Frequency of Routine and Flooding-stage Observations for Precise Annual Total Pollutant Loads and their Estimating Method in the Yodo River

9 572 0
Frequency of Routine and Flooding-stage Observations for Precise Annual Total Pollutant Loads and their Estimating Method in the Yodo River

Đang tải... (xem toàn văn)

Thông tin tài liệu

ABSTRACT The Yodo River, which has two big tributaries and an annual mean flow of 268m3/s, flows out of Lake Biwa, runs through the Amagase Dam and flows into Osaka Bay. Point and non-point pollutant loads in Kyoto City and its adjacent cities discharge into the river. The frequency of our routine observation (once every three days) during the period from late April to late November 2003-2005 was ten times that of official routine water quality monitoring in Japan. We also carried flooding-stage observations at three transverse points across the river during the routine observation period in 2005. The total pollutant loads during five flooding events were larger than those observed by the routine observation during the periods. The mean pollutant loads during observation periods of every three days, every six days, every nine days and semimonthly data were different from those derived from monthly data. The annual total loads of runoff pollutant in 2003-2005 were estimated by extrapolating to both sides of the observation period and by compensating during flooding-stages using regression equations between flow and pollutant loadings.

Journal of Water and Environment Technology, Vol. 6, No.2, 2008 - 93 - Frequency of Routine and Flooding-stage Observations for Precise Annual Total Pollutant Loads and their Estimating Method in the Yodo River Senichi EBISE * , Hironori KAWAMURA ** *Department of Civil and Environmental System Engineering, Setsunan University, Neyagawa, Osaka, 572-8508, Japan **Technical Division, Kyowa Co. Ltd., Senba, Chuo-ku, Osaka, 542-0081, Japan, ABSTRACT The Yodo River, which has two big tributaries and an annual mean flow of 268m 3 /s, flows out of Lake Biwa, runs through the Amagase Dam and flows into Osaka Bay. Point and non-point pollutant loads in Kyoto City and its adjacent cities discharge into the river. The frequency of our routine observation (once every three days) during the period from late April to late November 2003-2005 was ten times that of official routine water quality monitoring in Japan. We also carried flooding-stage observations at three transverse points across the river during the routine observation period in 2005. The total pollutant loads during five flooding events were larger than those observed by the routine observation during the periods. The mean pollutant loads during observation periods of every three days, every six days, every nine days and semimonthly data were different from those derived from monthly data. The annual total loads of runoff pollutant in 2003-2005 were estimated by extrapolating to both sides of the observation period and by compensating during flooding-stages using regression equations between flow and pollutant loadings. Keywoods: annual total load, flooding-stage observation, routine observation, pollutant, nutrient INTRODUCTION The Yodo River, which flows out from the Lake Biwa (surface area: 674 km 2 ; watershed: 3714 km 2 ), is one of the largest river in Japan, with a watershed of 7214 km 2 . It runs through the Amagase Dam and is then joined by two large tributaries, the R. Kidu (1596 km 2 ) and R. Katsura (1100 km 2 ), inflowing at 36 km upstream of the river mouth at Osaka Bay in the Seto Inland Sea (Fig. 1). The Yodo River has an annual mean flow, 268 m 3 /s and a ratio of maximum to minimum flow of 114, the flow being relatively stable flow (Ebise et al. (1993, 2002)). Point and non-point pollution sources in Kyoto City (population: 1.47 million) and its adjacent cities discharge into the Yodo River and its tributaries. In the lower reaches of the river, 80 m 3 /s is taken as a raw water for domestic and industrial use by the residents and factories in Osaka City (population: 2.64 million), Kobe City ((population: 1.53 million), and its satellite cities. Address correspondence to Senichi Ebise, Department of Civil and Environmental System Engineering, Setsunan University, Email: ebise@civ.setsunan.ac.jp , Received October 29, 2008, Accepted December 3, 2008 Journal of Water and Environment