Temporal Changes in Environmental Health Risks and Socio-Psychological Status in Areas Affected by the 2011 Tsunami in Ishinomaki , Japan

On March 11 2011, a tsunami caused by a magnitude 9.0 earthquake devastated the northeastern coast of Honshu, Japan.The present study was conducted to assess environmental health risks of the areas affected and socio-psychological status of the dislocated people in Ishinomaki. Samples of sludge, water, flies and rodents were collected in 20 urban neighborhood associations affected by the tsunami in July and August 2011, and in August 2012. A socio-psychological survey was conducted in two urban and one rural temporary housing complexes in 2012. Animal feed concentrates and fish from damaged factories were scattered along the coast which caused a strong odor and great number of flies. Removal of fish and feed along with spraying insecticides reduced the odor and the number of flies by August 2011. The sludge and water samples contained potentially hazardous bacteria, but none were highly pathogenic. Heavy metals in sludge were not in alarming quantities. A rodent was captured in one unit in August 2011, and monitoring in two units found that the log number of rodents captured increased significantly over time (slope=0.08, p = 0.005). In temporary housing complexes, those who originally lived in rural fishing villages wished to return to their homes more (64.2%, 9/14) than in urban areas (30.6%, 11/36, p = 0.06). Risk factors for depression included absence of friends (p = 0.011) or trusted person to counsel in the housing complexes (p = 0.003) and illness of the respondent or a family member (p = 0.003). In conclusion, overall environmental health risk was acceptable for living, and monitoring of rodents population was recommended. In addition, psychological and economical support was needed for evacuees in temporary housing complexes.


