Six Multi-Soil-Layering systems with combinations of two type of soils, Andisol (volcanic ash soil) and Entisol (granitic sandy soil), two particle size classes of zeolite, 1-3 and 3-5mm in diameter, and various hydraulic loading rates, 1 to 4 m3 m-2 day-1, were constructed and were compared for polluted river water treatment. Removal rates of BOD, T-N and T-P were 72.2 to 83.5 %, 22.4 to 50.5 %, and 51.9 to 66.8 %, respectively. Entisol was more efficient for T-N and T-P removal than Andisol. Zeolite of 1-3mm in diameter was more efficient for removal of all the pollutants than that of 3-5mm. These materials might contribute more effective contact of inlet water to the various components for purification in the MSL systems. Lower loading rate was more efficient for BOD and T-N removal. The loading rate of 2 m3 m-2 day-1 was the most efficient for T-P removal. However, differences in removal rates of these pollutants of river water among the MSL systems were small considering with the differences of hydraulic loading rates. The difference in BOD removal rates between MSL 1 and 6 was only 10 % against to 400 % difference in the loading rates between them. Although clogging was observed at the loading rate of 4 m3 m-2 day-1, the system was recovered after 2 months rest. These results indicated that higher loading rate, 4 m3 m-2 day-1, was more advantageous and practical for direct river water treatment
- 97 - DIRECT TREATMENT OF POLLUTED RIVER WATER BY THE MULTI-SOIL-LAYERING METHOD Tsuigiyuki Masunaga*, Kuniaki Sato*, Takayuki Zennami**, Syunitsu Fujii** and Toshiyuki Wakatsuki* *Faculty of Life and Environmental Science, Shimane University, Matsue 690-8504, Japan **Fujii Consulting Engineer and Associates, 1349 Higashitsuda, Matsue 690-0011, Japan ABSTRACT Six Multi-Soil-Layering systems with combinations of two type of soils, Andisol (volcanic ash soil) and Entisol (granitic sandy soil), two particle size classes of zeolite, 1-3 and 3-5mm in diameter, and various hydraulic loading rates, 1 to 4 m 3 m -2 day -1 , were constructed and were compared for polluted river water treatment. Removal rates of BOD, T-N and T-P were 72.2 to 83.5 %, 22.4 to 50.5 %, and 51.9 to 66.8 %, respectively. Entisol was more efficient for T-N and T-P removal than Andisol. Zeolite of 1-3mm in diameter was more efficient for removal of all the pollutants than that of 3-5mm. These materials might contribute more effective contact of inlet water to the various components for purification in the MSL systems. Lower loading rate was more efficient for BOD and T-N removal. The loading rate of 2 m 3 m -2 day -1 was the most efficient for T-P r emov al. However, differences in removal rates of these pollutants of river water among the MSL systems were small considering with the differences of hydraulic loading rates. The difference in BOD removal rates between MSL 1 and 6 was only 10 % against to 400 % difference in the loading rates between them. Although clogging was observed at the loading rate of 4 m 3 m -2 day -1 , the system was recovered after 2 months rest. These results indicated that higher loading rate, 4 m 3 m -2 day -1 , was more advantageous and practical for direct river water treatment. KEYWORDS Andisol; direct river water treatment; Entisol; Multi-Soil Layering (MSL) method; zeolite INTRODUCTION High performance of Multi-Soil Layering (MSL) method on water purification was exhibited in various studies on on-site domestic wastewater treatment (Luanmanee et al. 2001, Wakatsuki et al. 1993, 1998), cafeteria wastewater treatment (Attanandana et al. 2000) and laboratory scale experiments of high speed treatment at a hydraulic loading rate of 2 m 3 m -2 day -1 (Masunaga et al. 1998). The principle of MSL method was control and enhancement of purification function of soil resources by means of its structure (Fig.1a) and addition of functional materials (Wakatsuki et al. 1998, 1999). Several advantages were recognized in MSL method: (1) simultaneous removal of organic materials such as BOD & COD, nitrogen and phosphorous from wastewater; (2) prevention the system from clogging which was the limitation of traditional