Technology, Vol. 6, No.2, 2008 - 94 - Furthermore, the Yodo River delivers the greatest loads of pollutants and nutrients into Osaka Bay (Database, (2004)). Therefore, it is necessary to estimate precisely the annual total loads of pollutants in the Yodo River and their runoff characteristics (Ebise et al. (2001, 2003, 2004 and 2006)). Here, pollutants contain organic matter, nutrients, SS and other measured substances. Fig. 1 - Map of the Yodo River and its tributaries. METHODS AND MATERIALS Sampling points Over the last ten years, we have been carrying out observations of the Yodo River every third day from three points on the Yodo River New Bridge (nearby our university campus) during the period from late April to late November. We paid special consideration to the frequency of our routine observation and the transverse distribution of water quality in broad area of flow (about 240 m). The frequency of our observations (once every three days) was ten times that of official routine water quality monitoring program in Japan. This frequency means that the observation day of the week shifts successively, thus, allowing us to observe changes in patterns of factory production and lifestyle as point pollution sources in the catchment. Especially, with regard to non-point pollution sources over one year, we also carried out five flooding-stage observations during the routine observation period to estimate more precisely the annual total loads of runoff pollutants in 2005 (Ebise et al. (2001, 2003, 2004), Inoue et al. (2002)). Three transverse sampling points at the Yodo River New Bridge were set as each center of trisections in the flowing water width. Analytical Methods After measuring the water temperature, electrical conductivity, DO and turbidity of the three samples individually at the observation point and bringing the samples back to our laboratory, we Journal of Water and Environment Technology, Vol. 6, No.2, 2008 - 95 - combined them into a composite sample and analyzed inorganic ions and organic matter. The water samples taken at the three points spanning the river during the five flooding-stages were analyzed individually, and the representative concentrations were determined as the arithmetical means for the three points. The inorganic ions (Cl - , NO 3 - , NO 2 - , SO 4 2- , NH 4 + , Na + , K + , Mg 2+ and Ca 2+ ) in the filtered sample (obtained using 1-µm pore size glass-fiber filter) were measured by ion chromatography (Yokogawa, IC-7000). TOC (all components) and DOC (dissolved components were analyzed using by a TOC meter (Shimadzu TOC-5000A). COD (T-COD (all components)) and D-COD (dissolved components) were also analyzed by a titration method with KMnO 4 after standing the sample in a hot water-bath for 30 minutes under acid conditions. COD is one of the environmental standard items analyzed in marine areas in Japan. Suspended solids (SS) was measured using the above-mentioned glass-fiber filter, and chlorophyll-a was analyzed by the SCOR/UNESCO method (SCOR/UNESCO, (1966)) with 0.45-µm pore filter. POC and P-COD were estimated as (TOC-DOC) and (T-COD-D-COD) respectively. OBSERVATION EVERY THREE DAYS Our observations conducted every three days in 2005 (total 75 times) were done to estimate the annual total pollutant loads during observation periods and to clarify the optimum frequency of routine observation for precise estimation of annual total pollutant loads. The flow in the both sides of the routine observation period in 2005, mostly winter, was stable. Fig. 2 shows the changes in TOC and DOC concentrations with the change of flow in 2005, and Fig. 3 shows the changes in TOC and T-COD loadings with the change of flow. Both parameters were depended mainly on the flow. 0 1 2 3 4 5 6 4/21 4/27 5/3 5/9 5/15 5/21 5/27 6/2 6/8 6/14 6/20 6/26 7/2 7/8 7/14 7/20 7/26 8/1 8/7 8/13 8/19 8/25 8/31 9/6 9/12 9/18 9/24 9/30 10/6 10/12 10/18 10/24 10/30 11/5 11/11 11/17 11/23 11/29 Concentraition (mg・l -1 ) 0 200 400 600 800 1000 1200 1400 Flow (m 3 ・s -1 ) TOC DOC Flow Fig. 2 - Changes in TOC and T-COD concentrations (2005). Journal of Water and Environment Technology, Vol. 6, No.2, 2008 - 96 - 0 250 500 750 1000 1250 1500 4/21 4/27 5/3 5/9 5/15 5/21 5/27 6/2 6/8 6/14 6/20 6/26 7/2 7/8 7/14 7/20 7/26 8/1 8/7 8/13 8/19 8/25 8/31 9/6 9/12 9/18 9/24 9/30 10/6 10/12 10/18 10/24 10/30 11/5 11/11 11/17 11/23 11/29 