1 Introduction

By law, disaster coordination in Iceland is under the umbrella concept of Civil Protection. It is the responsibility of the police (Act on Civil Protection, Lög um almannavarnir 2008), at both the local (local chiefs of police) and national levels (National Commissioner of the Icelandic Police). In 2002, prompted by information from the scientific community, the local chief of police requested funding from the Ministry of Justice for a risk analysis and organized preparedness activities for a possible eruption of the Eyjafjallajökull volcano and flooding due to melting of the icecap (Thorkelsson 2013). In 2010, the police chief coordinated various response activities (Thorkelsson 2013). Previous research on the response to the 2010 Eyjafjallajökull volcano (EFJV) eruption shows that whilst emergency management agencies in Iceland responded immediately to ensure people’s safety and disseminated geophysical information during the crisis, specific information with respect to ongoing problems associated with the ash was lacking, including specialized information for farmers (Bird and Gisladottir 2012).

Delivering the right information to the right people, in the right format, in the right place and at the right time is a key disaster-related activity (Iannella and Henricksen 2007), requiring good organization, concerted efforts and allocation of resources, along with knowledge of hazard phenomena (Alexander 2006). This study looks for explanations on why information to the farmers was inadequate, and what could have been done differently, in order to learn lessons for livelihood professions in general. The conceptual framework used here is a livelihood-disaster perspective. As implied, the perspective is a combination of a livelihood perspective and a disaster perspective. The Sustainable Livelihoods Framework (SLF) (DFID 1999) is used to define the farming perspective and guided the data collection. The eight disaster-related objectives of Disaster-Function Management (Thorvaldsdóttir and Sigbjörnsson 2014) define the disaster perspective and guide the analysis. The combined perspective is presented in a Venn diagram. Information and data are collected for each component in the Venn diagram from data sources on farming and the EFJV. The data sources include scientific literature and secondary information available prior to the eruption, field reports and interviews. Information from a reconnaissance mission performed by farming actors four weeks after the onset of the EFJV eruption provides insight into the problems and response within the farming community. Eleven actors identified in the report are interviewed for information about their pre-eruption, real-time and post-eruption activities. After the interviews, seven actors were defined as direct farming actors and included in further analysis, based on each disaster-related objective.

The results show that farming actors were not included in any pre-eruption coordination by the police, did not perform any pre-event disaster-related activities specific to the pending EFJV eruption due to limited knowledge of the risk of an eruption, but did respond to the eruption. An analysis of the information shows that risk analysis, mitigation and preparedness activities performed by farming actors before the eruption could have improved the response of the farming actors and thus better addressed the needs of farmers.

2 Method

The method involves four steps (Fig. 1). First, a livelihood-disaster perspective is developed (Sect. 2.1). The perspective is a combination of a livelihood and disaster perspective. The former is based on the SLF and the latter on the DFM objectives. The second step is identifying data sources for each component of the livelihood-disaster perspective in relation to farming, Eyjafjallajökull, and the 2010 EFJV eruption (Sect. 2.2). The third step is to extract relevant information from the data sources, thus presenting the farming-EFJV eruption perspective (Sect. 3). The final step is analysing the information in Sect. 3 for each DFM objective (Sect. 4).

Fig. 1
figure 1

The method (numbers relate to sections)

Full size image

2.1 Livelihood-disaster perspective development

2.1.1 Livelihood perspective

The Sustainable Livelihoods Framework (SLF) (DFID 1999) is used to outline a livelihood perspective. The SLF “presents the main factors that affect people’s livelihoods and typical relationships between these”. The “aim is to help stakeholders with different perspectives to engage in structured and coherent debate about the many factors that affect livelihoods, their relative importance and the way in which they interact”. The framework captures how people’s decisions on their livelihood are influenced by regulatory opportunities and restrictions that are produced by transforming actors, and the vulnerability context within which they live.

