Laserfiche WebLink
<br />Summary <br /> <br /> <br />For Flushing Flow 2. the greatest accuracy in matching actual pool bed change <br /> <br /> <br />was obtain from the HEC-6 robust simulation, where more than 50% of the bed changes <br /> <br />were predicted over the length of the entire pool. <br /> <br />Discussion <br /> <br />Although HEC-6 produced reasonable, first-order approximations of the changes <br /> <br /> <br />in bed elevation due to scour and till after the sediment release from Halligan Reservoir, <br /> <br /> <br />the most accurate results were generated only after a data set collected over a one-year <br /> <br /> <br />period of time was used. Weekly and biweekly field outings over the course of one year <br /> <br /> <br />were necessary to produce a greater than 50% accuracy in the model output. Default <br /> <br /> <br />simulations produced minimal accuracy, and therefore limit the management capabilities <br /> <br /> <br />of the model to situations with adequate field data. At a minimum, the two input <br /> <br /> <br />parameters that cannot be defaulted are thickness of sediment accumulated above the bed. <br /> <br /> <br />and allowance for deposition and erosion. HEC-6 did simulate the best results for the <br /> <br /> <br />entire Tick Pool under Flushing Flow 2. relative to the scour and fill that was quantified <br /> <br />in the field, This condition supports statements in the program manual that longer <br /> <br /> <br />duration tlows are more accurately modeled than short duration discharges. <br /> <br /> <br />Sediment transport modeling using GSTARS 2.0 requires a much higher level of <br /> <br /> <br />expertise. and has many program bugs that need to be rectified. For example, GST ARS <br /> <br /> <br />2.0 cannot accommodate drastic grain size changes over short distances such as occur <br /> <br /> <br />along the North Fork where silt-filled pools are bounded up and downstream by cobble <br /> <br /> <br />and boulder bedded rimes (F. Simoes, personal communication, 1998). At this point in <br /> <br /> <br />time. it appears that the conceptual physical model from which the program was built <br /> <br /> <br />does not represent the North Fork system. and that a different model would have to be <br /> <br /> <br />developed specifically for bedrock channels. Until then, field monitoring of channels <br /> <br /> <br />remains the most reliable way to quantitY channel recovery, and sediment transport <br /> <br /> <br />modeling supplies a first order approximation to assess the potential effects on channel <br /> <br /> <br />morphology from reservoir sediment releases, especially if long-term predictions must be <br /> <br />made, <br /> <br />13 <br />