Introduction
A magnitude 9.0 earthquake occurred offshore of the northeast coast of Honshu, Japan at 14: 46 pm on March 11, 2011, generating a devastating tsunami that destroyed many towns and villages near the seashore in Iwate, Miyagi and Fukushima prefectures (Shibahara, 2011).The earthquake was named the 2011 Off the Pacific Coast of Tohoku Earthquake by the Japan Meteorological Agency (JMA, 2011) and the Japanese cabinet named the collective disasters caused by this earthquake the Great East Japan Earthquake (PMJHC, 2011).The total number of deaths caused by the disaster totaled 15883, and 2681 people were still missing by April 1, 2013(NPA, 2013).The coastal city of Ishinomaki in Miyagi Prefecture (Figure 1) was severely affected by the tsunami, which flooded about 13% of the city where 70% of the population lived (Ueda, Hanzawa, Shibata, & Suzuki, 2012).Table 1 summarizes the damage and health problems experienced by inhabitants of Ishinomaki.The large scale tsunami warning was announced at 14: 49 pm and at 15: 26 pm, more than 8.6 m of tsunami reached to Ishinomaki (JMA, 2011).In the affected areas by the tsunami overall in Japan, 62.6% of people started evacuation before the large scale tsunami arrived, 10.6% after arrived, and 26.8% did not evacuate, based on a study with 4,421 respondents (MLITT, 2011).Ishinomaki lost the lives of 3506 people, which accounted for 22% of total deaths in Japan (Table 1).The tsunami necessitated the evacuation of 31.6% of the city population (peak on March 17, 2011) to disaster shelters.Immediately after the tsunami, the Self Defense Force of Japan, together with foreign rescue teams including US military forces, began rescuing people trapped in debris, searched for those who were missing or presumed dead and restored life lines to the affected areas.There were domestic and international financial and material donations through United Nations, Red Cross and other NGOs.Volunteers and NPOs also played critical roles in restoring the city, providing meals and caring for evacuees.Evacuees were encouraged to move to rental flats and rapidly constructed temporary housing complexes, and all disaster shelters were closed by October 11, 2011, 7 months after the earthquake.The supplying period of the temporary houses had been for two years but was extended another year until 2014 (Miyagi Prefecture, 2012).The government is reclaiming uplands for relocation from coastal residential areas and creating employment opportunities (Reconstruction Agency, 2013).The other political supports include partial financial supports for lost and damaged houses, exemption of tax and interests for a double loan (a loan for new house construction in addition to the remaining loan), and supports for families with children and those who lost employment, and mental and physical health (Miyagi Prefecture, 2012).
The clinical phases of natural disasters can be classified as follows (Aghababian & Teuscher, 1992;Kouadio et al., 2012): phase 1 (impact phase, 0-4 days), initial treatment of disaster-related injuries is provided; phase 2 (post-impact phase, 4 days to 4 weeks), the first waves of air-, food-, and/or water-borne infectious diseases emerge as a substantial portion of the population is displaced into unplanned and overcrowded shelters; and phase 3 (recovery phase, after 4 weeks), infections with long incubation periods or latent types can become clinically apparent.
In Ishinomaki, in the phase 1, dehydration, vomiting and diarrhea were frequently observed due to lack of clean water and hygiene.In the phase 2, seven cases of tetanus, two cases of legionellosis, small outbreaks of diarrhea, respiratory diseases and deep vein thrombosis resulted from dehydration and restricted movement in crowded disaster shelters were reported (Table 1).There were 259 disaster shelters in Ishinomaki.In the tsunami-flooded areas, shelters were very crowded with approximately 1000 evacuees per shelter.Most of the shelters provided 10 square feet per person for many weeks and at most 20 square feet per person for several months (Picture 1) (Ueda et al., 2012).
While public health authorities concentrated on the health status and treatment of evacuees in the disaster shelters, environmental hygiene received little focus.Poor environmental hygiene can be a source of infectious and non-infectious diseases not only for evacuees in shelters, but also for those who return to damaged homes or move into rented rooms or temporary housing complexes (Picture 2).The most frequently reported infectious diseases associated with a tsunami and floods in the world are diarrhea, hepatitis A and E, acute respiratory infections (ARIs), measles, meningitis, tetanus, cholera, and leptospirosis (Kouadio et al., 2012).Although international data are scarce, levels of heavy metals were increased by the tsunamis in 2011 in Japan (Baba & Sera, 2012) and in 2004 in India (Ranjan, Ramanathan, Singh, & Chidambaram, 2008) and such contaminants in soil might pose public health risks unless long-term exposure and internal accumulation are prevented.The present study assesses environmental health risks due to microbiological agents, wildlife and chemical hazards in the areas of Ishinomaki affected by the tsunami to provide information that can be used to plan reconstruction policies.As a risk is assessed by the combination of probability of occurrence of a scenario that will affect humans and the size of the impact (Vose, 2008), we also conducted a sociological survey to assess the desires of evacuees in temporary housing complexes to return the areas affected by the tsunami.The mental health status of the residents was also surveyed to provide adequate and timely policy support.
The locations and needs of evacuees, the progress of restoration, and environmental hygiene obviously changed over the course of the study period.The present paper describes the temporal dynamics of general observations, entomology, wildlife, microbiology, chemistry and socio-psychology in the areas affected by the tsunami in Ishinomaki.
www.ccsen d is erring O1 V. 9 and 3, but ugust ft for to be great in July 2011.Flies were also captured using water traps with a few drops of surfactant in a yellow plastic container for 1 day and 1 night in one unit each strata where the number of flies was great.All the samples were chilled and packaged in Ishinomaki and sent to Rakuno Gakuen University (RGU) for morphological, microbiological, and chemical analyses.
Rodents were sampled in the 20 units in August 2011, using Sherman Traps (H.B. Sherman Traps Inc., FL, USA).Follow-up sampling was conducted at the unit 17 (Figure 1) in the western stratum, where a rodent was trapped in August 2011 (see Results), in November 2011 and in March and May 2012.In November 2012, sampling was conducted at the unit 17 again, and a grassland with a pine woods near the unit 9 (Figure 1), both of which were ecologically suitable habitats for rodents -proximity to water source such as canal and sea, secondary woods, and compiled debris.Collected rodents were euthanized in the field, chilled, and sent to RGU for microbiological tests.

