Abstract Penetration experiments were carried out for a period of over two years at an experimental field located in Kusatsu City. The field was divided into two tanks and each tank filled with different types of soil. Polluted water discharged from the road surface was supplied to the field during storm events. The effluent from the field was composed of Surface and Infiltrated flow. The flow was gauged at three points and water sampled for quality analysis. In order to analyze measures against clogging, soil in one tank was replaced to investigate the recovery of soil contents, penetration efficiency, and purification efficiency. The other tank was researched continuously for 26 months to analyze the change of purification efficiency with precipitation characteristics and the continuous use the facility. Sustainability was analyzed by modeling long-term mass balance in the facilities using water quality data (COD, N, P and SS) and comparing the results with actual measured data of the change of pollutants in the soil. The facility achieved high pollutant removal efficiencies of between 50% - 90% during the survey. Simulation results using water quality data compared well with actual measured data of change of pollutants in soil
Journal of Water and Environment Technology, Vol.3, No.2, 2005 - 263 - EMPIRICAL STUDY ON EFFICIENCY AND SUSTAINABILITY OF SOIL PURIFICATION FACILITY FOR DISCHARGED POLLUTANTS FROM ROAD SURFACE K. Yamada 1 , T. Shiota 1 , Y. Nagaoka 1 , M. Shiono 2 and K. Nishikawa 3 1 Department of Environmental Systems Engineering, Ritsumeikan University, 1-1-1 Nojihigashi, Kusatsu, Shiga, 525-8577, Japan (E-mail: yamada-k@se.ritsumei.ac.jp) 2 Agrex Co. Ltd., STS Bldg., 5-12-8, Nishinakajima, Yodogawa-ku, Osaka-shi, Osaka 532-0011, Japan (E-mail: masayuki_shiono@agrex.co.jp) 3 NJS Consultants Co. Ltd., 9 -15 Kaigan 1 –Chome, Minato-ku, Tokyo 105-0022, Japan (E-mail: kouichi_nishikawa@njs.co.jp) Abstract Penetration experiments were carried out for a period of over two years at an experimental field located in Kusatsu City. The field was divided into two tanks and each tank filled with different types of soil. Polluted water discharged from the road surface was supplied to the field during storm events. The effluent from the field was composed of Surface and Infiltrated flow. The flow was gauged at three points and water sampled for quality analysis. In order to analyze measures against clogging, soil in one tank was replaced to investigate the recovery of soil contents, penetration efficiency, and purification efficiency. The other tank was researched continuously for 26 months to analyze the change of purification efficiency with precipitation characteristics and the continuous use the facility. Sustainability was analyzed by modeling long-term mass balance in the facilities using water quality data (COD, N, P and SS) and comparing the results with actual measured data of the change of pollutants in the soil. The facility achieved high pollutant removal efficiencies of between 50% - 90% during the survey. Simulation results using water quality data compared well with actual measured data of change of pollutants in soil. Keywords Road surface runoff, soil purification facility, model INTRODUCTION Lake Biwa is the largest lake in Japan, and it covers 1/6 of land area of Shiga Prefecture. The water quality of the lake has been leveling off since 1980, and does not meet environmental standards (Figure 1). Thus, despite various measures undertaken to control pollution such as setting of environmental and effluent standards and construction