COMMUNITY ONSITE ANAEROBIC SEW AGE TREATMENT IN A UASB-SEPTIC TANK SYSTEM: SYSTEM BEHAVIOUR DURING WINTER PERIOD IN PALESTINE

The effect of the Mediterranean climate low temperature period on the performance UASB­ septic tank treating concentrated domestic sewage (COD ~ I 00mg/L) in Palestine was 3 studied. Two community onsite UASB-Septic tank reactors (capacity: 0.8 111 , height: 2.5 m) were operated in parallel over a year and monitored over the colder half of the year. The reactors were operated under two different HRT (2 days for RI and 4 days for R2) at average sewage temperature of 17 .3 °C with 12 and 25 °C extreme values. The COD,, in the raw sewage represented a high fraction of the total COD, viz. about 43.7% from the COD,01. RI achieved mean removal efficiencies for COD,m, CODsus, CODcol, CODdis, of 51 %, 83%, 20% and 24% respectively and BOD5 and TSS average removal efficiencies of 45% and 74% respectively. R2 achieved mean removal efficiencies for COD,0., COD,u,, CODcol, CODdis, of 54%, 87 %, I 0% and 28%, respectively with BOD5 and TSS average removal efficiency of 49% and 78% respectively. The sludge production from both reactors was very low with VS/TS ratios of 67.9 and 67.02 of the sludges in RI and R2, respectively. Based on the removal efficiencies and sludge production it can be concluded that the here researched UASB-Septic tank system is a potential compact and effective community onsite pre­ treatment unit for domestic wastewater in Palestine. The reactor can be designed at either 2 or 4 days HRT, but slight preference might be given to 4 days HRT when sludge production is of concern.


Studies with UASB-Septic tanks treating domestic sewage are scarce, and to our knowledge so far only a one research project had been conducted on the use of a UASB-septic tank system for the onsite sewage treatment at Dutch and Indonesian ambient conditions by Lettinga and his co-workers [I,2](Bogte et al., 1993; Lettinga et al., 1993). Nonetheless, the system has not been applied in other countries of different environments and sewage characteristics nor it has been optimised. For instance, in Palestine and Jordan in the Middle
East, sewage is characterised with high COD concentrations exceeding sometimes 1500 mg/L with high fraction of CODss (up to 70-80%) [3,4] .. In addition, in the Middle East region, the ambient temperature fluctuates over the year coupled with sewage temperature fluctuation in the range 15 and 25 °C, during winter and summer times respectively. Previous research has demonstrated that the performance of single stage UASB systems at low temperatures (5-20 0 C) is severely limited by the slow hydrolysis of entrapped solids that accumulate in the sludge bed [5]. This phenomenon is of particular concern when the reactor is fed with highly concentrated raw sewage at low temperatures. The use of the UASB system for the treatment of sewage with relatively high COD concentration is still undergoing trials and argued that such knowledge is important to improve the reliability of anaerobic processes. This is because knowledge of the perfonnance of anaerobic reactors treating municipal wastewater in extreme situation is limited.
The main objectives of this research were to asses the process performance of the community onsite UASB-septic tank for the treatment of concentrated domestic sewage under rather low temperature conditions and also to increase the knowledge on the system design. In view of that, two UASB-septic tank reactors treating domestic sewage in Palestine had been operated under ambient conditions at HR Ts of two and four days. The reactors were operated for a year and monitored for the last six months which are the cold period of the year.

Experimental set-up.
Two UASB-septic tank reactors, namely RI and R2, were installed in parallel at the centralised sewage treatment plant of Al-Bireh City/ Palestine. The reactors were made of galvanized steel with working volumes of 0.8 m 3 (height 2.50 m; diameter 0.638 m). Nine sampling ports were installed along the reactor height at 0.25 m for sludge sampling, with the first port at 0.15 m from the bottom of the reactors. The influent was distributed in the reactor through Polyvinyl Chloride (PVC) tube with 4 outlets located 5 cm from the bottom, Biagas was passed through a 16% NaOH solution for CO 2 scrubbing, and then methane quantity was continuously measured by wet gas meters. Schematic diagram of the experimental set-up is presented in Figure  The UASB-septic tank reactors were started up during spring, ca. April, The reactors were operated in parallel and continuously at ambient environmental conditions for a whole year. The reactors were fed with domestic sewage pre-treated with screens and grit removal chamber. The sewage was pumped every five minutes to a holding tank (200 L plastic container), with a resident time of about 5 minutes, where the reactors were fed and the influent was sampled. The reactors were inoculated with anaerobic sludge obtained from a household cesspit (RI and R2 respectively with IO and 20% of the reactors volume), and operated in parallel at ambient temperature. Grab samples of raw sewage and reactors effluents were collected and analysed two to three times a week after six moths of staring up the reactors, viz, starting in October. Samples were stored at 4 °C till being analysed. Daily monitoring included wastewater and ambient temperature and biogas production measurements, The atmospheric pressure was measured in situ.
Biodegradability of the effluent COD and sludge stability of both reactors were measured twice in duplicate using 500 mL working volume batch reactors incubated at 30 °C for a period of 120 days as described by [3).