Microbiological Tests
In 2011, bacteriological tests were performed on the sludge and water samples collected for total bacteria, enterobacteriaceae, Salmonella, spore-forming bacteria, Vibrio, and Aeromonas.Bacteriological follow-up in August 2012 comprised assessments of spore-forming bacteria.
Total bacteria, enterobacteriaceae, Salmonella, and spore-forming bacteria were counted in serially diluted samples of sludge and water.Total bacteria was counted after incubation on brain heart infusion agar in 5% CO 2 at 37 °C for 24 h.Enterobacteriaceae, which can indicate contamination with sewage or feces, were cultured on DHL agar at 37 °C for 24 h.Salmonella, which is a common cause of severe food poisoning, was enriched using Hajna tetrathionate broth at 37 °C for 18 h, left at ambient temperature at around 25 °C for 7 days, and inoculated onto Mannitol Lysine Cristal Violet Brilliant Green (MLCB) agar (Nissui, Tokyo, Japan) at 37 °C for 24 h.Aerobic spore-forming bacteria were cultured on Trypticase soy agar (Becton, Dickinson and Company, MD, USA) supplemented with 5% sheep blood (5% sheep blood agar) at 37 °C for 24 h.Anaerobic spore-forming bacteria were cultured on 5% sheep blood agar anaerobically for 72 h.Aerobic spore-forming bacteria were identified as Bacillus spp.and the anaerobic spore-forming bacteria were Clostridium spp.Bacillus spp.includes B. anthrax, which causes cutaneous and intestinal diseases, and B. cereus, which can cause food poisoning due to enterotoxin.Clostridium spp.includes C. tetanus, which causes tetanus after invading skin wounds, and C. perfringens, which causes food poisoning.C. perfringens isolated by anaerobic incubation on 10% egg yolk CW agar (Nissui) at 37 °C for 72 h was identified based on Nagler's reaction and inhibition by antiserum.Vibrio and Aeromonas can cause food poisoning and V. cholerae serogroups O1 biotype El Tor and O139 cause cholera, and they are frequently associated with sea foods in developed countries and contaminated water in developing countries (Sack, Sack, Naire & Siddique, 2004).Sludge samples suspended in phosphate-buffered 2% NaCl and water samples enriched with alkaline peptone broth at 37 °C overnight were inoculated onto ES Vibrio agar plate (Eiken Chemical Co., Ltd., Tokyo, Japan) and incubated at 37 °C overnight to determine the presence of Vibrio and Aeromonas.Isolated colonies were pure cultured on ES Vibrio agar plate once again and passaged onto 2% NaCl Trypticase soy agar.The biochemical characteristics of pure colonies were determined using TSI, LIM, and oxidase tests, and bacteria were identified using API20E (bioMérieux Inc., Durham, NC, USA).Bacteria that were identified as V. cholerae were tested for sero-groups O1 and O139, as well as the ctxA virulence gene using PCR.
A pool of 10 flies sampled by sweeping in July 2011 was prepared for each of the five sites.These five pools were served for bacteria counting of total bacteria, enterobacteriaceae, and Salmonella.
Sludge samples were virologically tested for hepatitis E virus (HEV) and Norwalk-like virus (NLV), which are hazardous to human health and can be found in the environment.The HEVs associated with human hepatitis are classified into four genotypes (Lu, Li & Hagedorn, 2006).Genotypes 1 and 2 cause waterborne outbreaks in developing countries and genotypes 3 and 4 are considered to be zoonotic and transmitted through the consumption of uncooked or undercooked contaminated meat (Mitsui et al., 2004;Wong, Purcell, Sreenivasan, Prasad, & Pavri, 1980).Total RNA was extracted from sludge suspensions using QIAamp Viral RNA Kits (QIAGEN, Hilden, Germany).The HEV RNA of the 5′terminal region of ORF1 was detected by semi-nested RT-PCR (Kanai et al., 2009) using the One Step RT-PCR Kit (Qiagen).NLVs which belong to the Caliciviridae family are major causes of acute nonbacterial gastroenteritis and a major public health concern.The NLV GI ORF1-ORF2 junction region was amplified by PCR using three forward primers for G1FF corresponding to nucleotides (nt) 5075 to 5097 in Norwalk/68, and the reverse primer, G1SKR (Kojima et al., 2002).The NLV GII ORF1-ORF2 junction region was also amplified by PCR using three forward primers for G2FB, corresponding to nt 4922 to 4941 in the Camberwell virus, and the reverse primer, G2SKR (Kojima et al., 2002).The RNA samples were reverse-transcribed using Transcriptor reverse transcriptase (Roche, Basel, Switzerland) and a random primer (2.5 μM).All representative cDNA samples were then amplified by PCR.