of wastewater treatment plants, the water quality has not improved. This is attributed to diffuse sources of pollution which are a significant source of pollution to the lake (Figure 2). Among diffuse sources of pollution, those from urban areas have increased due to increase in number of impermeable residential areas and surfaces. Road surface runoff contributes a high proportion of this pollutant load and therefore requires appropriate measures. Soil penetration is considered as one of the measures. Figure 1. Trend of water quality of Lake Biwa, Shiga Prefectural Government (2004) 0 0.01 0.02 0.03 1985 1995 2005 T-P(mg/L) Environmental Standard T-P ( 0 1 2 3 4 5 1985 1995 2005 COD(mg/L) Environmental Standard COD ( 0 0.1 0.2 0.3 0.4 0.5 1985 1995 2005 T-N(mg/L) Environmental Standard T-N ( year year year 0 0.01 0.02 0.03 1985 1995 2005 T-P (m g/ L South Lake North Lake Journal of Water and Environment Technology, Vol.3, No.2, 2005 - 264 - Figure 3. Experimental facilities In this study, purification of storm water runoff from road surface using a soil penetration demonstration experiment facility was investigated. The decrease of penetration capacity and purification efficiency during continuous use of the facility was investigated. The facility consisted of two tanks. In one of the tanks, continuous use of the facility and change in purification efficiency depending on precipitation characteristics were analyzed with the results of experiments carried out over a period of 26 months starting in 2001. The soil in the other tank was replaced to consider measures against clogging by analyzing the recovery of soil contents, penetration efficiency and purification efficiency. Sustainability was analyzed by modeling long-term mass balance in the facilities using water quality data (COD, N, P and SS). The calculated results were then compared with actual measured data of the change of pollutants in the soil. This study was a continuation of earlier studies whose results are reported elsewhere (Sugihara et al., 2000; Nishikawa et al., 2001; Nishikawa et al., 2002). Figure 2. Pollutant load going into Lake Biwa (Shiga Prefectural Government, 2004) OUTLINE OF SURVEY Experimental field and soil penetration facility The outline of the soil penetration facility is shown in Figure 3 and Table 1. The facility consisted of two parallel soil tanks; Tank A and Tank B (Figure 3) each with soil of 45 cm depth. Tank A consisted of two types of soil, 5 cm deep Akadama soil in the top layer and 40 cm deep pit sand in the bottom layer. Tank B consisted of only pit sand. Soil replacement was carried out in Tank B in 2003. Runoff water from 750 m 2 of road surface area flows into the soil penetration facility. Inflow water leaves the facility as Overflow or Infiltrated water. The bottom of the facility is covered with a water proof sheet to comprehend the water balance. Water that penetrates through the soil is gathered at the bottom of the facility and drained by a drainage pipe provided at the bottom. This water is referred to here as Infiltrated water. As Inflow water increases, surface flow goes over the weir and is drained. This is called Overflow. Outline of measurement The amount of Inflow water, Infiltrated Natural Urban Industrial Agricultual Land Total amount of Loads (t/day) Water flow Akadama soil Pit sand Points of Soil sampling Points of Flow measurement and Water sampling Soil 5cm 3.5m 3.5m 0.45m Inflow water Infiltrated water Overflow water Weir Weir Overflow water Infiltrated water 7m Inflow Tank A Tank B Soil penetration facility 24.5m 2 ×2 Road Surface 