preview

Evaluation Of The Model Calibration Results Essay

Better Essays

The model calibration results are summarized in Figure 3. The modeled volumes are compared with monitored volumes for all the 67 events at PSW45 and results are shown in the top plot in Figure 3. The slope of linear regression between modeled and monitored volume is about 0.9 with R square error equal to 0.9 as well, indicating a good fit of modeled results to monitored results from a volume point of view. In addition, only 3 events out of 67 events (4.4%) are in the outlier of 95% prediction interval, which is another indicator for a good fit. Modeled peak flow and modeled time to peak results are also compared with the monitored results at PSW45, shown in Figure 3 respectively. For the peak flow comparison, 4 events out of 67 events (6.0%) are in the outlier of 95% prediction interval, and for the time to peak comparison, only 2 events out of 67 events (3.0%) are in the outlier of 95% prediction interval, both indicating a good fit of modeled results to monitored results. The cumulative frequency distribution (CFD) of modeled volume is compared with that of monitored volume and results are shown in the bottom plot in Figure 3. There is a 3% difference between the total modeled volume and the total monitored volume throughout the entire period from 11/10/2011 to 12/31/2012, which once more shows a good fit for all small, median, and large storm events through the entire monitoring period. After calibration, the baseline model is properly set up and ready to model the impact

Get Access