The SLF consists of five main components (see Fig. 2): (1) outcomes of livelihoods, (2) strategies to reach these goals, (3) livelihood assets (human, natural, physical, financial and societal), (4) transformation structures (hereafter termed actors as the study is limited to actors and their activities) and policies affecting the assets and, finally, (5) issues of vulnerability, such as shocks (including natural disasters), monetary trends and changing seasons.

Fig. 2
figure 2

The Sustainable Livelihoods Framework (Based on Fig. 1 DFID 1999)

Full size image

The SLF allows for the perspectives of numerous actors that all relate to livelihood and thus create a general livelihood perspective. The SLF defines actors by types of organizations “that set and implement policy and legislation, deliver services, purchase, and trade and perform all manner of other functions that affect livelihoods”. Table 1 (extracted from DFID 1999) depicts actor types and levels, and various types of units for actor analysis. All of these help define actor perspectives.

Table 1 Actor categories and aspects of analysis (DFID 1999)
Full size table

Various disaster studies have used a sustainable livelihoods approach, such as Cannon et al. (2005) (reducing social vulnerability from a development agency perspective), Twigg (2004) (project cycle approach from a primarily humanitarian and development agency perspective) and studies specifically on volcanic risk (see overview presented in Kelman and Mather 2008). Odero studied the SLF (2006) and suggests that relevant, accurate and timely information be added as the sixth asset. This is a key aspect of this study.

2.1.2 Disaster perspective

The disaster perspective is presented through a set of disaster-related objectives associated with Disaster-Function Management (DFM) (see Table 2). DFM is derived from Comprehensive Emergency Management, which was defined in 1979 (NGA 1979) as the management of disaster-related activities in four phases: mitigation, preparedness, response and recovery. However, since then the concept of phases as the appropriate approach to disaster management has been challenged (Neal 1997; McEntire et al. 2002) and the idea of functions or functional areas suggested (McEntire et al. 2002; Quarentelli 1996). Thorvaldsdottir and Sigbjörnsson (2014) outline how current scientific and practical knowledge has surpassed the four-phase framework and describe why and how phases are replaced by objectives for management purposes. DFM is set within classical management theory, governed by “management by objectives”, and has been introduced as a contribution towards a new paradigm for addressing disaster-related management. Classical management theory states that “management by objectives”, as a concept, is an integrated management system to manage an organization by aligning the entire managerial effort with specific performance goals (Ghuman 2010). Applying “management by objectives” to coordination creates a coordination process that meets chosen objectives, thus giving the process a clear focus.

Table 2 Disaster functions and objectives (Thorvaldsdottir and Sigbjörnsson 2014)
Full size table

2.1.3 Livelihood-disaster perspective

The livelihood-disaster perspective is presented by a Venn diagram consisting of three overlapping rings (Fig. 3). First, the five basic components of the SLF are placed in the diagram. The Livelihood Strategies and Livelihood Outcomes are placed at the centre, representing the desired and actual outcomes when adopting the SLF. The three remaining components, Assets, Shock, and Transporting Structures and Processes (actors), are placed in each of the outer rings. The overlaps are filled with aspects that link the two basic SLF components on either side. The linking components are as follows:

Fig. 3
figure 3

Venn diagram for livelihood-disaster perspective

Full size image

  • Actor daily activities, linking Transporting Structures and Processes (actors) and Assets, addressing how actors manage assets.

  • Disaster-related activities, linking Transporting Structures and Processes (actors) and Shock, addressing how actors manage shock-related issues.

  • Damages, injury and disruptions, linking Assets and Shock. When addressing future disasters, the link is vulnerability to damages, injury or disruptions. When addressing actual disasters, the link is actual damages, injury or disruptions

2.2 Data sources

The following data sources were identified as relevant to the respective components in Fig. 3 for the farming-EFJV context:

  1. 1.

    Farming livelihood strategies

    1. a.

      Report on government farming strategy (Jóhannesson and Agnarsson 2005)

    2. b.

      Discussions with a teacher at the Agricultural University of Iceland

    3. c.

      Discussions with a farmer in the affected area

  2. 2.

    Shock (volcanic ash)

    1. a.

      Farming next to a volcano, literature-based discussion

    2. b.

