Introduction
Floods are among the natural disasters that cause death, disease, and environmental damage [1, 2, 3, 4]. According to the Emergency Events Databases report in 2022, floods are one of the ten most destructive disasters worldwide, with the highest number of deaths, total affected, and economic losses [5]. The most meteorological disasters in Asia are floods [3]. As one of the countries in the Middle East and South Asia, Iran is affected by floods; in 2019, almost 213 cities were involved, and 77 people died [6]. Floods cause people’s migration and create economic and cultural challenges as they destroy farmland and residential homes [7]. Both structural and non-structural interventions, such as forecasting, warning, and emergency evacuation, help mitigate flood consequences; however, non-structural techniques are more efficient and cost-effective in decreasing flood consequences [7, 8]. Evacuation is especially important during the preparation phase. It is an essential and effective measure to move residents from high-risk locations to safer areas [6]. During flooding, early, timely, and safe evacuation is necessary to reduce the effects of the flood [9]. Although evacuation helps protect people, animals, and documents, individuals are unwilling to evacuate [10] voluntarily. Various factors, such as past experiences with disasters, risk assessment, and evacuation plans, can predict the behavior of communities toward evacuation [11]. Meanwhile, Factors such as a lack of an early warning system, resources, planning, risk perception, and cooperation are known reasons for delaying evacuation [12]. Successful evacuation is critical for saving lives from future floods, which can be accomplished by analyzing evacuation behaviors [13].
Although various factors are helpful in evacuation, limited qualitative studies have been conducted in this field [14, 15]. Knowledge of public risk perception helps to reduce flood risks, and qualitative research helps to understand these ideas [16]. This is the first study in Iran to discover the factors affecting evacuation in floods. Since few studies have been done on flood evacuation in Iran and the authors’ experience with floods showed that people did not leave their homes despite the evacuation order, this research was conducted to explain the factors affecting flood evacuation.

Materials and Methods
Study design
Content analysis with a conventional approach was applied in the study. The steps used in this study include record, transcript, and coding. Then, information analysis was done to extract and categorize flood evacuation factors.

Study setting
The present study was conducted among stakeholders, including managers and citizens of Poldokhtar City, Iran. In 2021, Poldokhtar, one of the cities of Lorestan Province in Iran, was affected by floods (Figure 1). 
Study participants
A total of 27 stakeholders, including 10 citizens and 17 executive managers, were selected by the purposive sampling method. Interviews were carried out in two categories, namely, people and managers. In selecting the subjects, the participants were chosen from flood-affected areas among people. In choosing the managers, people involved in flood management were chosen. These persons included individuals working in cooperative and support centers in disaster management, such as the Red Crescent, the Governorate, the municipality, and the Water and Sewerage Organization. In choosing the participants, the most diversity was considered. The criteria for entering the study included the flood experience and the ability to communicate.

Data collection
This study used semi-structured interviews [17]. Interviews continued until information saturation was reached [18]. The absence of data addition for the study’s development was the saturation criteria. The duration of each interview was between 40 to 70 min based on the participants’ agreement. Individual interviews were conducted at the time and place accepted by the participants in places that they suggested. Written consent was obtained from the interviewees for audio recording. Interviews with citizens and managers were based on the central question of the research. Some questions are as follows: 
1. Please describe your experience of flooding.
2. Please explain what factors you think are effective in an emergency evacuation.
3. What is your suggestion to managers and people for the emergency evacuation of residents involved in floods?
 Leading and exploratory questions were also used as follows: 
Pleasee explain more?
What did you mean?
Why? 
These questions were used to discover the depth of the phenomenon. Interviews were implemented at the first opportunity and typed using the Word software, Version 23.

Data analysis
Content analysis was done to extract categories and themes by carefully and continuously reviewing primary data. Granheim and Lundman method was used for data analysis [19].