Chemical Analysis
Water samples collected in June and August 2011 were passed through a syringe filter (0.45 μm) and analyzed for anions and cations using DionexIC-20 (Thermo-Fisher Scientific, Waltham MA, USA) and PIA-1000 (Shimadzu, Kyoto, Japan) ion chromatographs.
Sludge samples were dried, homogenized, and analyzed for cadmium (Cd), mercury (Hg), lead (Pb), and arsenic (As) using particle-induced X-ray emission (PIXE) at the Nishina Memorial Cyclotron Center (Iwate, Japan).The estimated detection limits for the toxic elements Cd, Hg, Pb, and As were 20, 3, 3, and 1.4 ppm, respectively.A detailed description of the measurement setup is provided elsewhere (Sera & Yanagisawa, 1992).
A control unit was selected from an unaffected residential area called Hebita (shown in Figure 1) as a reference for chemical analysis.The levels of all radioactive materials measured by the government surveillance remained below the standard levels established by the government and this topic was not further studied.

Socio-Psychological Survey
Findings of environmental damage and mental health that had been recorded in Excel data spread sheets by students, staff, and faculty members of RGU who had volunteered in Ishinomaki since May 2011 were qualitatively reviewed by the authors, and a summary was prepared in a participatory manner (Mariner & Paskin, 2000).This database was created in Ishinomaki in order to accumulate information so that a new volunteer group arrived in the base camp can efficiently follow up the activities done by the previous group.Participatory appraisals (Mariner & Paskin, 2000) were conducted in one rural and two urban temporary housing complexes and plans for future accommodations, current health status, stress, and the living environment of the residents were assessed in July 2012.
A questionnaire was designed based on the results of the participatory appraisals to understand the factors associated with reduced mental health in detail.K6 (Kessler et al., 2003) values were collected to screen for a serious mental illness (SMI) among residents in the temporary housing complexes.Respondents with K6 values > 13 were considered to have SMI, referring a cut-off point suggested by Kessler et al. (2003).The Japanese translation of questions for K6, which was validated through the backtranslation procedure and with high value of the areas under receiver operating characteristic curves (AUCs): 0.94 (95% confidence interval 0.88-0.99)(Furukawa, 2007), was used in the present study.The questionnaires were distributed to 50 households each at the two urban complexes and to all 44 households in the rural complex.To ensure anonymity, the respondents returned sealed questionnaires to RGU by mail.

