Laserfiche WebLink
<br />:~y ~ <br />",.... <br />i:."~, <br />'. <br />"t' <br />it, <br />t <br />" <br /> <br />~. <br /> <br />. <br />1,; <br /> <br />.;".. <br />;'~.~ . <br />~ <br />:".(.\1 <br />'~"- <br />;;. <br />'!~' <br />:..~ <br /> <br />c <br /> <br />It can be seen that, while peak flow was predicted with relatively good accuracy, <br />total peak volume was over.-estimated. <br /> <br />CONCLUSIONS <br /> <br />CUHP and the computer version CUHPF/PC can be calibrated to <br />accurately predict hydrologic conditions. Frequency analysis using AMS data for <br />the drainage basin proved to be very useful in determining the appropriate <br />variables for calibrating the model. Most variables describing a drainage basin <br />can be easily determined by observation and careful examination. As would be <br />expected of any hydrologic forecasting model, CUHP was most sensitive to <br />effective percent imperviousness and infiltration rates. The infiltration rate is a <br />function of the type of soil in the basin and antecedent moisture conditions (either <br />from precipitation or irrigation). Denver is very arid and irrigation rates are <br />generally matched by evaporation rates, leaving very little antecedent moisture. <br />Soil types can vary dramatically within even a small watershed, making <br />classification of infiltration rates based on soil types exceedingly difficult. In <br />general, this study found that the use of Horton's equation produced more <br />accurate results than the use of constant infiltration rates. In the end result, the <br />difference between SCS soil types had less impact on the hydrologic outcome <br />than did the percent effective imperviousness. This is due, in part, because of <br />the intense nature of Denver rainstorms over a relatively short period of time. <br />This phenomenon tends to lessen the impact of varying infiltration rates. <br />One troubling aspect of the CUHP calibration was the fact that the model <br />could not adequately forecast both the short return period events (2- and 5-year) <br />and the long return period events (50- and 1 DO-year), simultaneously. The model <br />could only be calibrated to predict one or the other (see graph 4). While this <br />investigation focused on the short return periods, the failure of the model to <br />produce consistent results across a large precipitation range implies a problem <br />with the hydrologic algorithms. The dynamics of frequent rainfall events have <br />been studied in great detail. However, the unpredictability and randomness of <br />large storm events (Le. the 1 DO-year storm) make practical real-time analysis <br />nearly impossible. It is quite possible that the underlying theory of infiltration <br />rates (i.e. Horton's equation), retention behavior and flow mechanics have a very <br />different relationship under the dynamic flood conditions of the 1 DO-year storm. It <br />is certainly not uncommon, in all disciplines of science, for analytical tools to <br />have inherent difficulties measu,ring or predicting things both very small and very <br />large, simultaneously. <br /> <br />The calibration of the CUHP model is not a practical exercise for many <br />real life circumstances. When data is available calibration is the best way to get <br />accurate results from the CUHP program. Unfortunately, data is more frequently <br /> <br />Application and Evaluation of CUHP <br /> <br />Page 24 of 52 <br />