750m 2 T-P T-N 13% 10% 28% 20% 29% 56.1 (t/day) '95 15% 21% 34% 14% 16% 44.7 (t/day) '05(Forecast) 16% 6% 26% 16% 36% 21.6 (t/day) '95 27% 11% 8% 39% 15% 20.1 (t/day) '05(Forecast) 16% 5% 39% 27% 13% 1.35 (t/day) '95 36% 8% 19% 22% 15% 1.08 (t/day) '05(Forecast) COD Table 1. Outline of experimental facility (m 2 ) (cm) (cm) (cm) (cm) Tank A 25 45 3 Akadama sand (5) Pit sand (40) Tank B 25 45 3 Lower layer Soil type Pit sand (45) Surface area Depth Height of weir Upper layer Journal of Water and Environment Technology, Vol.3, No.2, 2005 - 265 - water and Overflow water were measured by a flow meter. In addition, water quality analysis, soil content measurement and water permeability tests were carried out. There were 219 rainfall events over the duration of the experiment. Flow measurements for Inflow water, Infiltrated water and Overflow water were undertaken for all the rainfall events using an automatic flow meter installed on site. Sampling for water quality analysis was done for 14 of the rainfall events (Table 2). SS, T-COD, T-N, T-P, D-COD, D-N and D-P were targeted in water quality analysis. Soil was sampled from both the surface and at a depth of 5cm for soil content analysis. In Tank A, which consisted of two types of soil, only Akadama soil was sampled. Sampling was done at three sites at the upstream, mid-stream and downstream as shown in Figure 3. Soil sampling was done 22 times. Carbon, nitrogen, phosphorus and particle size were targeted in soil content measurement. Sediments were collected from the road surface and analyzed for nitrogen, phosphorus and carbon content. Permeability tests were done at three locations near the points used for soil sampling. To measure permeability, the amount of water that flowed through a pipe installed in the soil over a fixed period of time was determined. Permeability tests were undertaken 25 times (Table 2). Table 2. Characteristics of selected storm events 9/5/01 - ss1 ps1 - - - - - 10/31 - ss13 ps16 - - - - - 9/10 qs1 - - 24.5 18.55 1.32 15.0 2.4 11/1 qs9 - - 33.5 10.50 3.19 6.0 8.1 9/12 - ss2 ps2 - - - - - 11/2 - ss14 - - - - - - 10/9 qs2 ss3 ps3 52.5 14.17 3.71 39.0 5.3 11/4 - - ps17 - - - - - 10/12 - ss4 ps4 - - - - - 2003 11/4 - ss5 ps5 - - - - - 2/25 - ss15 - - - - - - 11/28 - ss6 ps6 - - - - - 3/18 - - ps18 - - - - - 11/29 qs3 - - 7.5 10.33 0.73 12.0 22.4 4/10 - ss16 - - - - - - 11/30 - ss7 ps7 - - - - - 5/12 - - ps19 - - - - - 12/18 - - ps8 - - - - - 5/30 qs10 - - 19.5 8.00 2.44 1.5 4.1 3/5/02 qs4 - ps9 52.0 15.67 3.32 18.0 5.2 6/12 - ss17 - 4.0 3.00 1.33 0.5 1.8 3/7 - ss8 ps10 - - - - - 6/16 qs11 - - 20.0 6.00 3.33 3.0 0.6 6/22 - ss9 ps11 - - - - - 6/30 - - ps20 - - - - - 7/13 qs5 - - 6.0 3.00 2.00 21.0 2.3 7/29 qs12 ss18 - 31.5 17.00 1.85 3.0 5.9 7/23 - ss10 ps12 - - - - - 8/5 - - ps21 - - - - - 8/22 - - ps13 - - - - - 8/24 - ss19 ps22 - - - - - 8/29 - ss11 - - - - - - 9/29 - ss20 - - - - - - 9/6 qs6 - - 14.5 8.83 1.64 15.0 8.6 10/5 - - ps23 - - - - - 9/19 - ss12 ps14 - - - - - 10/13 qs13 - - 15.5 8.67 1.79 2.5 17.5 9/26 qs7 - - 5.5 3.50 1.57 6.0 9.3 10/29 - ss21 ps24 - - - - - 10/15 qs8 - - 5.5 0.50 11.00 18.0 7.2 11/3 qs14 - - 20.5 20.00 1.03 3.5 11.8 10/25 - - ps15 - - - - - 11/18 - ss22 ps25 - - - - - Water quality survey ID Ave. rainfall intensity (mm/hr) Antecedent days (day) Date Precipi tation (mm) Duration (hr) Soil content survey ID Water permeability survey ID Max. rainfall intensity for 10 min (mm/hr） Antecedent days (day) Date Water quality survey ID Soil content survey ID Water permeability survey ID Precipi tation (mm) Duration (hr) Ave. rainfall intensity (mm/hr) Max. rainfall