EVALUATING BANKFULL DISCHARGE HYDRAULIC GEOMETRY AS A DIAGNOSTIC TOOL IN WARNER CREEK, CATSKILL MOUNTAINS, NY

The Catskill Mountains of New York State supply up to 90% of the unfiltered drinking water for New York City. The eastern portion of the Catskill watershed is subject to periodic excessive turbidity sourced from glacial deposits exposed through stream erosion. An increasing trend in frequency of high flow events has resulted in increased erosion and turbidity which is indicative of an unstable stream system. A diagnostic tool for assessing stream stability is the assumed channel forming flow (bankfull discharge, Q_bf). A regression curve that predicts stable channel dimensions (area, width, and depth) as a function of drainage area is estimated based on morphological indicators of an established hydrologic regime. This typically corresponds to flows with a recurrence interval averaging 1.5 years. This concept is assumed to be valid for alluvial channels in equilibrium and can be difficult to apply in mountain streams with steeper slopes and non-alluvial boundaries such as exist in the Catskills. Catskill regional regression curves are available to confirm Q_bf hydraulic geometry (HG_bf) in ungauged streams where morphologic indicators can be identified. This study evaluates whether Q_bf indicators can be identified and hydraulic geometry curves developed in Warner Creek; a 23 km² Catskill mountain stream.

Nineteen channel cross sections were surveyed at locations representing a range of drainage areas. Sites were selected with potential Q_bf stage indicators such as depositional surfaces and slope breaks. The HG_bf and drainage area data was log-transformed and simple power function regression was used to generate Warner Creek HG_bf curves. The curves yield R² correlation values of 0.8 for area, 0.7 for width, and 0.6 for depth. These seem to validate the application of the HG_bf concept in this fluvial setting. Bankfull depths plotted lower and widths higher than established regional curves indicating active channel adjustment following headcut migration. Generally higher slope of the regression curve from Warner Creek indicates more rapid change than overall in the region. Cross sectional area is consistent with predicted curve values suggesting the channel has proper functioning capacity even though it is clearly adjusting to hydrologic trends.