loading(g・s -1 ) 0 200 400 600 800 1000 1200 1400 Flow(m・s -1 ) T-COD TOC Flow (7/5 ) T-COD 3283g・ s -1 TOC 3614g・ s -1 (7/8 ) T-COD 2057g・ s -1 TOC 2380g・ s -1 (7/1 1) TOC 1711g・ s -1 (9/9) T-COD 2576g・s -1 Fig. 3 - Changes in TOC and T-COD loadings (2005). On the basis of the data obtained by routine observation every three days, we were also able to use data sets for observations every six days, every nine days, semimonthly or monthly. We decided to take their observation times larger as possible in all data sets, and so treated their first observation time as the first observation data in the observation every three days. If each pollutant loading at one observation day should be assumed to change linearly to the loading at next observation day, each pollutant load during the interval of both observation days could be calculated by a formula of trapezoid area. Fig. 4 shows the mean loads of runoff pollutants during these observation periods and it can be seen that the total pollutant loads became smaller than those estimated by observation every three days. The highest mean pollutant loads derived from monthly observation data were different from those derived from more frequent observations. Fig. 4 - Mean pollutant loads during the observation period by five routine observations (2005). 0 10 20 30 40 50 60 70 80 90 100 T-COD D-COD TOC DOC Cl (× 1/10) NO3-N NH4-N Mean Load (kg/d) 3 days 6 days 9 days 15 days 30 days Flow(m 3 ・s -1 ) Journal of Water and Environment Technology, Vol. 6, No.2, 2008 - 97 - FLOODING-STAGE OBSERVATIONS In Japan, there are many heavy storm events and storm runoff events which bring increases in pollutant loads during flooding-stages in large river catchments. As the traverse water mixing in the wide Yodo River was relatively weak during a flooding stage. Therefore, we conducted continuous observations at three transverse points over the river during five periods of flooding in the routine observation period. The frequency of flooding-stage observations was set close interval before the flow peak and thereafter long interval. Such continuous flooding-stage observations had never been conducted previously for the Yodo River. The R. Katsura in the western upstream catchment and the R. Kidu in the eastern upstream catchment flow into the Yodo River nearly 14 km upstream from the Yodo River New Bridge. Moreover, the urban discharge of small tributaries close to cities located upstream flows into the Yodo River from both sides. Fig. 5 shows changes in TOC concentrations at the three transverse points with the change of flow during one of the five flooding events. The data reflected the differences in traveling time of river water from their main pollutant sources and the lack of uniformity of rainfall distribution in the broad catchment. It was particularly evident that the values and peak times of pollutant concentrations during the flooding events differed among the three transverse sampling points. In this broad river, the degree of transverse water mixing is small, as the waterway tends not to meander. 0 1 2 3 4 5 0 24 48 72 96 120 144 168 192 216 240 264 288 Concetration(mg・l -1 ) 0 200 400 600 800 1000 1200 1400 1600 Flow(m3・s -1 ) Left Center Right Flow 6/29 9:00 7/1 9:00 7/3 9:00 7/5 9:00 7/9 9:00 7/7 9:00 7/11 9:00 Fig. 5 - Changes in TOC concentrations at three points during a flooding event (2005). Flow(m 3 ・s -1 ) Journal of Water and Environment Technology, Vol. 6, No.2, 2008 - 98 - Table 1 - Comparison of total pollutant loads by routine and flooding-stage observations during five flooding-stage periods (2005), (Load: 10 6 kg (Chl-a: 10 3 kg), Flow: 10 6 m 3 ) Observation Flow T-COD D-COD TOC DOC SS Chl-a Cl - NO 3 - -N SO 4 2- PO 4 3- -P Na + Flooding-stage Obs. 1158 2.9 2.0 3.2 2.6 20.6 6.6 12.4 0.7 12.5 0.07 11.1 Routine Observation 947 2.3 1.5 2.5 2.0 11.4 4.4 10.9 0.6 12.2 0.08 9.9 During the two big flooding-stages, cooperative water discharge occurred through opening of the weir at Lake Biwa and the gates of the Amagase Dam. Theses discharge brought high pollutant loads and a high flow level in relatively long periods. The total loads of most pollutants recorded during the five flooding-stage observation periods became larger than those during only routine observations during the flooding periods (shown in Table1). In particular, peak times of flow and pollutant