water treatment system by soil such as soil trench method and land treatment system; (3) much higher loading rate, 1000 to 4000 l m -2 day -1 , than traditional soil trench system, 10 to 40 l m -2 day -1 (Perkins, 1989); (4) MSL system can be composed from locally available resources (Attanandana et al. 2000, Luanmanee et al. 2002). From these advantages, the MSL was appeared to be applicable for on-site treatment of wastewater from households, restaurants and so on in developed and developing countries. However, polluted public water environments such as river, lake, pond and so on have to be treated at the same time in order to ameliorate water environmental quality in total. In this study, application of the MSL method to direct treatment of polluted river water was discussed in terms of its performance and adaptability to improve the public water environment. - 98 - MATERIALS AND METHODS Description of MSL system and experimental condition The MSL systems were constructed at Uya River in which 600 m 3 water flow a day in average, in Oki Island, Shimane prefecture, Japan. About 1400 households are located along the river and have discharged gray water directly and toilet water with imperfect treatment to the river, which caused pollution of the river. Figure 1. (a) Structure of MSL system and (b) Water treatment flow. Table 1. Size of zeolite, type of soils and hydraulic loading rate for each MSL system. Size of zeolite Type of soil Hydraulic loading rate (m 3 m -2 day -1 ) system (mm) Jan.1999- July 2000- Oct. 2000- MSL 1 1-3 Andisol 1 2.7 4 MSL 2 1-3 Andisol2 2.7stopped MSL 3 1-3 Entisol 1 2.7 4 MSL 4 1-3 Entisol 2 2.7 4 MSL 5 3-5 Andisol2 2.7stopped MSL 6 3-5 Andisol 4 2.7 4 Figure 1a shows the structure of an on-site MSL system used for this study. The MSL system was composed of soil mixture layers and zeolite. Six sets of MSL systems were constructed and operated with six different conditions (Table 1). Two types of soils: Andisol, humus rich volcanic ash soil, and Entisol, humus poor granitic sandy soil (MASA soil), and two sizes of zeolite with its diameter of 1-3 and 3-5mm, were used for comparison of efficiencies of materials. Soil mixture layers composed of soil (Andisol or Entisol), saw dust, granular iron and charcoal at the weight ratios of 67.5 (78.3), 11.25 (7.5), 11.25 (7.5) and 10 (6.7)%, respectively (values in parentheses in case using Entisol). Soil mixture layers were packed into jute bags for easy construction and were placed 10 cm thickness being arranged in a brick pattern surrounded by zeolite layers. An aeration pipe (2.5 cm in diameter) was installed between the second and the third soil mixture layers. Two porous pipes for inlet of river water were installed on the surface of zeolite layer. Hydraulic loading rate was fixed from 1 to 4 m 3 m -2 day -1 for each MSL system from Jan 1999 to July 2000, and then it was changed as shown in Table 1. From October 2000 to March 2001, four MSL systems were selected to be applied the hydraulic loading rate of 4 m 3 m -2 day -1 since the capacity of the inlet pump was limited to apply it to all the six MSL systems. River water was pumped up to a sedimentation tank with its volume of 0.4 m 3 whose mean retention time was 0.4 to 0.5 hour, then it was pored into a storage tank and was distributed into the six MSL systems by water pumps (Fig. 1b). River water discharged into the MSL systems was moved down gravitationally and treated water was returned to the lower position than the water uptake position of the river. The MSL systems were operated with no aeration from January 1999 to March 2001. P P 10 10 10 10 10 10 10 10 5 10 10 70 70 20 70 20 2035 35 180cm 105 cm 10 85cm - 99 - Sampling and Analyses River water, pre-treated water by sedimentation tank and treated water were collected at around one month interval for analyses of following properties: pH by glass electrode method, ORP, BOD and suspended solids (SS) by Standard Method for the Examination of Water and Wastewater (1995), COD by potassium dichromate method, NH 4 -N by Nessler method, NO 2 -N by diazotization method, NO 3 -N and T-N (after potassium