COMPARISON OF DISCHARGE ESTIMATION METHODS WITHIN A FLUVIAL SYSTEM: A CASE STUDY

In studying the impacts of extreme flooding events, stream power (Ω) and unit stream power (ω) are often used to characterize an event and provide information about a given system. Both of these metrics rely heavily on stream discharge (Q), which can sometimes be difficult to obtain in ungauged basins. This poster explores two separate methods for independently estimating discharge based on rainfall accumulation: one field-based and one model-driven. Both methods are applied to a site along the Green River in Colrain, MA. These discharge estimates are then compared to measured discharge values from a nearby USGS stream gauge (located approximately 1 mile upstream). The field method relies on physical measurements of a river’s bankfull width (BFW), average depth, local bed slope, and mean grain size (D50, which we obtain through Wolman pebble counts coupled with equations that relate a characteristic flow velocity to a given mean particle size). The modeling method utilizes several regression equations based on mean slope, drainage area, and annual precipitation data, one from Jacobs (2010), and others that are incorporated into the USGS StreamStats application. By combining average monthly rainfall data within a defined artificial watershed these methods can be used to calculate expected 2, 5, 10, 25, 50, or 100-year-recurrence-interval events. The Q2 (or 2-year flood event, with 0.5 percent probability of recurrence) roughly corresponds to a flow at BFW and is used as an analog to represent this channel-forming flow. Field sampling is the backbone of direct observational research, while computer modeling reduces cost and increases the range and size of areas that can be screened or assessed. Generation of comparable data between methods validates modeling techniques and suggests that models can yield physically meaningful results.