Veldhuisen, C. and Haight, R., 2001. Predicting 100-year peakflows for small forested tributaries of the Skagit and Samish Rivers. Skagit System Cooperative, La Conner, WA. pp. 16.

Despite their infrequent occurrence, major peakflows (>25-year recurrence, or annual probability <4%) play a major role in damaging road crossings and other stream-associated structures (Furniss et al. 1998), often resulting in both serious economic costs and significant aquatic habitat impacts (Hayman et al. 1991). An easily applied method for estimating streamflow rates during high-magnitude events can be valuable for a variety of design applications, including road crossing structures and habitat enhancement projects. Though various peakflow magnitudes (e.g. 2-year, 10-year, or 100-year) are relevant to different projects, the 100-year magnitude often represents the upper design standard for hydraulic capacity, and has recently been adopted as the regulatory design standard for forest road crossings (previously 50-year), as well. Realistic peakflow estimates for small watersheds (basin area <10 mi2) are particularly useful, since small streams constitute the largest number of road crossings across the landscape, and thus the largest number of potential failure sites.
I b. Peakflow Prediction Methods
Peakflow response varies considerably between small watersheds and it is generally recognized that such differences reflect the influence of various watershed conditions including soils, vegetation and land use in addition to precipitation inputs (Dunne and Leopold 1978). However, because the level of effort required to account for all such basin attributes in peakflow estimation is prohibitive for most basins, regional empirical regression equations are commonly used for ungaged streams. Such prediction equations are developed from historic peakflow data collected at comparable gaged basins, and generate peakflow estimates on the basis of one or two key basin attributes, typically basin area and annual precipitation. Because only two basin attributes are accounted for directly, the accuracy of the peakflow estimates depends to a large extent on the similarity of the ungaged stream being analyzed to the pool of gaged streams that was used to develop the regression equation.
I c. Project Objective
The objective of this analysis is to identify a simple and relatively accurate method for estimating the 100-year peakflow for small, forested tributaries to the Skagit and Samish Rivers (Water Resources Inventory Areas #3 & 4), the watersheds of interest to the Skagit System Cooperative (S.S.C.). The primary application of this method is for the design of forest road crossings. The accuracy of peakflow estimates is especially important for very small watersheds (basin area <1.0 mi2 or 640 acres), since these stream crossings typically utilize culverts, which are amenable to simple hydraulic design. In contrast, road crossings of larger streams normally utilize bridges that provide such abundant channel clearance that capacity analysis is unnecessary.