41b52d6c4acbc7af78e0d414ccaf3876.ppt

- Количество слайдов: 25

ESEL Chung-buk National University Environment System Engineering Laboratory Identifying of a Pollution Delivery Coefficient for a stream water quality Analysis model introducing a Watershed Form ratio 2007. 6. 7. Ha Sung-Ryong, Park Jung-Ha Department of Urban Engineering Chungbuk National University, Korea

ESEL Chung-buk National University Environment System Engineering Laboratory Contents £ £ £ £ £ Introduction Objective of the research Conventional simple rate method (SRM) Nonlinear Regression Method (NRM) Limitation of Nonlinear Regression Method (NRM) Innovation Process for estimating K – The bypassed load – The leaking load – The weighting factor of flow rate – The characteristic of the nitrogen wash-off Study watershed Result and discussion Conclusions References

ESEL Chung-buk National University Environment System Engineering Laboratory Introduction £ TMDL(Total Maximum Daily Load) in Korea - TMDL has been established for watershed-based water quality management since 1998. - Observation of water quality and flow, researching present condition of the pollution sources, calculation of the discharge pollution load, water quality modeling, checking for approaching a water quality standard, allocation of pollution load. £ Delivery Coefficient, K – relation between discharge and delivery pollution load - Using for environmental volume characteristics of the watershed

ESEL Chung-buk National University Environment System Engineering Laboratory Introduction Conventional simple rate method( Son et al. , 1995) – A general method for calculating delivery coefficient which dividing delivery pollution load by discharge pollution load. – This method can not only reflect variation of environmental volume in the watershed but calculate delivery pollution load for unobserved watershed. £ Method for calculating pollution load reduction coefficient(Ha et al. , 1998) – Monte-carlo simulation £ Nonlinear Regression Method(Ha et al. , 2001, 2003, 2004, 2005) - Deducing nonlinear regression method through relation between observed delivery pollution coefficient of water quality observation points and calculated pollution run-off characteristic coefficient of watershed. - Definition delivery coefficient using watershed form ratio - In case of calculated discharge pollution load of watershed less than observed delivery pollution load, impossible to define nonlinear regression method £ Suggest a technical Innovated method for estimating the pollution delivery coefficient (Park et al. , 2007) – Analyzed the cause of that discharge pollution load less than observed pollution load. And suggested innovation method. £

ESEL Chung-buk National University Environment System Engineering Laboratory Objective of the research £ The objective of this study is to identify what the reason causes the limitation of NRM and suggest how we can purify the process to evaluate a pollution delivery coefficient using many field observed cases

ESEL Chung-buk National University Environment System Engineering Laboratory Delivery Function of Pollution Load £ Conventional simple rate method (SRM) to determine a delivery coefficient, K. PM = P O × K ---- Eq. (1) where PM : the pollution load monitored at a river mouth with enough observed data. PO : the total pollution load discharged from a specific catchment area. (to be available by unit loading factor application)

ESEL Chung-buk National University Environment System Engineering Laboratory Limitation of Conventional SRM Conventional method adapt the similar K to nearby watersheds even though the different circumstances. How can we get K of watershed which has no WQMS? K of watershed which has water quality monitoring station (WQMS) can be calculated using SRM Can it be similar K?

ESEL Chung-buk National University Environment System Engineering Laboratory Nonlinear Regression Method (NRM) PM = P O × K ---- Eq. (1) K = e -Φ ·Sf ---- Eq. (2) where Sf : the watershed form ratio of the watershed Φ : the retention coefficient of the watershed ---- Eq. (3) where, L : a sum of stream length and A : a area of specific watershed. £ Using GIS spatial analysis of GRID DEM (Digital Elevation Model), the L, A and Sf can be derived as a deterministic variable of NRM.

ESEL Chung-buk National University Environment System Engineering Laboratory Nonlinear Regression Method (NRM) Φ = a ( Sf ) b where, a and b : pollution load characteristic function of watershed £ Deducing nonlinear regression method through relation between watershed form ratio and retention coefficient.