      Literature survey for an overview of volcanic activity in Iceland

    3. c.

      Literature survey for an overview of the EFJV hazard and eruption 1991–2010

  3. 3.

    Damages, injuries and disruption

    1. a.

      Literature survey on the vulnerability of physical and natural farming assets due to ash

    2. b.

      Actual damages, injuries and disruption described in Farmers’ Reconnaissance Report (2010)

  4. 4.

    Actor identification and classification

    1. a.

      Farmers’ Reconnaissance Report (2010)

    2. b.

      SLF actor classification (Table 1)

  5. 5.

    Actor daily activities

    1. a.

      Daily objectives stated on actor websites

  6. 6.

    Actor disaster-related activities

    1. a.

      Interviews with actors

    2. b.

      Literature on the EFJV before and after the eruption

    3. c.

      EFJV activity reports acquired from websites

    4. d.

      Act on Civil Protection

  7. 7.

    Farming livelihood outcome (situation in 2010)

    1. a.

      Farmers’ Reconnaissance Report (2010)

  8. 8.

    Farming assets

    1. a.

      Farming assets are apparent in the text, e.g., animals, buildings, land, but they are not presented here in a separate discussion.

2.2.1 Farmers’ Reconnaissance Report 2010

The key source of information about the status of farmers and their farms and for determining relevant actors is a report from a reconnaissance (abbreviated recon) performed 11–12 May 2010, approximately 4 weeks after the explosive episode of the eruption began (Farmers’ Recon Report 2010). The Farmers’ Association, The Agricultural Association of South Iceland and other regional farmers’ associations, and the Ministry of Agriculture organized the recon. Its purpose was to bring relevant actors together to share information and ideas, collect information, maximize the use of resources and plan assistance to the affected farmers (Farmers’ Recon Report 2010). The first day was spent on site visits and the second on planning and developing recommendations. Approximately 35 people were divided into 2–3 people teams and visited approximately 120 farms (see Fig. 4). The organizers decided on the following questions to ask during the site visits:

Fig. 4
figure 4

The area of reconnaissance a Location of 120 farms visited during the 11–12 May 2010 reconnaissance. The Icelandic volcanic system and plate boundaries are depicted in the map in the upper left corner. The rectangle in this map shows the reconnaissance area. The main figure is an enlarged view of this area. All the buildings in the recon area are farm buildings. b Ash from Eyjafjallajökull Volcano, 17 April 2010. NASA image by Jeff Schmaltz, MODIS Rapid Response Team. Source: http://earthobservatory.nasa.gov/NaturalHazards/view.php?id=43690, extract 6 May 2014

Full size image
  1. 1.

    How many animals are on the farm (cows, sheep, horses)?

  2. 2.

    What is the housing need for animals (cows, sheep with lambs, horses)?

  3. 3.

    Is there sufficient feed and (uncontaminated) running water for all the animals?

  4. 4.

    For how many days will the current quantity of hay last?

  5. 5.

    Has the farmer secured extra hay, if needed?

  6. 6.

    If the eruption ends within 15 days, how many animals need to be transported, and how much hay is needed?

  7. 7.

    If the eruption continues into the summer, how many animals need to be transported, and how much hay is needed?

  8. 8.

    What type of solutions does the farmer recommend?

  9. 9.

    What damage has occurred that will be insured by the Farmers‘ Insurance Fund?

  10. 10.

    Is extra manual labour required?