Statistical Analysis
The number of rodents captured in a unit was log e -transformed, as a count follows Poisson distribution whose link for regression is logarithm, and a regression was performed with the explanatory valuable the month elapsed from occurrence of the tsunami when captured.The units where rodents were not captured were excluded from the analysis.
Log 10 -transformed concentrations of bacteria were compared among samples of water and sludge that had been collected at different times using paired student t tests.Means and confidence intervals were calculated in log 10 scale.The prevalence of C. perfringens, Vibrio spp., and Aeromonas was compared using Chi-square test and Fisher's exact test was applied when at least one cell in 2  2 tables included an expected frequency below 5.
Ion concentrations were log-normally distributed and log e -transformed data collected at different times were compared using paired Student's t tests.The ratio of Na + and Cl -are useful for determining proximity to seawater (Yoshii et al., 2012) and the value of 0.56 for seawater was calculated based on data provided by Sverdrup, Johnson and Fleming (1961).The degree of proximity to seawater according to the ion composition of the samples is therefore expressed as a ratio of sample and seawater Na + /Cl -(a ratio of 1 suggests that the Na + /Cl - values of the seawater and the sample are identical).Ratios were also log-normally distributed and compared using paired student t tests of log e -transformed data.A value of 0.1 mg/kg was assigned to samples with values below detection limits to log e -transform log-normally distributed heavy metal concentrations for statistical analysis.These log e means and confidence intervals were transformed back to the original scale for presentation.

Transe
The density of enterobacteriaceae and Clostridium in surface water samples significantly decreased (3.4 to 2.7, p = 0.006 and 3.3 to 2.7 mean Log 10 CFU/mL, p = 0.023, respectively) between July and August 2011.In contrast, the total bacterial count (5.4 to 6.3 mean Log 10 CFU/mL, p < 0.001) and the density of enterobacteriaceae (4.4 to 4.9, p = 0.046) in sludge significantly increased between July and August 2011.The concentrations of Bacillus (mean Log 10 CFU/ml: 4.9 to 3.8, p < 0.001) and Clostridium (5.0 to 4.5, p = 0.004) significantly decreased between August 2011 and August 2012.Salmonella was not identified in any of the water or sludge samples.Flies carried high numbers of bacteria; the mean Log 10 CFU/mL of total bacteria was 8.7 (95%CI: 7.3-10.1)and that of enterobacteriaceae was 8.0 (95%CI: 6.6-9.5, data not shown).

Chemical Analysis
Table 4 shows changes over time in mean ion concentrations between July and August 2011.In July, the mean ratio of sample Na + /Cl -to seawater Na + /Cl -was 1.01 (95%CI: 0.78-1.30)and the composition of the sampled surface water was similar to that of seawater.The surface water with similar composition to seawater was geographically evenly distributed in the studied areas (black dots in Figure 2A).In August 2011, the concentration of Na + remained (p = 0.65) but that of Cl -significantly decreased (p = 0.007), which resulted in a significant increase in the ratio of sample Na + /Cl -to seawater Na + /Cl -(2.61, p = 0.007).The composition of surface water had considerably changed in all tested areas by August 2011 (Figure 2B).  ) in July and A detection limit Cu, Hg, Pb, N fined in the S 0 mg/kg, resp nd confidence i in unit 1 of t vels of Pb and tectable and th According to the qualitative records collected from the RGU students during volunteer activities, the contents of stress dynamically changed over time.Shocks due to having lost homes, jobs, family, and friends were prevalent among those living in disaster shelters and damaged houses in May 2011.Many people felt guilty for not having died themselves, considering those who had died.The amounts of debris and sludge were overwhelming, but joint cleaning activities with volunteers offered a distraction.Long waiting list to enter in a room in the temporary housing complexes and inconvenience of many of the complexes due to the remote locations induced stress in May and June.Lost community function due to the absence or relocation of community leaders and board members was a huge obstacle for those who returned to damaged homes, where they lived without basic supplies including food, clothing, and hygiene products.
According to the participatory appraisals, in September 2012 the focus of mental stress shifted to the imminent future.Several political leaders announced closing dates of the temporary housing complex, but in reality, planning, upland reclamation, housing construction, and policy support for the relocation of evacuees were delayed.Obstacles for elderly respondents included ineligibility for loans to construct houses.Even middle-aged individuals worried about how to pay for a second loan while still paying original loans for homes that were destroyed by the tsunami.