ESEL Chung-buk National University Environment System Engineering Laboratory Limitation of Nonlinear Regression Method (NRM) £ £ Calculated discharge pollution load < observed delivery pollution load, Delivery coefficient > 1, Retention coefficient < 0 In this case, can not define linear regression method - retention coefficient of watershed showed negative - retention coefficient of period which large-scale flow showed negative - retention coefficient of T-N compared with the other water quality items showed negative

ESEL Chung-buk National University Environment System Engineering Laboratory Innovation Process for estimating K Nonlinear Regression Method (NRM) retention coefficient, Φ Application of bypassed load Application of leaking load Application of flow weight Application of nitrogen wash-off retention coefficient, Φ Φ≥ 0 Φ<0 Re-checking of observed water quality Delivery coefficient, K-NRM Delivery coefficient, K-Innovation RMSE analysis

ESEL Chung-buk National University Environment System Engineering Laboratory The bypassed load £ The bypassed load at a waste water treatment plant Ph = Pi × bypassed load rate Φ = -ln ( Pm /(Po + Ph ))×Sf Where, Ph : the bypassed load Pi : pollution load transported through pipeline bypassed load ratio : data suggested from Total Water Pollution Load Management Plan for Geum River (Chung-buk province, 2005)

ESEL Chung-buk National University Environment System Engineering Laboratory The leaking load £ The leaking load while pollution load transported through pipeline Pk = total pollution load of sewage catchment area – total inflow pollution load of waste water treatment plant Φ = -ln( Pm /( Po ) / Pk )) × Sf Where, Pk : The leaking load

ESEL Chung-buk National University Environment System Engineering Laboratory The weighting factor of flow rate £ The cause of that observed pollution load of raining season showed high is flowed accumulated pollution load on surface out of dry season at a time. £ thus, apply flow weight for each watershed Flow weight = monthly average flow / yearly average flow Monthly average discharge pollution load = monthly average discharge pollution load before improvement× flow weight

ESEL Chung-buk National University Environment System Engineering Laboratory The characteristic of the nitrogen wash-off £ £ Retention coefficient of T-N water quality showed negative through compared with the other water quality items, it's cause of nitrogen's discharge characteristic. Plenty of nitrogen element might be separated from soil by increased flow at raining season. Nitrogen is closely connected with infiltration water of farmland. we applied base-flow for definition weight of nitrogen which moved by infiltration water. Baseflow rate = monthly baseflow / yearly baseflow Pf = Pi × Baseflow rate Φ = -ln ( Pm /(Po +Pf ))×Sf

ESEL Chung-buk National University Environment System Engineering Laboratory Study watershed: Miho stream and Geum River upstream • Area: about 4, 709 Km 2 • Division of watersheds: 191 (data source: Korea Water Resources Corporation) • 15 WQMS was used to calculate the relationship between Sf and f, base on mean of watershed area.

ESEL Chung-buk National University Environment System Engineering Laboratory Result and discussion £ Comparison of nonlinear regression curve Before improvement After improvement

ESEL Chung-buk National University Environment System Engineering Laboratory Result and discussion £ Comparison of nonlinear regression curve Before improvement After improvement As a result, the critical point of nonlinear regression method(NRM) which retention coefficient showed negative has improved.

ESEL Chung-buk National University Environment System Engineering Laboratory Result and discussion £ Non-exceedance probability of Retention coefficient (Φ) Before improvement After improvement

ESEL Chung-buk National University Environment System Engineering Laboratory Result and discussion £ Non-exceedance probability of Retention coefficient (Φ) Before improvement After improvement Retention coefficient calculated by Improved nonlinear regression method(NRM) has shown stabilized value.

ESEL Chung-buk National University Environment System Engineering Laboratory Result and discussion £ Non-exceedance probability of Delivery coefficient (K) Before improvement After improvement

ESEL Chung-buk National University Environment System Engineering Laboratory Result and discussion £ Non-exceedance probability of Derivery coefficient(K) Before improvement After improvement Retention coefficient calculated by Improved nonlinear regression method(NRM) has shown stabilized value.

ESEL Chung-buk National University Environment System Engineering Laboratory Result and discussion £ RMSE comparison analysis(Delivery coefficient, K) Month BOD T-N T-P before after Jan 0. 051 0. 049 0. 964 0. 952 1. 222 1. 170 Feb 0. 052 0. 044 0. 262 0. 430 0. 237 0. 258 Mar 0. 048 0. 047 1. 340 1. 323 0. 265 0. 268 Apr 0. 103 0. 040 0. 389 0. 328 0. 127 0. 096 May 0. 162 0. 102 0. 160 0. 183 0. 053 0. 038 Jun 0. 262 0. 219 0. 139 0. 266 0. 084 0. 058 Jul 1. 164 0. 916 1. 172 0. 795 0. 240 0. 080 Aug 4. 653 4. 767 4. 734 4. 540 0. 220 0. 081 Sep 0. 482 0. 519 0. 538 0. 666 0. 210 0. 047 Oct 0. 324 0. 342 0. 889 0. 777 0. 353 0. 350 Nov 0. 204 0. 085 0. 396 0. 414 0. 158 0. 161 Dec 0. 190 0. 129 0. 475 0. 382 0. 175 0. 177 RMSE 합 7. 695 7. 260 11. 459 11. 055 3. 344 2. 785 – As a result of analysis, delivery coefficient after improvement has more accurate value than before. – RMSE analysis is compared delivery coefficient calculated by nonlinear regression method with improved nonlinear regression method based on observed delivery coefficient.