Rigor
Reliability was recognized as an alternative to validity and reliability [20]. In the credibility, the researcher has a long-term involvement with the subject. This involvement has been from initial conceptualizations to frequent meetings with participants, implementation, and processing of findings with experts in the crisis field. As an executive director, the researcher has 20 years of experience related to disaster management. Also, he has proper communication and interaction with the participants, which has led to in-depth interviews and reliable data. In dependability, checking the findings was done by reviewing observers and participants. Supervisors reviewed the interviews, codes, and categories to check the analysis process and declare their accuracy. Qualitative research experts performed the process of coding and monitoring the themes. For conformability, the findings were given to participants who had not participated in the study to confirm appropriateness. For the transformability of the findings, the maximum diversity was done. The participants worked in organizations such as the Iranian Red Crescent, the governorate, the municipality, and the Water and Sewer Organization. They were different in terms of work experience, education, and gender. The generalizability of the study is cautious only for flood-prone areas, such as Poldokhtar (Iran). Lessons learned, outcomes, and methodology of the present study may be practical for mountainous regions with monsoon rains like Poldokhtar City, Iran. Finally, the experiences of participants were considered in this research. For this research to be repeatable, the processes were documented. The study processes were documented for reproducibility.

Results 
Based on gender, most participants were male, with a maximum age range of 31 to 40 years and 5 to 10 years of experience in disaster management (Table 1).


Also, the number of initial codes gotten the interviews was 1057 (607 for executive managers and 405 for citizens). At first, the regulations were split into ten categories and 50 sub-categories. After content analysis, the codes were merged into four main categories and ten sub-categories, including human factors (a subcategory of individual characteristics and risk perception), social factors (a sub-category of cultural conditions and financial status), geographical factors (a sub-category of urban location and climate conditions), and infrastructure factors (a sub-category of facilities related to transportation tools and routes, urban constructions, communication infrastructure, and meteorological facilities) (Table 2).



Human factors
Individual characteristics
The participants believed that individual characteristics, such as age, gender, and disability, were not considered during evacuation.
“They did not pay attention to the fact that it might be a child or a disabled or a deaf that did not understand the warnings...” (Citizen, male, 34 years old).

Risk perception
In the interviews, the lack of flood experience in the public’s mind, the lack of evacuation of offices near residential houses, and previous false warnings were misleading factors for evacuation. These factors made individuals mistrust the warnings.
«As elders and elderly people say, we have never experienced a flood of this magnitude... the officers told us to evacuate the house in previous floods, but nothing happened. The offices themselves were not empty either” (Citizen, male, 61 years old).

Social factors
Cultural conditions
Poldokhtar City, Iran, consists of a large and old multi-tribe settlement. The traditional context of kinship and nomadic relations has been one of the reasons for not welcoming tents. For the above reasons, people who decided to settle preferred to live in the homes of friends and acquaintances rather than in temporary camps.
«We preferred to seek refuge in our relatives. Life in the tent or camp is not pleasant to us… relatives and friends who were in the safe areas of the city called us and invited us to their houses.... living in relatives’ house is better and easier than living in a tent …” (Citizen, female, 55 years old).

Financial situation
Some participants stated that families with financial resources left their homes faster; however, people who did not have the cost of transportation or were afraid of stealing their property and livestock preferred staying in their homes. These cases were the reasons for not evacuating.
“Emptying home and shop objects and shifting of cattle needed a vehicle that we did not have … we had to stay in the house because we had no car and we didn’t have money for renting our belongings … it was stolen from the neighbor’s house when they were not there ...” (Citizen, female, 64 years old).

Geographic factors
Urban location
Mountain cities located at the confluence of rivers are immersed during monsoons.
“Our city is at the exit point of the catchments of Lorestan Province, Iran, and other provinces … steep mountainous slopes without vegetation cause flood in these parts ...” (Executive manager, male, 44 years old, 15 years job experience).

Climate conditions
Poldakhtar City, Iran, has a warm climate, and although there is little rain, it is exposed to floods due to its location between two valleys and vegetation.
“It is true that our city does not rain much, but when it rains, it floods ...” (Executive manager, male, 50 years old, 25 years job experience).

Infrastructure factors
Transportation tools and routes
The lack of trucks to transport household items was one of the problems. People who did not have a car had to pay an immense cost for transportation. Also, there were no machines to remove the mud. In cities with only one entrance route parallel to the river and no alternative route, the relief process becomes difficult in case of heavy rains and the destruction of the main route. These cases were considered as reasons for non-evacuation.
“We did not have any car to take our household items out of the house to a safe place, and we did not have any tools to remove the mud from our house ...” (Citizen, female, 56 years old).
“One of the two bridges over the river, where people used to travel, was broken ... there was no other emergency route to leave the city ... the city was surrounded by water ...” (Citizen, male, 28 years old).