peroxodisulfate digestion) by cadmium reduction method, PO 4 -P and total T-P (after potassium peroxodisulfate digestion) by ascorbic acid method. In addition to above regular sampling for the determination of the seasonal fluctuation of river water and treated water, we determined the daily fluctuation of them by those sampling at 6 hour interval in 7th to 8th November 1999. Soil and zeolite were sampled from the MSL systems after finishing the experiments, and their chemical compositions are being analyzed. Results of total carbon and nitrogen (T-C, T-N) were shown in this paper as a preliminary study on purification mechanisms of MSL system for river water treatment. RESULTS AND DISCUSSION Fluctuation of river and treated water quality Figure 2. Fluctuation of BOD, T-N and T-P concentration (mg l -1 ) in river and treated water. a - c: daily fluctuation in 7th to 8th Nov. 1999; d - f: seasonal fluctuation (R.W: river water, P.W: pretreated water by sedimentation tank) e (T-N) d (BOD) 51.3 20 16 12 8 4 0 10 8 6 4 2 0 f (T-P) 1.76 1.6 1.2 0.8 0.4 0 a (BOD) 0 4 8 12 16 b 0 1 2 3 4 5 6 7 (T-N) 0 0.1 0.2 0.3 18 0 6 12 c (T-P) R.W P.W 123456 1 '99 35 7 911 1 '00 35 7 911 ( ) 37.5 56.7 1 '01 3 ( ) - 100 - Figs. 2a to 2c show the daily fluctuation of concentration of BOD, T-N and T-P in river and treated water. As the source of BOD in river water was households, its concentration seemed to fluctuate synchronizing the daily life cycle of peoples with 1-2 hrs time lag. That is, the highest concentration of BOD was observed in 0:00 when the water discharged from kitchen and bathroom flowed in river, and its lowest concentration was observed in 6:00 when very few water was discharged from houses. In spite of such fluctuation of BOD in river water, the BOD concentration of treated water was very small in all the MSL systems. Fluctuation of T-N in river water also showed similar trend to BOD, therefore it was thought to be affected mainly by daily life cycle. T-N concentration of treated water, however, fluctuated rather than that of BOD, except that of MSL 1 and 3 which were loaded river water at the minimum rate of 1 m 3 m -2 day -1 . Pattern of fluctuation of T-P concentration in river water was different from that in BOD and T-N. MSL 3 and 4 showed similar pattern of pre-treated water. Factors other than daily life cycle might affect it. The pattern in treated water was different among the MSL systems. However, T-P concentration in both river and treated water was very low comparing to BOD and T-N. Pattern of daily fluctuation of BOD and T-N was considered basically same throughout year, however those concentration ranges, including T-P, differed as shown in Fig.2d-f. Fig. 2d to 2f show the seasonal fluctuation of BOD, T-N and T-P concentrations in river water and treated water from January 1999 to November 2000. Since BOD concentration was influenced by the peoples’ life cycle as shown in Fig.2, water samplings were done at the time as same as possible, i.e. at around noon. In spite of the similar sampling time, BOD concentration in river water fluctuated very much. BOD concentration in treated water, however, fluctuated rather small to the daily fluctuation. Pattern of T-N fluctuation in river water was similar to that of BOD, but not for that in treated water. There was some difference in the fluctuation patterns between river water and pre-treated water by sedimentation tank, which was partly due to the organic sediments accumulated in the tank. Such sediments might occasionally release nitrogen and increase its concentration in loading water. Fluctuation of T-N in treated water showed similar pattern of that in pre-treated water, and was more fluctuated than that of BOD. Nitrogen discharged into MSL systems mainly consisted of NH 4 -N, NO 3 -N and organic nitrogen that considered being the rest of T-N minus inorganic nitrogen such as NH 4 -N, NO 3 -N and NO 2 -N (Table 2). Considering the data in Table 2, some of the organic nitrogen in river water was removed and transformed into NH 4 -N, and then NH 4 -N was mostly removed by its adsorption on zeolite and by its nitrification into NO 3 -N with