When asked what prompted the recon, the organizers stated they had gradually realized their own lack of understanding of the magnitude of the problems amongst farmers. Four explanations are stated here. First of all, questions arose about housing and feeding sheep. It was now spring so sheep were inside in pens, and lambing season was approaching. Sheep are usually kept inside during the winter months and put outside 3–4 days after their lambs are born. Then, after all lambs are born, the ewes, lambs and rams are driven to the highlands for the summer, and in the autumn the sheep are collected and the lambs slaughtered. The ash could hinder farmers putting the sheep outside and render the highlands unusable. This meant that the farmers had to somehow cope with an expanding sheep population inside the pens, causing overcrowding. Second, field staff members assisting the affected farmers, for example, from South-Agri, were beginning to show fatigue and needed support. A multi-organization recon would provide an overview of the situation and generate ideas on how to support them. Third, there was concern over the possible impact if the wind changed direction. The winds had been northerly; if they changed to easterly winds, new farming areas would be affected. The organizers wondered what problems this would create, and how they should respond to assist farmers. A recon would allow organizers to understand the problems and concerns of the farmers (i.e., obtain the farmer’s perspective) and seek their opinion on possible solutions. Finally, the organizers wanted to support the affected farming population by showing their concern. The eruption was still going on, and the end of the eruption could not be predicted.

2.2.2 Interviews

Interviews (Aug–Nov 2013) provided information about the activities of the actors in respect of the eight objectives of DFM as related to EFJV 1991–2013. A representative of each actor was asked eight questions; one question on each of the eight DFM objectives: what activities did you perform towards this objective in relation to the EFJV hazard/eruption? Discussions with a teacher at the Agricultural University of Iceland and with a farmer in the affected area were also used to generate information.

3 Farming-EFJV perspective

This section presents extracts from the data sources listed in Sect. 2.2 relevant for farming in the vicinity of the EFJV. The tables and figures created herein relate to the Venn diagram in Fig. 3, as depicted in Fig. 5.

Fig. 5
figure 5

Information defining a farming-disaster coordination perspective in Iceland regarding the 2010 EFJV eruption

Full size image

3.1 Farming livelihood strategies

“Livelihood strategies” is the term used to designate the range and combination of activities and choices that people make/undertake in order to achieve livelihood goals (DFID 1999). Livelihood goals are broader than a pure business goal. For example, investments, such as in technology, may be for the purpose of increased production to increase income or to ease physical labour and shorten working hours, leading to healthier and longer lives, giving opportunities to spend time on other things, whether farming-related or leisure (Jóhannesson and Agnarsson 2005). Livelihood goals are therefore affected by an individual’s motive for choosing a particular livelihood. Farming is the business of growing crops and/or raising livestock (Jóhannesson and Agnarsson 2005). There are many motives for choosing farming, such as family tradition, wanting to be your own boss, desire to work with animals or desire to work with machinery (Gislason 2013). Attachment to the land can also influence choosing farming as a livelihood (Eyvindsson 2013). A farming strategy is based on decisions on a variety of livelihood variables that characterize farming, including (Jóhannesson and Agnarsson 2005):

  • Type of production

  • Type of livestock

  • Proximity to market

  • Level of technology

  • Level of debt

The farming market is unique in that its products are seen as essential for human survival. This is reflected in product demand being relatively insensitive to market price (Jóhannesson and Agnarsson 2005). Government strategy in Iceland regulates to some degree both the supply and price of these products through various regulatory tools, such as production quotas and subsidies, and includes ensuring food security on the island and avoiding total migration to the capital (Jóhannesson and Agnarsson 2005). Government agencies monitor animal welfare and production and control the risk of disease by establishing control areas. Live animals, hay and other farm products may not move between these areas. The 2014 annual meeting of the Farmers’ Association concluded that the association should work with the government on a long-term farming strategy in association with the specialized farming organizations (for cattle, eggs, pig farming, etc.) (Bændabladid Newspaper 2014), showing farmers’ interest in influencing government farming strategy.

3.2 Shock (volcanic ash)

3.2.1 Volcanic activity in Iceland

Iceland is one of the most active and productive volcanic regions in the world. Its eruption frequency is up to 20 events per century (Thordarson and Höskuldsson 2008). Eruptions are broadly grouped into effusive eruptions, where over 95 % of the erupted magma is lava, explosive eruptions if more than 95 % of the erupted magma is tephra and mixed eruptions if they include tephra but less than 95 % (Thordarson and Larsen 2006). Explosive events in Iceland are dictated by sub-glacial events and are more common than effusive activity or mixed eruptions (Thordarson and Larsen 2006). A summary of verified eruptive events is presented in Table 3. A total of 2400 eruptions have been identified from post-glacial times (last 11,000 years). In historical times (last 1100 years), a total of 192 single eruptive events have been verified, based on volcanic products, and 33 were mentioned in oral accounts. In addition, 13 fires (multiple events) lasting for months or years have been verified in historical accounts.