Urban constructions
The increase in population and the lack of land for settlement caused the construction along the river. On the other hand, most of the participants mentioned that the city had many drought years. Over the years, rivers have been changed to recreation centers.
“We didn’t have rain for many years, so we thought we could build on the riverside…” (Citizen, male, 70 years old).

Communication infrastructure
The destruction of the communication infrastructure due to the proximity to the river and the lack of a satellite communication network made it difficult to give information, early warning, and evacuation.
“The telephone lines were cut, and we didn’t know how to inquire about our relatives ....” (Citizen, male, 33 years old).

Meteorological facilities
There were no meteorological facilities for flood forecasting and river flow estimation. On the other hand, meteorological reports were received from meteorological stations in other provinces. The participants believed the delay and inconsistency in the notification were significant reasons for non-evacuation. 
“We did not have a satellite network to find out about weather forecasts and to give early warning … The news reached us so late or contradictory that we missed the time to pay attention to the early warning ... we had no source to know the amount of possible rainfall or the river’s flow.”(Executive manager, female, 39 years old, 8-year job experience).

Discussion
In this study, the factors influencing evacuation included human, social, geographical, and infrastructure issues. Human factors had two sub-categories, namely individual characteristics and risk perception. Burnside, in a study investigating the influence of information and risk perception on the storm evacuation decision-making of New Orleans residents in the US, found that having the necessary information and understanding of the risk of disasters are essential factors in making decisions on evacuation [21]. Lechowska believes that the perception of risk as one of the human factors is effective in flood management [22]. People who ignore the flood risk face a big concern and challenge in flood management. Knowing the factors affecting the mental understanding of the flood risk can solve this trouble and create the necessary motivation in the residents of areas at risk to take measures to diminish the negative impacts of floods and increase the resistance of communities against the risk of floods [12, 23, 24]. Jang et al. also discussed the importance of knowledge, attitude, and evacuation behavior in terms of gender, education level, and local conditions of the region [25]. In the current study, individual factors and attitudes are factors influencing evacuation. People in flood-prone areas could not leave their residences due to their dependence on their place of residence, lack of vehicles to carry goods and livestock, and physical problems.
Social factors had two sub-categories: Cultural conditions and financial situation status. Heath believes that the lack of finances and the conditions of the families affect the emergency evacuation, such that the number of livestock, the presence of older people and children, and the lack of understanding of the conditions for leaving their property are effective on the speed of the evacuation process [26]. Emphasizing the importance of family financial conditions, Lim et al. showed that households should have suitable vehicles for long-term evacuations [27]. Armenakis et al. believe that a flood map alone cannot determine and assess risks, and socio-economic factors are critical to risk assessment. Therefore, an approach integrating flood maps, socio-economic parameters, infrastructure, and services is essential [28]. According to Munyai et al., social and economic elements are more vulnerable to floods than environmental elements, while social aspects are higher than economic ones [29].
In this study, the geographical factors had two sub-categories: Urban location and climate conditions. Yari et al. paid attention to the principles of urban and construction safety and the improvement of urban flood management by responsible and participating organizations to manage floods and reduce mortality [30]. Galateia et al. state that social vulnerability and community conditions can turn a natural event into a deadly disaster. Also, climatological factors, such as rainfall, hydrological parameters (humidity and soil permeability), and geomorphological characteristics (housing, depth, and type of soil) are helpful in flood disaster management [31]. Ibarra believes that short-term and medium-term corrective measures are necessary to control floods. Geographic measures include structural and non-structural mitigation. Drainage channels in flood-prone areas are among these measures [32]. In the present study, the lack of suitable urban infrastructure, including the wear and low height of bridges, and the lack of alternative transportation routes were among the factors affecting the evacuation.
Infrastructure factors had four subclasses of facilities related to transportation tools and routes, urban constructions, communication infrastructure, and meteorological facilities. Lane believes that land use and urban development can control floods. He also refers to using appropriate emergency facilities and equipment [33]. Managers and urban planners should pay attention to the equipment supply plan in the planning stage before disasters [34-41]. In the current study, the lack of transportation and dredging equipment has caused problems in the emergency evacuation route. Botzen et al. showed that construction buildings’ location and quality are essential in flood disaster management. These factors reduce the damage related to evacuation and prevent property damage. In addition, the inappropriate structure of the city is an influential factor in evacuation [42]. 

