5.1 Tidal patterns and physico-chemical water quality in multi-branched urban estuaries
Intensive short-term time scale field observations were conducted in the Arakawa, Sumida and Shakujii River estuaries. The unique hydrodynamic features, characteristics and phenomena of the multi-branched estuaries during a rainfall event (November 11, 2016), neap tide (July 3, 2017) and spring tide (July 11, 2017) were analyzed. A long term monitoring of water quality and scum coverage was also conducted (April to October 2018 and August to October 2017) to determine the effect of the estuary condition variations to scum generation and odor. The rainfall event showed that the Shakujii River, even with smaller catchment area, had a higher net SS transport than Sumida River. Salinity retention was evident in the bottom depth of Shakujii River, even with high rainfall discharge, due to topography. The neap and spring tide results showed that the Sumida River and Arakawa River followed the typical natural estuary trends on salinity, SSC, DO and velocity. The Shakujii River estuary on the other hand followed a typical artificial urban estuary with lower flow velocity during spring tide than neap tide and has hypoxic bottom depths with high SSC during spring tide. These unique characteristics in the Shakujii River are attributed to the channel slope that does not allow transit of water to upstream even during high tide.
The results from the rainfall and neap-spring field observations were further validated in the statistical analysis of the long term monitoring data. The DO and SSC had a good positive correlation in Sumida River but not in Shakujii River. Sumida River follows the usual trend in natural river estuaries where there is high flow velocity, SSC and DO particularly during spring tide. Although adjacent to one another, Shakujii River and Sumida River had different hydrodynamics and water quality trends. Rainfall and DO had a significant positive correlation in Shakujii River but not in Sumida River.
This shows that Shakujii River is highly affected by freshwater flow. The Shakujii River has a smaller catchment area yet, the net SS flux was higher than Sumida River
83
during rainfall. The deposition of the sediments could impose a problem on river water quality deterioration.
The Sumida River and Shakujii River nutrient and POM had different nature.
The POM and nutrient variability in Sumida River is regulated by tidal processes with added effects of tributary rivers. The Shakujii River has a smaller catchment area but the POM levels were similar to Sumida River during neap tide and spring tide and up to 2 times higher than Sumida River POM after rainfall. Shakujii River is highly influenced by freshwater with land derived nutrients and sediments from the combined sewer system outflow. These conditions may be typical in artificial urban estuaries.
The phenomena in the Shakujii River were made clear by studying the adjacent rivers. The main motivation of the study is to address the issue on scum generation and odor in the Shakujii River. From the long term monitoring of water quality and scum coverage, the following critical values were identified for generation of scum: salinity of below 1, SSC of approximately below 15 mg/L, DO less than 6 mg/L and scum appears 2-5 days after rainfall with intensity of less than 11 mm/hr.
5.2 Scum generation and odor analysis in the Shakujii River
The short term and long term data monitoring findings on hydrodynamics and water quality trends were integrated with the odor analysis. Several factors were identified which may cause scum generation and odor. Following the outline of scum generation from previous studies, quantitative values and factors were given correspondingly.
1) Rainfall: The long term monitoring shows that less than 11 mm/hr rainfall causes scum and scum tends to appear 2-5 days after rainfall. The days after rainfall accounts for the time when the scum develop sufficient buoyancy from anaerobic gas that are generated in the bottom sediments. The Shakujii River catchment area is fully served with combined sewer systems with 154 outfall chambers.
Particulates from aquatic plant detritus, sediment and sludge deposited in sewer pipes overflows from combined sewer systems after strong rainfall.
2) Deposition: High SS flux (21.38 kg/m/s) was advected downstream in the Shakujii River after heavy rainfall on November 11, 2016. The bottom depth of Sh-U during spring tide has high SSC levels (> 14 mg/L). The POM in the Shakujii River is as
84
high as Sumida River during neap tide and spring tide and the levels increase five times during the November 11, 2016 rainfall (23.5 mg/L). This may cause flocculation of the SS in the estuary. The flow velocity in Sh-U during neap tide (-0.09 to 0.16 m/s) and spring tide (-0.04 to 0.05 m/s) was very low. This stagnant water condition is favorable for scum generation. It was identified that scum is generated when SSC is approximately below 15 mg/L.