Questionnaire Results
In September 2012, 44 evacuees in the two temporary urban housing complexes and 16 in the rural complex responded to the questionnaire.The urban and rural response rates were 44% (44/100) and 36.4% (16/44), respectively, and the mean ages of respondents were 63.6 and 58.7 years, respectively.The urban relocation area planned by the authorities was inland Ishinomaki and only 30.6% (11/36) of the respondents wished to return to their own homes in urban areas, whereas 64.2% (9/14) of those in rural areas wished to return (x 2 = 3.48, df = 1; p = 0.06).Regardless the planned destination either original home or inland, 46.2% (12/26) of urban respondents felt that their wish would be unable to be realized and 58.3% (7/12) of these stated that financial problems were the main obstacle.Those in rural areas who chose relocation (35.8%, 5/14) wished to move to nearby residential areas with higher elevation after development of these areas by the authorities, and wished to continue working in the fishing industry.All of the respondents in the rural housing complex felt that this would be realized.
High K6 value which suggests SMI (K6 value > 13) was found in 15.9% (7/44) and 6.3% (1/16) of the urban and rural respondents, respectively.A total of 13.3% (8/60) had severe mental health conditions.This was higher than the prevalence of K6 above 13 at pre-tsunami status, 5% in 2010 (NCNP, 2013).The proportions were not significantly different between urban and rural respondents (p = 0.7).Table 6 shows univariate analysis of K6 values which indicate mental stress.Statistical significance was found in health problems (p<0.001),lack of friends (p = 0.003) or trusted person to counsel (p = 0.005), not participating in events (p = 0.02), and pest infestation (p = 0.03).The overall mean K6 value was 6.5 (data not shown).Multivariable analysis revealed that risk factors for a high K6 value were a lack of friends (p = 0.011) or trusted person to counsel (p = 0.003) and the illness of the respondent or a family member (p = 0.003, Table 7).Respondents with friends in the temporary urban housing complexes had significantly lower K6 values, indicating less mental stress (4.9 vs. 11.9;p< 0.001) than those without friends, whereas these values did not significantly differ in the rural complex (5.4 vs. 1.0, p = 0.4).Smilarly, K6 values were significantly lower among respondents in urban areas with trusted person to counsel than without (4.1 vs. 12.7, p = 0.002), but did not significantly differ in the rural respondents (3.3 vs. 6.2, p = 0.8).The proportions of respondents with friends (p = 0.2) and trusted counselors (p = 1) did not significantly differ between the urban and rural complexes.