a contribution of high nitrification ability of natural soil. NO 3 -N was, however, not fully removed from water, which caused high fluctuation of T-N in treated water. In terms of T-P, although concentration in treated water seemed to be affected by that in river water, T-P concentration in treated water changed in the range around 0.1 to 0.4 for all the MSL systems, except for in July 2000. For the other parameters, COD ORP and SS in both river and treated water followed the pattern of BOD in river water. NH 4 -N, NO 2 -N and NO 3 -N in river water followed the pattern of T-N, but those in treated water did not fluctuate so much. NO 3 -N seemed to affect pH. pH was a little bit high when NO 3 -N concentration was low, and vice versa. In all the parameters, the magnitude of those fluctuation was generally smaller in treated water than in river water throughout the operational periods, showing MSL possessed high stability in water treatment performance against to changes of river water quality or environmental condition such as temperature. Such stable performance was also observed in our previous study of on-site domestic wastewater treatment by the MSL method (Wakatsuki et al. 1993). This on-site MSL system has been operated since 1990, and it still showed stable performance even in 10th year. In addition to that, the MSL system that was even loaded at 4 m 3 m -2 day -1 showed stable performance in river water treatment in this study. This was owing to natures of soils such as high buffering capacity to chemical and/or physical condition change which affect activities of microorganisms and physical reactions contributing water treatment mechanisms in the MSL system. For example, pH change caused by inlet water quality or nitrification of ammonium can be easily buffered by exchange of H + and other cations on soil and zeolite surface, and temperature change was far smaller inside of the system than its surface, ex. 7 to 13 o C and –2 to 20 o C, respectively, in 22nd to 23rd March 1999. This was the remarkable advantage of the MSL system. Treatment efficiency of MSL systems in relation to component and hydraulic loading rate Table 2 shows mean water qualities and removal rates (%) of selected parameters during the periods that - 101 - different hydrologic loading rates were applied, from January 1999 to July 2000. For pH and ORP, no significant difference was observed among river water and treated water of MSL systems. Sedimentation tank was not effective to reduce all the pollutants contained in river water because of its small retention time around 20 to 30 minutes. Removal efficiency of BOD was higher in MSL systems with lower hydrologic loading rate, compared the each set of MSL systems, i.e. MSL 1 and 2, 3 and 4, 5 and 6, and was higher in those with smaller size of zeolite, 1-3 mm, compared MSL 2 and 4 to MSL 5. Although soil type did not influence the BOD removal rate, it did COD removal rate. Entisol seemed more suitable for COD removal. Andisol might release some nonbiodegradable organic materials such as humus, which reduced the COD removal efficiency at a condition of high loading rate in this study. In contrast to BOD and COD, SS removal rate was higher at a loading rate of 2 m 3 m -2 day -1 than at that of 1 m 3 m -2 day -1 as shown in the results of MSL 2, 4 and 5. The same trend was observed in our previous experiment with loading rates from 0.5 to 2 m 3 m -2 day -1 treating wastewater with SS of 28 mg l -1 in average (Masunaga et al. 1998). This phenomenon might be because SS accumulation near the surface of the system worked as a filter. However, higher loading rate for MSL6, 4 m 3 m -2 day -1 , caused a clogging of the system at the 13th month. Table 2. Mean river and treated water qualities, and removal rates of selected parameters from Jan.‘99 to July ‘00. pH ORP BOD 5 CODcr SS NH 4 -N NO 2 -N NO 3 -N T-N T-P (mV) (mg l -1 ) R.W 6.99 359 13.3 26.0 27.4 1.19 0.18 2.5 5.4 0.53 P.W 6 . 