Statistical data analysis
Statistical comparisons of means was followed by "Paired samples t-test" for the measured parameters of the two reactors using the SPSS program for windows-Release 11.0.0, SPSS  Table 2 and compared with other studies in Jordan and Palestine.

Figure 3, COD,,,, influent and efjluent concentrations and removal efficiencies for Rf (left) and R2 ( r ight)
During the cold period of the year both reactors RI and R2 were not sufficient for removing CODcol from the influent. The average removal efficiencies were 20% (32) and 10% (37) for RI and R2, respectively with insignificant (p>0.05) difference between both reactors, In addition to the low removal rate of COD co l, even negative removal efficiency had been observed, where this means that the effluent concentration of COD co l some times exceeds the influent concentration. The same results had also been observed by [7], in which is justified as the increase in the COD, 01 was generated from the COD"" that had been digested. The average CODdi s removal rates were 24% (15) and 28% (18) for RI and R2, respectively (see " Table 3

H¾
Kalmar ECO-TECH '07 KALMAR, SWEDEN, November 26-28, 2007 In this research COD dis in the effluent represents about 75.7% and 70.8% from COD , 0 , for RI and R2, respectively. Those results are in agreement with those reported by (4] of 50% CODdis out of the final effluent COD , 01 ofaa UASB reactor. [8] also reported about 50% of the COD,01 effluent from UASB reactor was as COD dis -The average VFA concentrations in the effluent of both reactors were 80 (37) mg/I and 69 (38) mg/1 with average removal efficiency of negative removal and 2 % (82) for RI and R2, respectively. The concentrations of the VFA represent about 30% and 27.8% from the COD dis so most of the COD dis was in a non-acidified form in both reactors. Wang (1994) found about 46% of the effluent COD, 01 after anaerobic sewage treatment could be attributing to non-acidified COD dis as proposed by Van der last and Lettinga ( 1992). This COD part is apparently anaerobically non biodegradable. The VFA concentration in the effluent was affected by temperature and the methanogenic conditions where the production of the VF A decreased during the winter period comparing to the results obtained during the hot period. The negative and low removal efficiency represents an increase in the VFA concentration which is mainly as a result of the predominant acidification process and apparently low methanogenic activities in the two reactors.
[I] reported that declining temperature resulted in reduced production of VF A accumulate at the reactors and a complete conversion of VFA in to CH4 was achieved during 3 to 4 month of second year of the UASB-septic tank operation, when the temperature increased to more than 15 °C.

Biodegradability of effluent
The anaerobic biodegradability of the COD in the reactors effluents were 47.64% and 41.7% of RI and R2 respectively after incubation period of 120 days at 30 °C for the two tests conducted at day 204 and 250 from the start up of the UASB septic tank reactors.

Some researchers have provided information about their systems that could lead to formulation of COD balance [I, 8-9]. The monthly COD mass balances over the two UASB septic tank reactors during the period of the research are presented in
I ; ., .,

Figure 6. Monthly COD mass balance of R 1 (le ft) and R2 (Right) ove r the total test pe riod as a pe rcee ntage of avee rage inf luee nt COD, 0 , and divide d ove r COD accumulate d, COD ejj1uent and CH 4 as COD
Through analyzing Figure 6 for both RI and R2 one can see that in the second month of monitoring and analysis (November), which took place when winter started, the accumulated COD started to decrease and washed out with the effluent COD without any change in the amount of methane production, In December when temperature started to decrease gradually, a reduction in the amount of COD accumulated and methane production was observed and so an increase in the amount of the effluent COD was observed, This observation during that month could be also proved by the lowest monthly COD removal efficiency during the period of the experiment about 41% and 40% for RI and R2 respectively. Later on in January the removal efficiency of COD increased accompanied with an increase in the methane production and the amount of COD accumulated. Moreover, later in February and March (the fifth and the sixth month of the analysis period) the amount of methane production stayed approximately constant while the accumulated COD started to decrease with an increase again in the COD washed out with the effluent. During the six month of the research the COD mass balance could be represented about 25.25% and 30% of the COD accumulated in the reactor R l and R2 respectively, and about 25.6% and 23.32% from the total COD entered the reactors R l and R2 respectively converted to methane. The proportion of COD accumulated and found in R2 which is relatively higher than R l justified the slightly better removal efficiency achieved in R2.