concentrations were occasionally not encountered during routine observations every three days. The increase of runoff pollutant loads was due to non-point pollution sources mainly in the urban and rural catchment. TOTAL RUNOFF POLLUTANT LOADS DURING OBSERVATION PERIOD Observations were conducted every three days during 2003 and 2004 in the same way. Fig. 6 shows a comparison of total pollutant loads during the 2003-2005 observation periods. For the Yodo River, 2005 was hydrologically a dry year (Q: 218m 3 /s), 2004 was a rainy year (Q: 434m 3 /s) and 2003 was an average year (Q: 236m 3 /s). During the observation periods, years with more rainfall had larger mean pollutant loads than years with less rainfall. The influence of yearly changes of pollution sources in the Yodo River was ignored for these three consecutive years. Therefore, we also recognized that non-point pollution sources occupied a larger proportion among all pollution sources in the Yodo River catchment. Fig. 6 - Comparison of total pollutant loads in the observation periods during 2003-2005. T-COD D-COD TOC DOC Na Cl NO3 SO4 Load(kg・d -1 ) 2003 2004 2005 1×10 5 2×10 5 3×10 5 4×10 5 5×10 5 6×10 5 7×10 5 0 Load(kg・d -1 ) Journal of Water and Environment Technology, Vol. 6, No.2, 2008 - 99 - The total loads of runoff pollutants during the observation period could be evaluated precisely by considering the transverse distribution of water quality concentrations and storm runoff loads. The total pollutant loads during the observation period in the dry year of 2005 and the average year of 2003 became smaller than those in the rainy year of 2004. Therefore, a precise estimation of annual total pollutant loads, even for a large stable-flow river, is required using both frequent routine observations and continuous observations during flooding events throughout the year. On this basis, we propose an estimation method using regression equations between flow and pollutant loadings. ESTIMATION OF TOTAL POLLUTANT LOADS USING REGRESSION EQUATIONS Estimation of the annual total pollutant loads is important for selecting measure to prevent water pollution in the Yodo River and Osaka Bay. Flow data were obtained every day from the relationship between flow and water-level. The annual total loads of runoff pollutants in 2003-2005, which flowed into Osaka Bay, were estimated by extrapolating to both sides of the observation periods using the regression equations between flow (Q) and pollutant loadings (L), L=a・Q n with high regression coefficients (shown in Table 2). As both sides of the observation periods during 2003-2005 were periods of stable and low flow in winter, the annual total pollutant loads could be estimated more precisely by the regression equations. Table 2 - Regression equations of pollutant loadings (regression coefficient). Empirically, if about five flooding event data (including two biggest flooding events) were obtained in a year, the reliability of the regression equations should become better (Ebise, (1990)). The total pollutant loads during the five flooding events in 2005 were larger than those observed by the routine observation. We applied the regression equations to estimate a precise annual total pollutant loads by extrapolating to both sides of the routine observation period and by compensating during flooding-stages using regression equations. Especially, in the case of an excess of 268 m 3 /s (mean flow) over the flow (Q>268 m 3 /s) and an excess of 120% over the previous day’s flow (Q>120%), compensation of pollutant loadings during the periods of flooding was done for one year. The compensated values of pollutant load during flooding became larger in the dry year of 2005, than in other years, and became the largest for TOC. The values for inorganic ion loads became smaller than those for organic matters because 2003 2.3184・Q 1.0237 (0.957) 2.3545・Q 0.9756 (0.960) 5.8022・Q 0.9695 (0.966) 3.1102・Q 0.9575 (0.974) 66.704・Q 0.6915 (0.941) 60.053・Q 0.7181 (0.957) 2004 0.2412・Q 1.3946 (0.834) 1.9175・Q 1.1014 (0.806) 4.5064・Q 0.9825 (0.888) 1.3812・Q 1.0613 (0.935) 177.97・Q 0.5422 (0.812) 134.07・Q 0.6025 (0.896) 2005 1.7510・Q 1.0712 (0.996) 2.3140・Q 0.9673 (0.997) 3.9714・Q 0.9655 (0.992) 2.7186・Q 0.9565 (0.999) 76.062・Q 0.6797 (0.986) 67.030・Q 0.7170 (0.997) Flooding-stage(2005) 1.5145・Q 1.1076 (0.952) 2.8183・Q 0.9424 (0.956) 3.6554・Q 0.9827 (0.976) 1.4946・Q 1.0904 (0.984) 85.376・Q 0.6637 (0.925) 74.094・Q 0.7042 (0.950) DOC Na + Cl - L=a・Q n L=a・Q n L=a・Q n Year T-COD D-COD TOC L=a・Q n L=a・Q n L=a・Q n Journal of Water and Environment Technology, Vol. 6, No.2, 2008 - 100 - of the small values of n in the regression equations. There were only small values between the two cases of Q>268 m 3 /s) and Q>120%, as shown in Fig. 7. The compensated values in the case of Q >268 m 3 /s became a little larger than those in the case of Q>120% for most pollutant loads. When the routine observation caught a high pollutant loading during flooding periods, the compensated value became small as the case of TOC in Fig.7. 0 20 40 60 80 100 120 140 160 T-COD D-COD TOC DOC Loading(kg/day) Control Q> 268m3/s Q>120% Figure 7 - Compensation of pollutant loads in annual mean loads during 2005. CONCLUSION We conducted observations every three days at three transverse points across the Yodo River during the period from late April to late November in 2003-2005. The observations were also conducted during five flooding events in 2005. The total pollutant loads during the five flooding events were larger than those observed only during the routine observations. The frequency of water quality monitoring was used to compare mean pollutant loads during periods of every three days, every six days and every nine days, semimonthly and monthly. As the monthly data indicated the largest frequency among the data for other periods, precise estimation of annual total pollutant loads in the Yodo River must be based on observation data excluding the monthly data. Total pollutant loads during routine observations in the dry year of 2005 became small, whereas those obtained in a rainy year became larger. The annual total pollutant loads were estimated using regression equations between flow and pollutant loadings with high regression coefficients. This method will enable precise estimation of annual total pollutant loads for previous years and other rivers based on these data and regression equations. 268m 3 /s Journal of Water and Environment Technology, Vol. 6, No.2, 2008 - 101 - REFERENCES Database (2004). Osaka Bay environmental database, Kinki Regional Development Bureau, MLIT, Japan, http://kouwan.pa.kkr.mlit.go.jp/kankyo-db/deta/b3_04ryu_1.html . Ebise S. (1990) . Regression models for estimating storm runoff pollutant loads, Proc. of Environmental and Sanitary Research, (JSCE), 20, 27-38. Ebise, S., Inoue, T. and Numabe, A. (1993). Runoff characteristics and observation methods of pesticides and nutrients in rural rivers, Water Science & Technology, 28, 589-593. Ebise, S. and Miki, K. (2001). Evaluation on runoff characteristics of heavy metals in the Yodo River and its tributary streams by high frequency observations, J. of Japan Society on Water Environment, 24, 715-723. Ebise, S., Inoue, T. and Numabe, A. (2002). Runoff characteristics of pesticides from paddy fields and reduction of risk to the aquatic environment, Water Science & Technology, 45, 127-131. Ebise, S., Fukushima, K. and Oike, N., (2003). Evaluation of runoff characteristics and risks of pesticides in the Yodo River and its tributaries, J. of Japan Society on Water Environment, 26, 699-706. Ebise, S., Oike, N. and Fukushima, K. (2004). Statistical analyses on runoff characteristics of dissolved heavy metals in the Yodo River and its tributaries, Environmental Science, 17, 49-59. Ebise, S., and Kawamura, H. (2006). Evaluation of runoff pesticides by high frequency observation and flooding-stage observations in Yodo River, J. of Japan Society on Water Environment, 29, 705-713. Inoue, T., Ebise, S. and Numabe, A. (2002). Runoff characteristics of particulate pesticides in a river from paddy fields, Water Science & Technology, 45, 121-126. SCOR/UNESCO (1966) SCOR/UNESCO Working Group on Photosynthetic Pigment, Monographs on Oceanographic Methodology, No. 1, P.69, (published UNESCO). . 2 3 4 5 6 4/21 4/27 5/3 5/9 5/15 5/21 5/27 6/ 2 6/ 8 6/ 14 6/ 20 6/ 26 7/2 7/8 7/14 7/20 7/ 26 8/1 8/7 8/13 8/19 8/25 8/31 9 /6 9/12 9/18 9/24 9/30 10 /6 10/12. 2.3545・Q 0.97 56 (0. 960 ) 5.8022・Q 0. 969 5 (0. 966 ) 3.1102・Q 0.9575 (0.974) 66 .704・Q 0 .69 15 (0.941) 60 .053・Q 0.7181 (0.957) 2004 0.2412・Q 1.39 46 (0.834) 1.9175・Q

Ngày đăng: 05/09/2013, 09:38

Từ khóa liên quan

Tài liệu cùng người dùng

  • Đang cập nhật ...

Tài liệu liên quan