalstream designs frequently are based on the bankfull- channel characteristics of a stream reach that is accommodating streamflow and sediment transport without excessive erosion or deposition. The bankfull channel is identified by the use of field indicators and confirmed with tools such as regional curves. Channel dimensions were surveyed at six streamflow-measurement stations operated by the U.S. Geological Survey in the Gettysburg-Newark Lowlands Section and Piedmont Lowlands Section of the Piedmont Physiographic Province in Pennsylvania and Maryland. Regional curves were developed from regression analyses of the relation between drainage area and cross-sectional area, mean depth, width, and streamflow of the bankfull channel. Regional curves were used to confirm the identification of the bankfull channel at a reference reach. Stream dimensions and characteristics of the reference reach were measured for extrapolation into the design of a steam-restoration project on Bermudian Creek in Adams County, Pa.

Dimensions for cross-sectional area, mean depth, width, and computed streamflow of the bankfull channel in all surveyed riffle cross sections in the reference reach were within the 95-percent confidence interval bounding the regression line representing bankfull channel geometry in the Lowland Sections of the Piedmont Physiographic Province. The average bankfull cross-sectional area, bankfull mean depth, and computed bankfull discharge for riffle cross sections in the reference reach ranged from 15.4 to 16.5 percent less than estimates determined from the lowland regional curves. Average bankfull channel width was about 2 percent greater than estimates. Cross-sectional area, mean depth, and computed streamflow corresponding to the bankfull stage at the reference reach were 31.4, 44.4, and 9.6 percent less, respectively, than estimates derived from the regional curves developed by Dunne and Leopold in 1978. Average bankfull channel width at the reference reach was 16.7 percent greater than the Dunne and Leopold estimate.

The concepts of regional curves and reference reaches can be valuable tools to support efforts in stream restoration. Practitioners of stream restoration need to recognize them as such and realize their limitations. The small number of stations available for analysis is a major limiting factor in the strength of the results of this investigation. Subjective selection criteria may have unnecessarily eliminated streamflow-measurement stations that could have been included in the regional curves. A bankfull discharge with a recurrence interval within the 1- to 2-year range was used as a criteria for confirmation of the bankfull stage at each streamflow-measurement station. Many researchers accept this range for recurrence interval of the bankfull discharge; however, literature provides contradictory evidence. The use of channel-characteristics data from a reference reach without any monitoring data to document the stability of the reference reach over time is a topic of debate.