TSS and VSS removal efficiency
Both reactors performed equally good as the achieved TSS removal efficiencies were 74 % (10) and 78% (11) by RI and R2 respectively but with no statistical significant differences (p>0.05) between the two reactors. The results reported for TSS in this research are better than the results reported in literature review for conventional UASB reactors that have treated domestic wastewater. The removal efficiencies averages for VSS for this research were 74% (I 0) and 78% (12) for RI and R2, respectively. However, R2 is significantly better than RI with respect to YSS removal efficiency ( p<0.05). The YSS and TSS concentrations and removal efficiencies for RI and R2 were stable regarding to the TSS concentrations measured at the effluent throughout the period of the research. The achieved TSS and YSS removal efficiencies are rather good because the reactors were operated during the cold period of the year, [9] pointed out that the decrease in temperature would be accompanied with direct increase of the wastewater viscosity and so increase the hydraulic shearing force on solid particles, Accordingly the solids particles will move out the reactor which directly reduces the removal efficiency of both reactors, The average VSS/TSS ratio for both of the reactor R I and R2 were 0,84 (0,06) and 0,81 (0.04) respectively, which is closed to the results reported by (AI-Shayah, 2005 6). However, the difference in removal efficiency of (NH/-N) were not statistically significant (p>0,05), That also holds true for the Nkj-N as Nkj-N was partially removed in the USAB reactors due to particulate N removal ( Table 3). The average removal efficiencies of Nkj-N were 17 % (7) and 15 % (8,3) for R I and R2, respectively, Moreover, the difference in removal efficiency of (Nkj-N) were not statistically significant (p>0.05).

Tota\ -P and Ortho phosphorous removal
The results presented in Table 3

1 pH in the UASB-septic tank reactors
In this research the pH mean value for the raw sewage Influent was 7,6 (0.28) and 7.44(0, 13) and 7 .4 7 (0, 16) for the effluent of R I and R2, respectively. The slightly lower pH values which was observed in the UASB effluent is expected in the anaerobic treatment where the buffering capacity in the raw domestic wastewater is enough to neutralize the production of volatile acids and carbon dioxide, which dissolved at the operating pressure [ 11], During the whole of the experiment was no observation for pH value out of the normal and optimum range where for R I the pH ranged from pH (7.14-7,68) and for R2 pH ranged from (7.1-7. 79).

2 General discussion
Fonn the average influent concentration of COD 101 (I 045 mg/I) during one year which was obtained by this research and the research don by Al-Shayah (2005) one can calculate the number capita equivalent to the COD 101 entered each reactor which is 3 capita and 2 capita for RI and R2, respectively see Table 4,7, About 1892,5 1/c,year and 1321.85 1/c,year of CH 4 in gaseous form were produced from RI and R2, respectively as measured from the gas meter see Table 4.7. Moreover, the annual specific sludge production for each capita was 2, I kg TSS/c,year and 2, 7 kg TSS/c,year for RI and R2, respectively see Table 5, The accumulated COD in the sludge after I year per person in the reactors equal 2,29 kg COD/c.year and l .86 kg COD/c.year in RI and R2, respectively see Table 4.7. These values were reasonable related to (Jewell, 1994) who reported that for each I 00 kg COD soluble treated there will be 5 kg COD converted to sludge as mentioned at Figure 2.1. Theoretically and regarding to [ 12], the amount of COD that present at the sludge equals 5% of the soluble COD. Regarding to this research the percentage was taken related to the total COD, relatively it is less than the results reached by [ 12] it was 4.5% and 4.8% from the total COD for RI and R2, respectively, In general the design criteria of community on site UASB-septic in Palestine are presented in Table 5. At Table 6 the OLR for RI with HRT of 2 days was 0,45(0.12) and for R2 with HRT of 4 days was 0.23(0,06),

CONCLUSIONS
• The here presented UASB-septic tank reactor is a very efficient technology for onsite sewage treatment in Palestine, It achieved the following main removal efficiencies: • The reactor can be adequately designed at either 2 or 4 days HRT, but preference might be given to 4 days HRT when sludge production is of concern 2 5 1 (9)  3 7) 28 ( 1 8)