ESEL Chung-buk National University Environment System Engineering Laboratory Conclusions This study analyzed limitation of nonlinear regression method(NRM) through application of NRM to real river, and then improved for nonlinear regression method. By improved nonlinear regression method, we could calculate more accurate delivery coefficient to verify that validity. £ £ The limitation of pollution load delivery function calculation method – Discharge pollution load < observed pollution load, – Delivery coefficient > 1, retention coefficient < 0 £ Application of innovated pollution load delivery coefficient calculation method for solving the limitation. – application of the bypassed load, – application of leaking load, application of flow weight, – application of nitrogen’s discharge characteristic weight £ As a result of apply innovated pollution load delivery calculation method to Keum river and Miho stream, we could verify that improved method more accurate than non-improved method through comparing of nonlinear regression curve, non-exceedance probability distribution analysis, RMSE analysis.

ESEL Chung-buk National University Environment System Engineering Laboratory References £ £ £ £ £ 1. 강선홍, 서동일, 1997, 초강천의 퇴적물 분석을 이용한 총인 및 질소의 유달특성평가, 대한상하수도학회지, 11(4), 99 -109. 2. 박대희, 2005, GIS-HSPF 기반 유역 오염유출 해석을 통한 하천 수질오염 영향도 평가, 충북대학교 박사학위 논문. 3. 배명순, 2003, 지형정보학을 이용한 오염부하유출계수 산정, 충북대학교 박사학위 논문. 4. 손명기, 서수용, 서동일, 1995, 금강 하류부의 오염물질 발생현황과 유달율을 고려한 유입오염부하량 산정, 대한환경공학회, 401 -404. 5. 이광호, 하성룡, 1993, 농경유역으로부터의 오염부하 장기유출해석, 한국수자원학회, 501 -508. 6. 이은실, 1998, 하천 수질 보전을 고려한 토지이용 적지분석, 충북대학교 석사학위 논문. 7. 충청북도, 2005, 충청북도 금강 오염총량관리 기본계획. 8. EPA, 1987, The Enhanced Stream Water Quality Models QUAL 2 E. 9. ESRI, 1994, Using ARC/INFO Network/TIN/COGO/Arc. Storm/Arc. Scan. 10. ESRI, 1995, Using GRID with ARC/INFO. 11. Horton, R. E. , 1945, Erosional development of streams and their drainage basins : Hydrophysical approach to quantitative morphology, Geol. Soc. Am. Bull. , 56, 275 -370. 12. S. R. Ha. , D. I. Jung, and C. H. Yoon, 1998, A renovated model for spatial analysis of pollutant runoff loads in agricultural watershed, Water Science and Technology, 38(10), 207 -214. 13. S. R. Ha, M. S. Bae, 2001, Effects of land use and municipal wastewater treatment changes on stream water quality, J of Environmental Monitoring and Assessment System, 70(1 -2), 135 -151. 14. S. R. Ha, M. S. Bae, 2003, Nonlinear regression approach to Evaluate Nutrient Delivery coefficient in Trans. Boundary Watershed with Observation Data Limited, J. of Environmental Science and Engineering, 5, 65 -71. 15. S. R. Ha, M. S. Bae, 2004, D. S. Kong, Nonlinear regression approach to evaluate nutrient delivery coefficient, 8 th International Conference on Diffuse Pollution. 16. S. R. Ha, M. S. Bae, 2005, GIS-based influence analysis of geomorphological properties on pollutant washoff in agricultural area, Water Science and Technology, 51(3), 301 -307. 17. M. S. Bae, S. R. Ha, 2006, Nonlinear regression approach to Evaluate Nutrient Delivery coefficient, Water Science and Technology, 53(2), 271 -279.