| Summary: | 碩士 === 國立臺北科技大學 === 土木與防災研究所 === 96 === In recent years, the Best Management Practices (BMPs) for non-point source pollution control are gaining in popularity in Taiwan. The constructed wetland system, one of the BMPs, possesses several advantages in many aspects, such as low cost, low energy consumption and no secondary pollutants, as well as provides ecological and landscape benefits. There are a lot of BMPs simulation models such as AGNPS, HSPF, MUSIC, SWMM, and VAST which are capable of simulating continuous rainfalls or single storm event. The above models have been applied in domestic watershed management for years. However, the above models cannot evaluate the removal efficiency of the structured-BMPs. Instead, the BMP Tool Box, a BMP module, is suitable for evaluating the arrangement and locations of several BMPs. However, the BMP Tool Box is still in its development stage, and further research is needed to establish the domestic parameters and evaluation techniques which are better suited for domestic applications.
There were some research using BMP Tool Box in the past, but few of them covered complete sets of model calibration and validation as well as domestic parameters. The present research is a case study of a combined constructed wetlands system, which was built near Donan Bridge and located on the bank of Peishih Creek by Taipei Feitsui Reservoir Administration in 2005. This research applies the BMP Tool Box to do model calibration and validation based on the long-term sampled data collected on site. The sensitivity analysis was also conducted and further evaluation of the BMPs removal efficiency with respect to different storm events developed by the so-called Alternating Block Method was also performed. Based on the long-term sampled data, the average removal efficiency for total phosphorus (TP) was calculated to be 64.4%.
Then, five site data of the TPs in 2006 were used to calibrate the simulation model; meanwhile, fourteen site data of the TPs were used to validate the simulation model. The results showed that the coefficients of the determination (R2) for the model calibration and validation of the inflow concentration were found to be 0.84 and 0.86, respectively. Moreover, the R2’s for the model calibration and validation of the outflow concentration were found to be 0.87 and 0.8, respectively. The above figures close to the ideal cases as far as the model simulation are concerned.
Furthermore, ten design storm events developed by the Alternating Block Method with recurrence interval 1.1 year, 2 year, 5 years, 10 years and 25 years combining with 2 and 4 hours of duration were used to simulate the removal efficiency of the TPs. The results showed that the best removal efficiency of TPs was found to be 76％ with recurrence interval of 25 years and 4 hours of duration. This research also found that the removal efficiency will be tremendously affected by construction and maintenance activities. Besides, the accuracy of the simulation will be much improved by continuous sampling. This research also helps build up the domestic design parameters for the other
reservoirs in Taiwan.
|
|---|