Conclusion
The human, social, geographical, and infrastructure factors were the influential factors of evacuation in the study; therefore, it is recommended that managers help the flood response stage by training manpower and increasing the risk perception, increasing the financial credit of cities exposed to floods, and providing meteorological equipment and an early warning system. It is suggested that future studies investigate the motivational factors of managers and people to improve the understanding of flood risk.

Study limitations
Since the interviews were conducted several years after the flood event, some experiences may have been forgotten. The small number of female participants, especially executive managers, was another limitation of this study.

Ethical Considerations
Compliance with ethical guidelines
This design has been approved by the Research Ethics Committee of Hamadan University of Medical Sciences (Code: IRUMSHA.REC.1400.039, No.: 140002281400). 

Funding
This study is a part of the MPH thesis of Changiz Ahmadi, approved by Department of MPH, School of Public Health, Hamadan University of Medical Sciences.

Authors' contributions
Conceptualization, methodology, data generation, data interpretation and writing the manuscript: All authors; Data collection: Changiz Ahmadi; Data analysis: Arezou Karampourian and Mohammad Reza Samarghandi; Final approval: All authors.

Conflict of interest
The authors declared no conflict of interest.

Acknowledgments
The authors thank the Hamadan University of Medical Sciences for approving the study and all participants in this study.