3) Stratification and Anoxic condition: The Shakujii River has hypoxic (< 2 mg/L) bottom depths and low flow velocity during neap and spring tide. There were high levels of POM and nitrogen during the neap and spring tide, high bottom depth SSC during spring tide and high SSC after heavy rainfall. These conditions altogether can lead to anaerobic decomposition of bottom channel deposits due to oxygen depletion.
4) Anaerobic gas generation: The strength of odor representation or component; odor index or the total strength; and odor similarity of scum, air, water and sediment were analyzed using the odor machine. The strongest odor component of scum and sediment was organic acid followed by hydrogen sulfide and sulfur. Hydrogen sulfide was the strongest odor component of water and air. The odor index of sediment was higher than air and water. Sh-U sediment had high odor similarity with scum even at low POM content. Methane and hydrogen sulfide were the two anaerobic gases that were explained in terms of hydrodynamics and water quality.
5) Scum surfacing: Velocity and salinity were identified as additional factors that affect scum surfacing along with other anaerobic gas generation factors. Low velocity is conducive for scum generation. Lower salinity conditions leads to higher methane gas generation and higher scum coverage.
The total nitrogen level in the Shakujii River (8.7 to 12 mg/L) during spring tide is almost close to Sumida River (8.6 to 14.5 mg/L). The elevated nitrogen (NO2+NO3 -N) levels are also high in Shakujii River (7.6 to 8.7 mg/L) and Sumida River (7.2 to11.7 mg/L). From this, it can be stated that denitrification can also cause odor in the Shakujii River. Denitrification is another factor that was identified. When oxygen is depleted, the bacteria turns to other electron acceptor like nitrate. Nitrate is consumed rapidly during denitrification: 𝑁𝑂3− → 𝑁𝑂2− → 𝑁2. The nitrification process: 𝑁2 + 3𝐻2 → 2𝑁𝐻3 and dissimilatory nitrate reduction to ammonia (𝑁𝐻3) or ammonium (𝑁𝐻4+): 𝑁𝑂3− → 𝑁𝑂2− → 𝑁𝐻3 occurs due to organic matter enrichment (Dang and Chen, 2017). After
85
reducing some other minor elements, the bacteria will turn to reducing sulfate: 4𝐻2+ 𝑆𝑂42− → 𝐻2𝑆 + 2𝐻2𝑂 + 2𝑂𝐻−. The by-product of these processes, 𝐻2𝑆 and 𝑁𝐻3 have strong characteristic odor. 𝐻2𝑆 and 𝑁𝐻3 can be detected by the odor machine used for odor analysis. The study however, does not cover the actual chemical components of air, water, sediment and scum. The strength of odor representation gives the probable components of the odor sample based on reference gases that are built in the odor machine.
Extensive studies are continuously done to generate more conclusive findings on scum generation and odor in the Shakujii River. The factors that cause scum generation and odor are shown in Fig. 5.
Fig. 5 The scum generation mechanism leading to odor with causative factors.
5.3 Future work and recommendations
1) Field work: The Shakujii River and Sumida River junction is an important area because during spring tide, the Sh-D and Su-D have a net SS transport that converges in the junction. It is recommended to study the phenomena occurring in the junction of the two rivers which may give good information on the unique trends in the Shakujii River most particularly.
2) Odor and scum analysis: Salinity is now known to either cause higher scum coverage at low level and higher odor index at high level. For the odor analysis of odor, it is recommended to conduct fieldworks or focus on the analysis of data for the variation of
86
odor during neap, intermediate and spring tide. It is also recommended to obtain organic sludge samples from the combined sewer systems. The sludge, scum and sediment samples are recommended to be analyzed chemically to identify the actual components.
This could validate the nutrient, organic matter and other contents of the substances that are discharged in the Shakujii River.
3) Simulation: The study generated immense amount of new and good quality data for the local community. The data could be used for simulation of the river conditions.
Salinity is more prominent in Sumida River two days after neap tide. This could be a good consideration for future simulation and data analysis. A comparison of simulation using Fantom and HEC-RAS may be conducted in the future. The study generated significant findings for the scientific community and may serve as basis to solve current and future river problems in urban estuaries. A mechanism in the future could be developed to extrapolate these site-specific data to urban estuaries with similar channel type and state condition.
87