Table 3 Volcanic activity in Iceland, verified events
Full size table

Four volcanoes or volcanic systems have been most active in historical times: Katla Volcano under the Mýrdalsjökull icecap (21 eruptions), Grimsvötn/Laki (~70), Hekla (23 eruptions) and Bardarbunga/Veidivötn (23 eruptions) (Gudmundsson et al. 2008). Both Hekla and Katla Volcanoes have produced ash layers over 20 cm thick over the area surrounding Eyjafjallajökull (Fig. 3 in Gudmundsson et al. 2008, reprinted as Fig. 6). Figure 6 shows “areas that may receive over 20 cm of tephra fall in major explosive eruptions indicated with circles around volcanoes or fissure swarms, where explosive activity is common or the dominant mode of activity. The radius of each circle is defined as the distance to the 20-cm isopach along the axis of thickness for the largest historical and prehistoric explosive eruptions of each volcano. Also shown are populated areas and the main route, Highway 1. The volcanic zones are shown with a shade of grey” (Gudmundsson et al. 2008).

Fig. 6
figure 6

Ash hazard map from 2008 showing 20-cm isopach radius (Fig. 3 in Gudmundsson et al. 2008)

Full size image

3.2.2 Eyjafjallajökull volcano

Eyjafjallajökull (glacier) is in South Central Iceland in the Eastern Volcanic Zone. Prior to the 2010 eruption, the volcano is believed to have erupted four times. The most recent was in 1821–1823. Chronicles provide relatively detailed accounts of a small intermittent explosive eruption, producing ash levels harmful to livestock (Larsen 1999). The ash led to considerable fluoride poisoning and negatively impacted farming activities in general, including reduction in livestock (Larsen 1999). Historical accounts indicate that there may have been an eruption in 1612 or 1613 (Vetter 1983) although the physical evidence of ash is inconclusive (Larsen 2013), and it may have been its more active neighbour Katla Volcano that the accounts refer to Sturkell et al. (2003). Dugmore et al. (2013) have recently identified and dated floods that verify flank eruptions in ca. 920 AD and in the sixth to seventh century. Dugmore et al.’s work is referenced as “unpublished results” in Gudmundsson et al. (2005), so it was known prior to the 2010 eruption.

A chronological outline of key geological and scientific events from 1991 to 2010 and post-eruption activities, based on Sigmundsson et al. (2010), Gudmundsson et al. (2010, 2012) and other references, is presented in Table 4. The recognition of geophysical precursors to volcanic activity is a primary challenge in volcanic monitoring (Vogfjörd et al. 2005). Prior to the 1990s, only about a handful of earthquakes associated with the Eyjafjallajökull volcanic system had been detected (Pedersen and Sigmundsson 2006); after that the situation gradually changed. The 2010 eruption was the culmination of two decades of intermittent volcanic unrest, providing ample precursors, many of which were captured by scientists. An effusive flank eruption began on 20 March, ended on 12 April and caused no damage (Gudmundsson et al. 2010). Increased seismic activity was noted by seismometers in EFJV at 2300 local time on 13 April. The summit eruption commenced at 0115. A plume was visible in the early hours of the 14th, and the first signs of glacial flooding were seen at 0650 at a gauging station (Gudmundsson et al. 2010). Concern arose about health risks from fallout, because ash can transport acids as well as toxic components, such as fluoride, aluminium and arsenic (Gislason et al. 2011). The summit eruption lasted 39 days, till 22 May. Persistent north-westerly winds transported the ash towards the southeast (Gudmundsson et al. 2012). The intensity of tephra fallout varied throughout the eruption. As explained by Gudmundsson et al. (2012, Fig. 4, reprinted as Fig. 7), Fig. 7 shows isopach maps (thickness in cm) of tephra deposition on land (a) during the first 3 days of the first explosive phase, erupting from a water-filled vent (14–16 April), (b) during the second part of the first explosive phase (17 April until early 18 April) and (c) total fallout on land in the eruption (14 April–22 May) and estimated fallout thickness (dotted lines) to the south and south-east of Iceland. Based on Fig. 7, the entire ash fallout in the recon area is estimated to be less than 5 cm thick over the main area. Less than 10 cm were measured at the foothills of the volcano (Gudmundsson 2014).