91 3 6 3 11 .2 (19.3) 26.4 (-3.5) 29.1 (-20.2) 0.93 0.19 2.4 5.4 (-5.4) 0.44 (13.0) MSL1 7.05 383 2.1 (83.5) 6.2 (78.4) 10.6 (72.1) 0.09 0.02 2.2 3.1 (39.0) 0.22 (55.0) MSL2 7.02 389 2.4 (80.6) 7.0 (76.0) 4.3 (84.0) 0.09 0.02 2.6 3.7 (26.1) 0.19 (60.7) MSL3 7.23 373 2.5 (79.9) 5.0 (84.9) 6.7 (77.8) 0.07 0.02 1.6 2.5 (50.5) 0.19 (56.9) MSL4 7.10 383 2.6 (78.3) 6.2 (80.0) 5.7 (82.9) 0.08 0.02 2.2 3.2 (36.6) 0.15 (66.8) MSL5 6.95 387 3.0 (75.4) 8.1 (71.2) 4.9 (79.6) 0.09 0.03 2.4 3.8 (24.2) 0.25 (52.7) MSL6 7.02 389 3.4 (72.2) 10.0 (62.7) 5.9 (74.5) 0.10 0.03 2.5 3.9 (22.4) 0.22 (51.9) Removal rates are shown in parentheses. R.W: river water, P.W: pre-treated water by sedimentation tank In terms of nitrogen treatment, all the MSL systems could treat NH 4 -N and NO 2 -N very efficiently. Mean NO 3 -N concentration was lower in MSL systems with lower loading rates, especially MSL used. There might be several factors affecting NO 3 -N removal, i.e. denitrification. Possible factors were longer retention time caused by lower loading rate, more efficient contact of inlet water to soil mixture layer brought by higher water permeability of Entisol. Hydraulic conductivity of Entisol was estimated to be 10 times larger than that of Andisol. Further study is needed to analyze and discuss those factors. Removal of T-N was less effective than that of the other parameters, which was probably due to excess aerobic condition in the MSL systems shown in ORP values near 400 mV. Open top of the MSL systems allowed air enter the inside of the systems and made aerobic condition. In addition, high dissolved oxygen concentration in river water also contributed to keep the systems aerobic. Practical way to improve T-N removal is to enhance the contact of polluted water and soil mixture layer whose inside was anaerobic. Difference of T-P removal rates in MSL systems showed similar trend to that of SS. Possible mechanism of phosphorus removal is that metal iron in soil mixture layer was oxidized to ferrous ion, which was transferred to zeolite layer and oxidized further to ferric ion and deposited on zeolite particles as hydroxides which could fix phosphate ion as well as iron and aluminum hydroxides as a component of clay in soil fix it (Wakatsuki et al., 1993). In this process, soil mixture layer was required to be in anaerobic condition to some extent for ferrous ion transport to zeolite layer. Loading rate of 2 m 3 m -2 day -1 might be appropriate to attain such condition, especially for the MSL systems used Entisol. Smaller size of zeolite also contributed higher T-P removal by means of contact area expansion of inlet water and hydroxides iron on the surface of zeolite. With regards to the removal rate of the pollutants in Table 2, the differences among the MSL systems were small considering with the difference of hydraulic loading rates. For example, difference in BOD removal rates between MSL 1 and 6 was only 10 % against to 400 % difference in the loading rates between them. This meant higher loading rate was more advantageous in terms of treatment speed or saving space of the system, which are related to a treatment cost. Although data were not shown in this paper, performance of 4 MSL systems with a hydrologic loading rate of 4 m 3 m -2 day -1 was examined from October 2000 to March 2001. Treatment performance of MSL6 was - 102 - basically the same with the results in Table 2. Removal efficiencies of T-N and T-P somewhat declined for MSL1, 3 and 4 because of the higher loading rate. However, those of BOD and SS were the same level with the lower loading rate applied before July 2000. MSL3 and 4 efficiently removed COD at the removal rates more than 70%, even though loading rate was increased. This result indicates that again Entisol was better soil component to remove COD than Andisol. Although removal efficiencies of T-N and T-P was declined for MSL1, 3 and 4, general performance of those systems was better than that of MSL6. It was because contact of inlet water and materials inside the system and filtering ability of SS in inlet water were improved by using smaller zeolite for those MSL systems. However, those MSL1, 3 and 4 started to cause clogging in the surface of the systems in January, but not for MSL6. It can be said that courser size of zeolite, 3-5mm in diameter in this study, is