intensity for 10 min (mm/hr) RESULTS AND DISCUSSION Survey results T-N content ratio. Figure 4 shows the results of nitrogen content in the soil. For Tank A the data correspond to the content in the top layer Akadama soil only while for Tank B the data correspond to content in pit sand. The initial nitrogen content of Akadama sand in Tank A was higher than that of pit sand of Tank B. The nitrogen content in both Tanks increased with time. In both Tank A and B, the content of nitrogen in the surface soil leveled off to a value comparable to that of sediment collected from the road surface (3.4 mg/g). The nitrogen content of soil at 5 cm depth remained constant from the start of the survey but begun to rise at a relatively slow rate at about the same time that the nitrogen content in the top soil begun level off. In Tank B, after the soil was replaced (at 645 days) the nitrogen content recovered to almost the same level as that at the initial state. The results of the subsequent measurements showed an increasing tendency in nitrogen content similar to the situation before soil replacement. Journal of Water and Environment Technology, Vol.3, No.2, 2005 - 266 - The above results show that the nitrogen content in the soil at 5 cm did not rise significantly above the initial content at the start of the survey. In other words, the results of the soil content survey showed that most of the nitrogen was stored in the upper 5 cm depth of the soil during 26 months (804 days) Figure 4. T-N content in soil Removal efficiency. For the purpose of evaluating pollutant removal efficiency, Gross Removal Efficiency (GRE) and Net Removal Efficiency (NRE) were defined as follows: GRE = (Pollutant load held in soil)/( Pollutant load of Inflow) (1) NRE = (Pollutant load held in soil)/( Pollutant load of Inflow – Pollutant load of Overflow) (2) GRE and NRE are shown in Figure 6. GRE ranged between 60 – 95% and 20 – 95% for Tank A and Tank B, respectively. Except for the first three initial surveys, the average GRE in Tank A were about 75% for T-COD and T-P, 65% for T-N and 90% for SS. In tank B, though the Overflow ratio (surface Overflow water / Inflow water) was comparatively high, the corresponding GRE values were about 60% for T-COD, 60% for T-P, 50% for T-N and 85% for SS before soil replacement. The GRE values after soil replacement showed almost similar values as those before soil replacement, contrary to the expectation that the removal efficiency would improve with soil replacement. In particular, for SS the average removal efficiency after soil replacement decreased to nearly 60%. On the other hand, NRE was higher than GRE by about 10% in Tank A, 20% in tank B before soil replacement and 5-20% in Tank B after soil replacement. The relatively higher NRE for Tank B compared to Tank A is due to the fact that the Overflow ratio was relatively higher in Tank B. As with GRE, the NRE in Tank B after soil replacement was lower than that before soil replacement. This could be attributed to the fact that the soil was unstable after replacement and therefore some of the soil may have flowed out together with infiltrated water. However, the removal efficiencies after soil replacement show an increasing tendency with time. In particular, the removal efficiencies for the last two surveys after soil replacement (survey No. wps13 and wps14) show a relatively high increase, indicating that the soil may have stabilized. There is need to further study the recovery of purification capacity. 