Discussion
The present study revealed the temporal dynamics of environmental health risks and socio-psychological status in areas of Ishinomaki affected by the tsunami between 2011 and 2012.While public health interests might focus on monitoring the incidence of infectious diseases, public health risks must be understood from the environmental viewpoint of tsunami-affected areas for mid-and long-term reconstruction planning.Japan has been promoting the relocation of residential areas to uplands to avoid future tsunami-related disasters (Reconstruction Agency, 2013).On the other hand, the present study showed that some populations would prefer to live in their original locations.The present environmental and socio-psychological risk assessment is important in providing real-time information to policy makers in Ishinomaki so that they can plan adequate policy support for those who lived in the tsunami-affected areas, as well as those in temporary housing complexes in Ishinomaki.
Common health problems that arose in the disaster shelters comprised respiratory and gastro-intestinal (GI) syndromes (IASR, 2011b).Norovirus infections were reported in Fukushima (IASR, 2011c) and Iwate Prefectures (IASR, 2011b).However, none of these outbreaks were large-scale considering the reported epidemic patterns (IASR, 2011b; ISAR, 2011c) and there might not be many spill-over infections from the environment, except for infections that developed during phases 1 and 2, when clean water was scarce under crowded conditions (Ueda et al., 2012).The risk of food poisoning was high in the affected areas due to the contamination of water, sludge, and flies with potentially hazardous bacteria.The disposal of fish carcasses, cleaning, and insecticide spraying dramatically decreased the number of flies, which adequately decreased health risks by August 2011.
Cholera frequently occurs during natural disasters.The global attack rate of V. cholerae O1 Cholera epidemic in Haiti after the 2010 earthquake and hurricanes was 488.9/10,000 inhabitants and the mortality rate was 6.2/10,000 inhabitants (Gaudart et al., 2013).Although V. cholerae and V. vulnificus was isolated in Ishinomaki, toxigenic V. cholerae was not.However in the USA, environmentally acquired V. vulnificus, V. parahaemolyticus, and non-toxigenic V. cholerae caused infections and deaths soon after Hurricane Katrina (CDC, 2005).This suggests that the bacteria isolated in Ishinomaki could have caused infections and deaths.
Early establishment of hygiene in the disaster shelters and relocation to temporary housing complexes might have prevented such outbreaks.Vibrio outbreak associated with the tsunami has not been reported in any of the affected areas of Japan.After the 2004 tsunami in Banda Aceh, Indonesia, acute jaundice potentially due to infection with water-borne hepatitis A (HA) and E was identified among displaced population (WHO, 2005).The present study detected HEV in only one sample and no outbreak occurred, probably because HA and HE are not endemic in Japan.Although the land in areas affected by the tsunami remained covered with seawater until July 2011, and the NH 4 + concentration suggested high urine contamination, hygiene at the disaster shelters in the flooded areas might have been maintained at a high enough level to avoid such waterborne outbreaks.On the other hand, the high NH 4 + concentration suggested a harsh living condition of the people remained in the damaged houses, who were later relocated to temporary housing complexes or restored toilet facilities.The increase of the Na + /Cl -ratio was probably due to Na + sorption onto soil's clay particles, while Cl -was leached by rainfall.Evacuees in the disaster shelters and damaged houses were relocated to temporary housing complexes by October 11, 2011, when they were finally released from an overcrowded, uncomfortable environment.
The most dynamic changes observed in the present study were in rodent populations.After the affected areas were cleaned and debris was removed, the original residential areas became covered with grass and the lands became favorable to rodents.These animals are recognized as important mammalian reservoirs of Leptospira spp.(Meerburg, Singleton, & Kijlstra, 2009) and shed infectious organisms in urine throughout their lifespan (Li et al., 2013;Vinetz, 2001).Leptospirosis can be fatal and an outbreak in the Philippines after the flood in 2009 was characterized by jaundice, anuria, and hemoptysis (Amilasan et al., 2012).Leptospira was not investigated in the present study, but future environmental surveys should include diagnosing leptospirosis in rodents to understand the health risks.
Notably high level of Hg (24.9 mg/kg) was observed in only one sludge sample.This might have been carried from soils beneath the sea.A study of the coast in Aomori, Iwate and Miyagi prefectures found a greater variety of heavy elements in sludge than in inland samples (Baba & Sera, 2012).Sediments in Japanese bays contain considerable amounts of toxic heavy metals (Kabir et al., 2006).A study of hair from individuals in areas affected by the tsunami in Iwate Prefecture did not find clear differences in heavy metal concentrations between before and after the tsunami (Sera et al., 2012).However, no data are available for individuals in Ishinomaki who lived in damaged houses contaminated with sludge, and the effect of exposure to heavy metal particles in dried sludge is unknown.Concentrations of Cd, Cr, Cu, Ni, and Pb increased from 7 to 35, 141 to 617, 32 to 132, 62 to 252, and 11 to 144 mg/kg, respectively, after the 2004 tsunami in India (Ranjan et al., 2008).Although we did not find such alarming concentrations, metal elements might be released into the ecosystem and create a threat in the event of changed geochemical status (Ranjan et al., 2008).Therefore, future monitoring, especially of agricultural lands, is recommended, as elements absorbed from agricultural products can enter the food chain.
The contents of mental stress considerably changed over time, but residents in temporary housing complexes remained highly stressed even at one and a half years after the tsunami.A published review has argued that 30-50% of individuals after a tsunami would experience moderate to severe psychological distress that might resolve with time, or mild distress that could become chronic (Carballo et al., 2005), which supports the findings in the present study.The risk factors for serious mental illness suggested that individuals without close human relationships might be vulnerable.Our comparison of associations between human relationships and mental stress in urban and rural areas also supported this assertion.Human relationships remained close in rural areas, where even those who felt isolated did not have increased mental stress.In contrast, evacuees were from many different places in urban complexes, and weaker human relationships affected mental health.A review paper which summarizes the impacts from tsunamis in the world has indicated that women, children, and elderly individuals are the most vulnerable to mental stress during a tsunami (Carballo, Heal, & Hernandez, 2005).The present study, however, did not find such a tendency.The prevalence of mental illness among such vulnerable populations might have been higher soon after the tsunami.