better component for direct treatment of polluted river water with regards to stable system operation reducing the frequency of clogging of the system. T- C a nd T-N contents in MSL systems after 2 years operation Figure 3 shows the T-C and T-N contents in soil mixture layers and zeolite layers at each depth of MSL1, 3 and 6. Since MSL6 was discharged the river water 4 times as much as MSL1 and 3 did, T-C content in MSL6 was higher than those in MSL1 and 3. This was the result of the accumulation of organic pollutants. However, T-N was not accumulated in MSL6. The level of T-N content was kept in original soil T-N levels in both Andisol and Entisol. MSL3 showed lower T-N content that the others. T-C contents tended to be lower in soil mixture layers near to surface of the system for all MSL. Oxygen supply from the top surface of the system supported high decomposition ability in soil mixture layers near the surface. In contrast, T-N contents tended to be lower in lower position of soil mixture layers. This was due to efficient denitrification in lower soil mixture layers where soil was kept as anaerobic. Difference of the treatment mechanisms of organic matters such as BOD and COD, and nitrogen, differed the distribution of T-C and T-N in soil mixture layers. While distribution patterns of T-C and T-N in soil mixture layers were different, those of T-C and T-N were same in zeolite layers. In addition, T-C and T-N contents in zeolite layers were not much different among the systems, even though amount of pollutants discharged into systems different between MSL6 and the others. Both T-C and T-N contents were higher in upper position of zeolite, which was due to an accumulation of organic pollutants and biofilm in upper part of the systems that was well observed during the zeolite sampling. Absorption of NH 4 might also contribute high T-N content in the upper part. These results indicated soil mixture layer strongly contributed the treatment mechanisms of filtering, decomposition and several chemical/biological reactions being changed its reduced/oxidized (and/or aerobic/anaerobic) condition at different depth, while zeolite mainly contributed as filtering medium. In addition to these, there are another treatment mechanisms to be investigated such as phosphorous removal. These will be discussed when all the soil and zeolite samples are analyzed. Figure 3. T-C and T-N contents in soil mixture layers and zeolite layers at each depth of MSL1, 3 and 6. Depths were the distances from the top surface of zeolite layer. MSL1 MSL3 MSL6 3 69 T-C 0 0.2 0.4 T- 0 0.2 0.4 0 0.05 0.1 T-C T- 0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-85 10-20 30-40 50-60 70-80 - 103 - CONCLUSION The results of this study revealed that MSL system could effectively work in polluted river water treatment with some advantages such as high stability of treatment ability against to climate and river water quality changes, and high treatment abilities at loading rate of 4 m 3 m -2 day -1 . The MSL system could remove nitrogen and phosphorous more efficiently than the methods currently often used in Japan such as course gravel, charcoal or various types of plastic modules. And it can be applicable to land limited area where wetland or reed bed system could not be applied (Lantzke et al. 1999). In addition, it was found an endocrine disrupting substance, 17 -Estradiol, could be also effectively removed from river water (Masunaga et al., 2001). In terms of clogging, although the MSL 6 caused it at 13th month, it was recovered by stopping the water loading for only two months. Although no clogging was observed for the other systems using smaller size of zeolite during the period with hydrologic loading rates lower than 2 m 3 m -2 day -1 , they caused clogging within 5 months after the loading rate was increased to 4 m 3 m -2 day -1 . From the results on clogging and on treatment performance in this study, it can be said that desirable treatment condition of polluted river water is treatment with higher loading rate by MSL system consisting of courser zeolite (3-5mm) and Entisol. In terms of clogging, it can be prohibited