Table 4 Geophysical events of EFJV from 1991 to 2013
Full size table
Fig. 7
figure 7

Ash distribution from the 2010 Eyjafjallajökull eruption (Fig. 4 in Gudmundsson et al. 2012)

Full size image

3.3 Damage, injury and disruption

Large explosive volcanic eruptions can potentially distribute heavy ashfall across large areas of agricultural land (Wilson et al. 2011a, b). Farming near an active volcano has both positive and negative aspects (Kelman and Mather 2008). Volcanic soils are often physically and chemically suited for growing crops, and long periods between eruptions make agriculture common in volcanic regions (Cronin et al. 1998, Wilson et al. 2011a, b). The negative aspects involve the susceptibility of assets, people and farming processes to damage, injury and disruption, respectively, due to eruptions. The problems the farmers faced in Iceland in 2010 were due to ashfall and remobilized ash. The nearest farms were roughly 20 km from the summit.

3.3.1 Pre-eruption knowledge on vulnerability

A literature survey of papers published before the 2010 EFJV eruption was done to gain understanding of the information available prior to the eruption on vulnerable conditions for farming assets due to ash (see Table 5). Two of the papers in Table 5 are published after the eruption: one relays information available before the eruption regarding lack of visibility during an eruption (Bird and Gisladottir 2012) and the other describes an eruption in 1991 (Wilson 2011a, b). The papers discuss the vulnerability of mainly natural and physical farming assets, such as livestock, farming processes, buildings, infrastructure, utilities and mechanical equipment. More recent papers illustrate the ever-expanding knowledge of volcanic vulnerability (see, for example, Wilson et al. 2012a, b and Magill et al. 2013).

Table 5 Farming vulnerability due to volcanic ash
Full size table

Impacts from past eruptions provide insight into the vulnerability associated with farming close to volcanoes (e.g. Kelman and Mather 2008, Lebon 2009). All forms of agricultural production are vulnerable to the physical and chemical effects of volcanic ashfall, including vegetation, soils, animal health, human health and infrastructure (Wilson et al. 2011a, b). The impact of eruptions in Iceland includes people being killed (by drowning, lightning and CO2 suffocation), livestock deaths (by tephra fall leading to lack of access to feed and fluorine poisoning) and crop damage from ash coverage, leading to farm abandonment and famine. Heavy tephra fall in past Katla eruptions has been known to obstruct visibility (Bird et al. 2010), causing total darkness during the day (Bird and Gisladottir 2012).

Fluoride is a potentially toxic element for people and animals and can severely impact sheep through diarrhoea, loss of appetite and general prostration and dental fluorosis at different concentration levels (Kristinnsson et al. 1997). The degassing of magma chemicals produces fluorine, sulphur and chlorine in the atmosphere (Thordarson et al. 1996). Soluble fluorine can be chemically adsorbed on the surface of tephra particles (Oskarsson 1980) and thus transported with the ash. Fluorine can also be transported by water due to leaching from tephra and gas particle disposition. Fluorine concentration has even been measured to rise 5 days after an eruption ended due to the melting of polluted snow and leaching of pollutants from volcanic ash into a river discharge area during a rainstorm (Gudmundsson et al. 1992). Fluoride can contaminate open drinking water sources. Outbreaks of fluoride toxicosis (fluorosis) in farm animals have repeatedly been observed in Icelandic sheep when pastures and drinking water have become contaminated with fluoride from volcanic eruptions (Kristinnsson et al. 1997), for example, from Hekla and Laki eruptions (Gudmundsson et al. 2008), and EFJV (Larsen 1999).