0 1 2 3 4 5 0 200 400 600 800 Days Content Ratio(mg/g) Tank A Days Nitrogen content (mg / g) 0 1 2 3 4 5 0 200 400 600 800 Days Content Ratio(mg/g Tank B Nitrogen content (mg / g) Surface 5cm DepthSurface (after soil replacement) 5cm Depth (after soil re placement) Journal of Water and Environment Technology, Vol.3, No.2, 2005 - 267 - Figure 6. Pollutant removal efficiency in Tank A (upper graphs) and Tank B (lower graphs) Simulation of long-term mass balance Flow. A conceptual diagram of mass balance in the soil treatment facility is shown in Figure 7. Simulation of long-term mass balance is considered for Tank A only where continuous survey was undertaken. The effect of rainfall characteristics on the mass balance was investigated by classifying the rainfall events into 7 types according to the average rainfall intensity and the precipitation (Table 3). Figure 7. Conceptual diagram of mass balance Road surface runoff (Ro) was calculated as follows: Ro = rr × (R - Iw) (3) Table 3. Classification of rainfall Net removal efficiency in soil Gross removal efficiency COD T-N T-P 0 20 40 60 80 100 wqs1 wqs2 wqs3 wqs4 wqs5 wqs6 wqs7 wqs8 wqs9 wqs10 wqs11 wqs12 wqs13 wqs14 Analysis No. 0 20 40 60 80 100 wqs1 wqs2 wqs3 wqs4 wqs5 wqs6 wqs7 wqs8 wqs9 wqs10 wqs11 wqs12 wqs13 wqs14 Analysis No. 0 20 40 60 80 100 wqs1 wqs2 wqs3 wqs4 wqs5 wqs6 wqs7 wqs8 wqs9 wqs10 wqs11 wqs12 wqs13 0 20 40 60 80 100 wqs1 wqs2 wqs3 wqs4 wqs5 wqs6 wqs7 wqs8 wqs9 wqs10 wqs11 wqs12 wqs13 ( ( 0 20 40 60 80 100 wqs1 wqs2 wqs3 wqs4 wqs5 wqs6 wqs7 wqs8 wqs9 wqs10 wqs11 wqs12 wqs13 Purification rate(%) 0 20 40 60 80 100 wqs1 wqs2 wqs3 wqs4 wqs5 wqs6 wqs7 wqs8 wqs9 wqs10 wqs11 wqs12 wqs13 wqs14 Analysis No. Purification soil(%) Removal Efficiency (%) Removal Efficiency (%) R O LO Ro LRo I LI (Rainfall) (Road surface runoff) (Pollutant load of road surface runoff) (Inflow water) (Pollutant load of inflow) (Overflow water) (Pollutant load of Overflow) D LD S LS F LF (Infiltrated flow) (Pollutant load of infiltrated flow) (Water held in soil) (Pollutant load of water held in soil) Infiltrated water Pollutant load of infiltrated water Road SurfaceRoad Surface Soil Field F=I-O LF=LI-LO S=F-D LS=LF-LD Rainfall t ype Precipitation R (mm) Ave. rainfall intensity I (mm/hr） A R ＜3.0 I ＜1.0 B C D 3.0≦R ＜20.0 I ＜1.0 1.0≦I ＜4.0 4.0≦I E F G 20.0≦R I ＜1.0 1.0≦I ＜4.0 4.0≦I Journal of Water and Environment Technology, Vol.3, No.2, 2005 - 268 - In Equation (3) rr is Road surface runoff ratio, R is precipitation (L) and Iw is Initial lost water (L). Initial lost water was calculated by the amount of initial rainfall and the measured value of Road surface runoff. According to Equation (1), rr was calculated as the slope of the regression equation of measured Road surface runoff and the difference between precipitation and Initial lost water. Typical calculation results for rr for rainfall of type A are shown in Figure 8. The calculated values of runoff ratios are given in Table 4. The relationship between values of Ro estimated by Equation (3) and measured values is shown in Figure 9. The gross quantity of Flow into soil (F) and Water held in soil (S) for one rainfall event were calculated from measured values of gross Inflow water (I), Overflow water (O) and Infiltrated water (D), as follows (all units in liters): F = I – O = I × (1 – ro) (4) S = F – D (5) In Equation (5), ro is Overflow runoff ratio of flow. The Overflow runoff ration of flow is related to Inflow water, cumulative rainfall, antecedent days of no rain and average of rainfall intensity. Multiple linear regression analysis with these four factors set as explanatory variables was used to calculate ro as follows (Figures 10 and 