References

1. National Disaster Management Authority (NDMA). National policy on disaster management 2009. India: National Disaster Management Authority (NDMA); 2009. [Link]
2. Saharia M, Jain A, Baishya RR, Haobam S, Sreejith O, Pai D, et al. India flood inventory: Creation of a multi-source national geospatial database to facilitate comprehensive flood research. Natural Hazards. 2021; 108:619-33. [DOI:10.1007/s11069-021-04698-6]
3. Huang SHK, Lindell MK, Prater CS, Wu HC, Siebeneck LK. Household evacuation decision making in response to Hurricane Ike. Natural Hazards Review. 2012; 13(4):283-96. [DOI:10.1061/(ASCE)NH.1527-6996.0000074]
4. Quandt A, Kimathi YA. Perceptions of the effects of floods and droughts on livelihoods: Lessons from arid Kenya. International Journal of Climate Change Strategies and Management. 2017; 9(3):337-51. [DOI:10.1108/IJCCSM-11-2014-0132]
5. Jones RL, Guha-Sapir D, Tubeuf S. Human and economic impacts of natural disasters: Can we trust the global data? Scientific Data. 2022; 9(1):572. [DOI:10.1038/s41597-022-01667-x] [PMID]
6. Peyravi M, Ahmadi Marzaleh M. The effect of the US sanctions on humanitarian aids during the great flood of Iran in 2019. Prehospital and Disaster Medicine. 2020; 35(2):233-4. [DOI:10.1017/S1049023X20000242] [PMID]
7. Khantong S, Ahmad MN. An ontology for sharing and managing information in disaster response: In flood response usage scenarios. Journal on Data Semantics. 2020; 9(1):39-52. [DOI:10.1007/s13740-019-00110-6]
8. Bin Wan Abdul Majid W, Brown E, Bin Osman S, Bin Asan G, Bin Osman A, Bin Samsudin R, et al. Flood forecasting and warning for Muar River: Non-structural measures for flood mitigation. Paper presented at: 37^th IAHR World Congress "Learning from the Past for the Future". 13-18 August 2017; Kuala Lumpur, Malaysia. [Link]
9. Wang Z, Huang J, Wang H, Kang J, Cao W. Analysis of flood evacuation process in vulnerable community with mutual aid mechanism: An agent-based simulation framework. International Journal of Environmental Research and Public Health. 2020; 17(2):560. [DOI:10.3390/ijerph17020560] [PMID]
10. Korolov V, Kurowska K, Korolova O, Zaiets Y, Milkovich I, Kryszk H. Methodology for determining the nearest destinations for the evacuation of people and equipment from a disaster area to a safe area. Remote Sensing. 2021; 13(11):2170. [DOI:10.3390/rs13112170]
11. Thompson RR, Garfin DR, Silver RC. Evacuation from natural disasters: A systematic review of the literature. Risk Analysis. 2017; 37(4):812-39. [DOI:10.1111/risa.12654] [PMID]
12. Ahmadi C, Karampourian A, Samarghandi MR. Explain the challenges of evacuation in floods based on the views of citizens and executive managers. Heliyon. 2022; 8(9):e10759. [DOI:10.1016/j.heliyon.2022.e10759] [PMID]
13. Wood N, Jones J, Schmidtlein M, Schelling J, Frazier T. Pedestrian flow-path modeling to support tsunami evacuation and disaster relief planning in the US Pacific Northwest. International Journal of Disaster Risk Reduction. 2016; 18:41-55. [DOI:10.1016/j.ijdrr.2016.05.010]
14. Alonso Vicario S, Mazzoleni M, Bhamidipati S, Gharesifard M, Ridolfi E, Pandolfo C, et al. Unravelling the influence of human behaviour on reducing casualties during flood evacuation. Hydrological Sciences Journal. 2020; 65(14):2359-75. [DOI:10.1080/02626667.2020.1810254]
15. Thi An T, Izuru S, Narumasa T, Raghavan V, Hanh LN, Van An N, et al. Flood vulnerability assessment at the local scale using remote sensing and GIS techniques: A case study in Da Nang City, Vietnam. Journal of Water and Climate Change. 2022; 13(9):3217–38. [DOI:10.2166/wcc.2022.029]
16. Borowska-Stefańska M. Assessment of material losses within areas exposed to floods in small towns in Łódź Region. Space-Society-Economy. 2016; 16:7-27. [DOI:10.18778/1733-3180.16.01]
17. Yaghmaei F, Mohammadi S, Alavi Majd H. Developing and measuring psychometric properties of “quality of life questionnaire in infertile couples”. International Journal of Community Based Nursing & Midwifery. 2013; 1(4):238-45. [Link]
18. Saunders B, Sim J, Kingstone T, Baker S, Waterfield J, Bartlam B, et al. Saturation in qualitative research: Exploring its conceptualization and operationalization. Quality & Quantity. 2018; 52(4):1893-907. [DOI:10.1007/s11135-017-0574-8] [PMID]
19. Graneheim UH, Lundman B. Qualitative content analysis in nursing research: Concepts, procedures and measures to achieve trustworthiness. Nurse Education Today. 2004; 24(2):105-12. [DOI:10.1016/j.nedt.2003.10.001] [PMID]
20. Behzad I, Elahe M. Patient safety culture and spiritual health in the operating room: An Iranian exploratory qualitative study. Journal of Religion and Health. 2022; 62(4):2359-74. [DOI:10.1007/s10943-022-01531-7] [PMID]
21. Burnside R, Miller DS, Rivera JD. The impact of information and risk perception on the hurricane evacuation decision-making of greater new orleans residents. Sociological Spectrum. 2007; 27(6):727-40. [DOI:10.1080/02732170701534226]
22. Lechowska E. What determines flood risk perception? A review of factors of flood risk perception and relations between its basic elements. Natural Hazards. 2018; 94:1341-66. [DOI:10.1007/s11069-018-3480-z]
23. Becker G, Aerts JCJH, Huitema D. Influence of flood risk perception and other factors on risk-reducing behaviour: A survey of municipalities along the Rhine. Journal of Flood Risk Management. 2014; 7(1):16-30. [DOI:10.1111/jfr3.12025]
24. Bradford R, O’Sullivan JJ, Van der Craats I, Krywkow J, Rotko P, Aaltonen J, et al. Risk perception-issues for flood management in Europe. Natural Hazards and Earth System Sciences. 2012; 12(7):2299-309. [DOI:10.5194/nhess-12-2299-2012]
25. Jang LJ, Wang JJ. Disaster resilience in a Hakka community in Taiwan. Journal of Pacific Rim Psychology. 2009; 3(2):55-65. [DOI:10.1375/prp.3.2.55]
26. Heath SE, Kass PH, Beck AM, Glickman LT. Human and pet-related risk factors for household evacuation failure during a natural disaster. American Journal of Epidemiology. 2001; 153(7):659-65. [DOI:10.1093/aje/153.7.659] [PMID]
27. Lim MB. Understanding evacuation decision, departure timing and destination choice of households in high flood risk areas using discrete choice model [PhD dissertation]. Pathum Thani: Thammasat University; 2016.
28. Armenakis C, Du EX, Natesan S, Persad RA, Zhang Y. Flood risk assessment in urban areas based on spatial analytics and social factors. Geosciences. 2017; 7(4):123. [DOI:10.3390/geosciences7040123]
29. Munyai RB, Musyoki A, Nethengwe NS. An assessment of flood vulnerability and adaptation: A case study of Hamutsha-Muungamunwe village, Makhado municipality. Jàmbá: Journal of Disaster Risk Studies. 2019; 11(2):1-8. [DOI:10.4102/jamba.v11i2.692]
30. Yari A, Ardalan A, Zarezadeh Y, Rahimiforoushani A, Soufi Boubakran M, Bidarpoor F, et al. Investigating the risk factors of flood deaths in Iran. Journal of Injury and Violence Research. 2022; 14(2 Suppl 1). [PMCID]