Conclusion
Considering the effect of removing sludge from the soil of affected areas and the limited number of people returning to their original residences, the assessed environmental health risk has been reduced to a level that is acceptable for living.However, the population of rodents is increasing, and these pests may harbor hazards for humans.Monitoring rodent population dynamics and the prevalence of zoonotic agents such as Leptospira spp.might be needed.Financial support for evacuees to construct homes is also recommended.Furthermore, mental support is needed for evacuees in temporary housing complexes and efforts should be directed towards enhancing closer relationships in such complexes.comparative data on heavy metal concentration in unaffected soil by tsunami.The inundation limit spatial data was produced by the Earth Environmental Engineering Group, Institute of Industrial Science, the University of Tokyo, under permission by the Geospatial Information Authority of Japan.The biggest thanks and our love go to the people in Ishinomaki participated in this study as well as all those who were affected by the Great East Japan Earthquake.We shall never forget about the victims of the disaster.

Figure
Figure Note: 2A s

Table 1
www.ccsen Note: Each 2. Study S ion sing complex, I family (taken i

Table 2 .
Log 10 CFU/ml and 95% confidence interval of total bacteria, enterobacteriaceae, Bacillus spp.and Clostridium spp. in surface water and sludge serotypes O1 or O139 and none had the ctxA gene which is associated with virulence.The prevalences of V. cholerae, V. fluvialis, and A. hydrophila were not significantly different.
b: significantly different between August 2011 and August 2012Table 3 shows the prevalence of Vibrio spp.and Aeromonas hydrophila in surface water and sludge samples.The numbers of samples are different from those in Table 2 because a different team collected the samples.One extra surface water sample was collected in August 2011.Although V. cholerae were isolated, they were not highly pathogenic

Table 3 .
Prevalence of Vibrio spp.and Aeromonas hydrophila Virological tests in July 2011 detected an HEV monoclonal band in one of 12 samples of sludge, but in none of 13 samples of surface water.Norovirus was not detected in any water or sludge samples in July 2011.Virological tests for water and sludge were not conducted in August 2011.Bornaviruses were undetectable in rodents sampled in July 2011.
*Highly pathogenic V. cholerae serotype O1 or O139 was not detected.None of V. cholerae had ctxA gene which is associated with the virulence.

Table 4 .
Changes of the mean ion concentrations (mg/L) and 95% confidence intervals in surface water between July and August 2011

Table 5 .
Changes of the mean heavy metal concentrations (mg/kg) and 95% confidence intervals in sludge between July and August 2011 Detection limit of Cd is 20 mg/kg and Cd was not detected from all the samples. *

Table 6 .
Univariate analysis for mental health

Table 8 .
Comparison of the associations between human relationships and mental stress (K6) in urban and rural areas *Note: total numbers of answer are different between the questions because of the answers not provided.