by operation planning like alternative use of two sets of the systems or settling the rest period of the system during the season in which water is less polluted. This procedure contribute to make treatment more stable. In soil types, although Entisol was found to be better for this time, the other soil types were also found to be usable with an improvement of those physical and chemical characters. Finally, although adaptability of MSL system to direct treatment of polluted river water was studied as a technique for remediation of public water environments in this study, MSL system can be modified (optimized) and be applied for other type of wastewater or polluted water by changing of its material composition as examined by Luanmanee et al. (2002) and Attanandana et al. (2000), and its structure. Based on this experiment, we tried a lager scale of on-site experiment at more polluted river water treatment with improvements of MSL system in Iizuka-city, Fukuoka, and the results showed better performance (Unno et al., 2003). REFERENCES Standard Method for the Examination of Water and Wastewater (1995). 19th edn, American Public Health Association/American Water Works Association/Water Environment Federation, Washington DC, USA. Attanandana T., Saitthiti B., Thongpae S., Kritapirom S., Luanmanee S. and Wakatsuki T. (2000). Multi- Soil-Layering systems for food service wastewater treatment. Ecological Engineering, 15, 133-138. Perkins R.J. (1989). Onsite Wastewater Disposal, Lewins Pub., pp.251. Lantzke I.R., Mitchell D.S., Heritage A.D. and Sharma K.P. (1999). A model of factors controlling orthophosphate removal in planted vertical flow wetlands. Ecological Engineering, 12, 93-105. Luanmanee S., Boonsook P., Attanandana T. and Wakatsuki T. (2002). Effect of organic components and aeration on the efficiency of a Multi-Soil-Layering system for domestic wastewater treatment. Soil Sci. Plant Nutr., 48, 125-134. Luanmanee S., Attanandana T. Masunaga T. and Wakatsuki T. (2001). The efficiency of a multi-soil-layering system on domestic wastewater treatment during the ninth and tenth years of operation. Ecological Engineering, 18, 185-199. Masunaga T., Wakatsuki T., Shirahama M. and Inaishi T. (1998). High quality water remediation by the Multi Soil Layering method. In Proceeding of the 12th International Conference on Chemistry for Protection of the Environment, Z. Cao and L. Pawlowski (eds.), Nanjing Univ. Press, pp. 303-309. Masunaga T., Sato K., Aoyama C., Zennami T., Fujii S. Nobe K. and Wakatsuki T. (2001). Development of Technology for Water Environment Remediation by Multi-Soil-Layering Method, - Direct Treatment of River Water – YOUSUI TO HAISUI, 43, 1033-1040. (in Japanese) Masunaga T., Sato K. and Wakatsuki T. (2002). Removal of Simazin, Fenitrothion, Napropamid and tetrachloroethylene by Multi-Soil-Layering Method. J. Jpn Soc. Water Env., 25, 361-366. (in Japanese - 104 - with English abstract) Unno S., Iyota K., Masunaga T. and Wakatsuki T. (2003). Study on direct treatment of river by Multi-Soil- Layering method and its characteristics of water purification. J Jpn Soc. Civil Eng., Division 2. (in press). (in Japanese with English abstract) Wakatsuki T., Esumi H. and Omura S. (1993). High performance and N & P - removable on-site domestic waste water treatment system by Multi-Soil- Layering method, Wat. Sci. Tech., 27(1), 31-40. Wakatsuki T., Masunaga T., Kita K. Shirahama M., Zennami T. and Harada T. (1998). Wastewater treatment using soil ecosystems – Pedosphere ecological engineering. YOUSUI TO HAISUI, Water and Wastewater, 40, 874-882. (in Japanese) Wakatsuki T., Mori J., Masunaga T. and Kudou K. (1999). Wastewater treatment based on soil ecosystems using Multi-Soil-Layering method. KANKYOU GIJUTSU, Environmental Technology, 28, 806-813. (in Japanese) . with no aeration from January 19 99 to March 20 01. P P 10 10 10 10 10 10 10 10 5 10 10 70 70 20 70 20 2035 35 18 0cm 10 5 cm 10 85cm - 99 - Sampling and Analyses. (BOD) 51. 3 20 16 12 8 4 0 10 8 6 4 2 0 f (T-P) 1. 76 1. 6 1. 2 0.8 0.4 0 a (BOD) 0 4 8 12 16 b 0 1 2 3 4 5 6 7 (T-N) 0 0 .1 0.2 0.3 18 0 6 12 c (T-P) R.W P.W 12 3456