31. Galateia Terti1, Isabelle Ruin, Sandrine Anquetin, Jonathan J. Gourley. Vulnerability situations associated with flash flood casualties in the United States. Paper presented at: American Geophysical :union: Fall Meeting. Aguat 2015; San Fransisco, USA. [Link]
32. Ibarra EM. A geographical approach to post-flood analysis: The extreme flood event of 12 October 2007 in Calpe (Spain). Applied Geography. 2012; 32(2):490-500. [DOI:10.1016/j.apgeog.2011.06.003]
33. Lane SN. Natural flood management. WIREs Water. 2017; 4(3):e1211 [DOI:10.1002/wat2.1211]
34. Sheikhbardsiri H, Yarmohammadian MH, Khankeh H, Khademipour G, Moradian MJ, Rastegarfar B, et al. An operational exercise for disaster assessment and emergency preparedness in south of Iran. Journal of Public Health Management and Practice. 2020; 26(5):451-6. [DOI:10.1097/PHH.0000000000000815] [PMID]
35. Rezaei F, Maracy MR, Yarmohammadian MH, Sheikhbardsiri H. Hospitals preparedness using WHO guideline: A systematic review and meta-analysis. Hong Kong Journal of Emergency Medicine. 2018; 25(4):211-22. [DOI:10.1177/1024907918760123]
36. Sheikhbardsiri H, Doustmohammadi MM, Mousavi SH, Khankeh H. Qualitative study of health system preparedness for successful implementation of disaster exercises in the Iranian context. Disaster Medicine and Public Health Preparedness. 2022; 16(2):500-9. [DOI:10.1017/dmp.2020.257] [PMID]
37. Khademipour G, Nakhaee N, Anari SMS, Sadeghi M, Ebrahimnejad H, Sheikhbardsiri H. Crowd simulations and determining the critical density point of emergency situations. Disaster Medicine and Public Health Preparedness. 2017; 11(6):674-80. [DOI:10.1017/dmp.2017.7] [PMID]
38. Sheikhbardsiri H, Khademipour G, Davarani ER, Tavan A, Amiri H, Sahebi A. Response capability of hospitals to an incident caused by mass gatherings in southeast Iran. Injury. 2022; 53(5):1722-6. [DOI:10.1016/j.injury.2021.12.055] [PMID]
39. Khademipour G, Sheikhbardsiri H. Disaster risk assessment of primary healthcare facilities in South East of Iran: A study with approach of disaster risk reduction. Disaster and Emergency Medicine Journal. 2022; 7(1):11-20. [DOI:10.5603/DEMJ.a2022.0002]
40. Nejadshafiee M, Nekoei-Moghadam M, Bahaadinbeigy K, Khankeh H, Sheikhbardsiri H. Providing telenursing care for victims: A simulated study for introducing of possibility nursing interventions in disasters. BMC Medical Informatics and Decision Making. 2022; 22(1):54. [DOI:10.1186/s12911-022-01792-y] [PMID]
41. Beyramijam M, Khankeh H, Shahabi-Rabori MA, Aminizadeh M, Sheikhbardsiri H. Hospital disaster preparedness in Iranian province: A cross-sectional study using a standard tool. American Journal of Disaster Medicine. 2021; 16(3):233-9. [DOI:10.5055/ajdm.2021.0406] [PMID]
42. Botzen WJW, Van Den Bergh JCJM. Managing natural disaster risks in a changing climate. Environmental Hazards. 2009; 8(3):209-25. [DOI:10.3763/ehaz.2009.0023]