11): ro= ( ) ( ) ( ) ( ) ( ) 2 4 3 3 2 2 5 1 5 103.7104.6101.2104.9102.1 −−−−− ×−+×−+×+×+× xxxx (6) In Equation (6), x 1 is Inflow water (L), x 2 is cumulative rainfall (mm), x 3 is the number of antecedent days of no rain (days), and x 4 is average of rainfall intensity (mm/hr). R 2 = 0.16 R 2 = 0.42 R 2 = 0.32 0.0 0.2 0.4 0.6 0.8 1.0 0102030 Ave. rainfall intensity(mm/hr ） overflow runoff Ratio(% Overflow runoff ratio, ro 0 5000 10000 15000 0 5000 10000 15000 Actual measurement （ L ） Estimated （ L ） Figure 9. Estimated and measured Road surface runoff Figure 8. Road surface runoff and precipitation (rainfall type A) Table 4. Road surface runoff ratio, rr y = 0.6632x R 2 = 0.8183 0 500 1000 1500 0 500 1000 1500 Precipitation reduced initial los t （ L ） Road surface runoff （ L ） Road surface runoff (L) Precipitation reduced by initial lost water (L) R 2 = 0.04 R 2 = 0.05 R 2 = 0.06 0.0 0.2 0.4 0.6 0.8 1.0 0102030 Antecedent days (day) overflow runoff Ratio(% × ～1000mm (cumulative rainfall) 2000mm～ △ 1000mm～2000mm Antecedent days (days) Ave. rainfall intensity (mm/hr) Figure 10. Overflow runoff ratio and antecedent days of no rain Figure 11. Overflow runoff ratio and rainfall intensity Rainfall type Runoff ratio, rr A 0.663 B 0.795 C 0.808 D 0.904 E- F 0.889 G 0.916 Estimated (g) Overflow runoff ratio, ro Journal of Water and Environment Technology, Vol.3, No.2, 2005 - 269 - Figure 15. Overflow runoff ratio of load and flow Figure 16. Estimated and measured values of pollution load (T-N) The relationship between gross Flow into Soil (F) and gross Infiltrated water (D) is shown in Figure 12, and the regression equation is given in below: D = 0. 783 × F-184.4 (7) The relationship between values estimated by the model and values measured by water quality survey and flow rate survey is shown in Figure 13. Figure 12. Infiltrated water and Flow into soil Mass balance of pollution load. Pollution load was calculated by the relationship between Road surface runoff (Ro) and Pollutant load from road surface runoff (LRo). As an example, the following relationship was established for T-N (Figure 14): LRo= 3785.0 5833.0 Ro× (8) The amount of pollutant load in overflow was calculated by the regression analysis of Overflow runoff ratio of flow (ro) and Overflow runoff ratio of load (Lro). The following relationship was established for T-N (Figure 15): Lro=0.6059 × ro (9) T-N R 2 = 0.5291 0 10 20 30 40 50 0 10000 20000 30000 40000 Surface runoff, Ro （ L) loading dose ： LRo(g) Pollutant load, LRo (g) 0 5000 10000 15000 20000 0 5000 10000 15000 20000 measured value （ L ） Estimated vslues （ L ） Estimated value (L) R 2 = 0.9451 0 5000 10000 15000 0 5000 10000 15000 20000 volume of Flow into s oil ： F （Ｌ） Infiltrated water volume ： D （Ｌ） Figure 14. Pollutant load and Road surface runoff Figure 13. Estimated and measured values of flow Overflow runoff ratio (flow), ro (%) Overflow runoff ratio (load) : Lro (%) R 2 = 0.966 R 2 = 0.939 R 2 = 0.8874 R 2 = 0.9284 0 10 20 30 40 50 0 20406080 △ T-P × SS T-COD T-N 0 5 10 15 20 0 5 10 15 20 measured load(g) Estimated load(g Estimated load (g) Measured load (g) × Infiltrated water △ Water held in soil Overflow water Flow into soil Measured value (L) Infiltrated water volume, D (L) Flow into soil, F (L) Overflow runoff ratio (load), Lro (%) × Pollutant load infiltrated △ Pollutant load held in soil Pollutant load in overflow Pollutant load into soil Journal of Water and Environment Technology, Vol.3, No.2, 2005 - 270 - Measured values of Pollutant load into soil (LS) and Pollutant load infiltrated (LD) were calculated from water