Laserfiche WebLink
discharges ranging from 100 cfs to around 400 cfs. Calibration measure- <br />ments were made at about 150 cfs. Average velocities for each channel <br />segment were calculated for each of the five flows and compared with the <br />measured values. From this data an analysis of the frequency and magni- <br />tude of errors could be conducted. The percentage error was calculated <br />as an absolute value by the following equation: <br />Error (%) = I-v X 100% (19) <br />?m <br />where, <br />V = the measured average velocity of the subsection <br />p = the predicted average velocity of the subsection <br />The number and percentage of predicted velocities within specified <br />percentage error bounds are shown in Table 3. The small number of <br />observations used in this trial suggest that further field testing of <br />this method is warranted. <br />Bovee, Gore, and S91verman (1977) conducted a limited field test of <br />the equation v. = a. Q i in the same stream reach for which Elser's <br />field test was conducted. In this sense, conditions for the two tests <br />were identical and, therefore, results from the tests are comparable. <br />Velocity predictions were made using a two-point rating curve - <br />system-and were applied to. specific locations on a transect, rather than <br />for average velocities for a channel segment. Velocity predictions were <br />made for flows ranging from 140 cfs to about 400 cfs. A total of 71 <br />velocity measurements and predictions are included in the analysis shown <br />in Table 4. Error calculations and presentations were made by the same <br />procedures used with Elser's data. <br />Both the above field tests were conducted during the developmental <br />phase of both approaches for instream flow use. Velocity predictions <br />for both studies were made in a highly turbulent, non-uniform riffle <br />area. Results could be expected to give higher accuracy in pools and <br />runs, although these areas have not been subjected to a field test. <br />RECOMMENDATIONS <br />The preferred approach in predicting either the stage-discharge <br />relationship or the velocity distribution would be to use the rating <br />curve approach for each type of prediction, using at least three points <br />which span the flows of interest. Outwardly, it might seem that this <br />approach requires an unreasonable amount of time and manpower. However, <br />it is a rare instream flow study that is conducted with only one trip to <br />the field. It is much more common to conduct numerous field trips to a <br />particular stream, for one reason or another. In this case, collection <br />of additional field data is a relatively small imposition. The cost of <br />three sets of data is small compared to the cost of one set which may <br />prove to be unreliable. <br />?9