quality analysis data. The relationship between the measured values and values estimated by the model equations above is shown in Figure 16. Comparison between simulation results based on water quality data and actual measurement data of pollutant load in soil. Gross water flow and gross pollutant load of road surface runoff, overflow water, flow into soil, infiltrated water and water held in soil were calculated for all the precipitation data (total precipitation 3,300 mm) from September 2001 to November 2003. The simulation results of gross water balance and gross pollutant load balance are shown in Tables 5 and 6, respectively. From the results of soil content survey before and after the simulation term, the contents of pollutants at 5 cm-depth and surface soil were calculated, assuming the following values: specific gravity as 2.65, void ratio as 50% and proportion of particles under 75 µm as about 35%. Comparisons of increase in soil pollutant content amounts are shown in Table 6. Simulation results of (LS) compared well with measured data. Table 5. Simulation results of gross water balance (mm) I O F D S Tank A 2768 936 1832 1335 498 Table 6. Simulation results of gross pollutant load balance and increased soil contents (g) T-N P-N T-P P-P Pollutant load of inflow water (LI) 1699.5 511.9 179.8 123.1 Pollutant load of overflow water (LO) 228.4 69.1 21.4 11.8 Model simulation Pollutant load of flow into soil (LF) 1471.1 442.8 158.3 111.3 Pollutant load of infiltrated water (LD) 423.6 57.0 20.6 9.2 Pollutant load of water held in soil (LS) 1047.9 385.8 137.8 102.2 Actual measurement Increase in pollutant amount in soil 1006.4 145.8 CONCLUSIONS Sustainability of a soil penetration facility for storm water runoff was analyzed by carrying out a mass balance analysis using data measured from surveys of permeability, flow rate and water quality. The simulation results were compared with actual measured data of pollutant content in soil. The following results were obtained: 1. Nitrogen accumulated in the upper 5 cm layer of soil in of the soil purification facility. 2. Average pollutant removal efficiencies of 50% - 90% were obtained for COD, N, P and SS. An obvious decline in pollutant removal efficiency was not seen in the facility over the two years the survey was carried out. However, further studies are needed to investigate recovery of the pollutant removal efficiency after soil replacement. 3. Mass balances of flow and pollutant load in the penetration facility were carried out. Regression equations relating flow and pollutant load to rainfall characteristics were established. REFERENCES Sugihara, M., Yamada, K., and Terada, A. (2000). Study on soil filtration treatment for pollutants from road surface during storm events. Proceedings, 3rd Annual Symposium, JSWE, 95-96. Nishikawa, K., Sugihara, M. and Yamada, K. (2001). Study on soil penetration facility for pollutants discharged from road surface. Proceedings, 56th Annual Meeting, JSCE, 488-489. Nishikawa, K., Sugihara, M., and Yamada, K. (2002). Study on removal by soil of pollutants discharged from road surface during storm events. Proceedings, 6th International Conference on Diffuse Pollution, IWA, 616-621. Shiga Prefectural Government (2004). http://www.pref.shiga.jp/index.html . 6.00 3. 33 3.0 0.6 6/22 - ss9 ps11 - - - - - 6 /30 - - ps20 - - - - - 7/ 13 qs5 - - 6.0 3. 00 2.00 21.0 2 .3 7/29 qs12 ss18 - 31 .5 17.00 1.85 3. 0 5.9 7/ 23 -. - - ps8 - - - - - 5 /30 qs10 - - 19.5 8.00 2.44 1.5 4.1 3/ 5/02 qs4 - ps9 52.0 15.67 3. 32 18.0 5.2 6/12 - ss17 - 4.0 3. 00 1 .33 0.5 1.8 3/ 7 - ss8 ps10 - -