According to Wilson et al. (2011a), the impacts and consequences of remobilized ash were seldom examined prior to the EFJV eruption. However, the seriousness of remobilized ash was already identified following an eruption of Vulcan Hudson in Patagonia in 1991 (Scasso et al. 1994). Remobilization is related to wind erosion. Problems associated with sand and dust production from volcanic eruptions, wind erosion and loss of ecosystems in Iceland are well known (Arnalds et al. 2001). Iceland has about 22,000 km2 of sandy deserts that are a major source of atmospheric dust. Icelandic dust is mostly basaltic volcanic glass, which is rather unique for global dust sources (Arnalds et al. 2001). The sandy areas have black surfaces due to their basaltic origin. The sand originates largely from glacial margins, glacio-fluvial deposits and volcanic eruptions, but also from sedimentary rocks (Arnalds 2010). Wind erosion has been extensively studied in agricultural fields, but knowledge of field conditions and wind erosion rates of fresh volcanic deposits under severe wind conditions is limited (Arnalds et al. 2013).

3.3.2 Actual damages, injuries and disruption

The main and most affected areas, according to the recon, are depicted in Fig. 4a, corresponding to the ash plume in Fig. 4b. There were less than 20 farms in the main affected areas, and less than 10 of these were in the most affected area (see Fig. 4a). The issues raised by farmers during site visits are divided into four categories (see Table 6): coping strategies, constraints to the coping strategies, complaints about the response and concerns over issues that they did not know how to deal with. The main factors amongst farmers, according to the recon, were questions related to the natural assets (air quality, the impact of chemicals on pastures and fodder, etc.), physical (housing of animals, etc.), farming processes (lambing, fencing, temporary pastures, etc.) and financial concerns. Financial worries included the overall impact of the damages and disruption on farmers’ ability to pay debts. For example, lambs are normally slaughtered in the autumn, but if the situation called for early slaughtering, lambs would be smaller than at the normal slaughtering time, and farmers were concerned about being compensated for such losses. Farmers’ Insurance Fund coverage was also a concern at the time of the recon. The Fund determines compensation on a case-by-case basis (i.e., per eruption, earthquake, snowstorm, etc.), depending on the typical types of damage each time. At the time of the recon, farmers were still uncertain about which types of damages would be compensated.

Table 6 Farmers’ viewpoint collected during recon
Full size table

3.4 Actor identification

The Farmers’ Recon Report mentions 11 institutions. Each of these was by default an actor relevant to the study, since they were either directly involved in the recon or referred to as an actor of relevance in the report. Each actor is characterized in Table 7 by name, type (public or mixed public and private) and level (executive or/and civil organization, both national and local). The Farmers’ Association, The South-Agri and other regional farmers’ associations, and the Ministry of Agriculture organized the recon. Representatives from the Food and Veterinarians Authorities, the Agricultural University of Iceland, police authorities, and the Rangarthing Eystra Municipality participated in the data analysis and development of recommendations. The Land Conservation Authorities, the Farmers’ Insurance Fund, the Icelandic Catastrophe Insurance and the Directorate of Labour are mentioned in the report.

Table 7 Actor characterization
Full size table

3.5 Actor daily activities and categorization

An actor’s daily activity is inferred from their general objectives described on the website (see Table 7). The relevance of the actors to the study is based on the relevance of their daily objectives to a farming livelihood perspective and falls into one of the two following categories:

  1. 1.

    Direct farming actors: those working directly for, or with, farmers for the sake of livestock or production matters, at any level.

  2. 2.

    Associated actors: those who deal with farmers for reasons other than farming, influence farmers’ lives and livelihoods, but are not involved in farming-related activities.

Seven of the 11 actors were direct farming actors (No. 1–7, Table 7). Five were on a public executive level, dealing with matters of government strategy, implementation and monitoring, government insurance companies and educational establishments. The other two were a national and a regional farmers’ civil organization (Farmers’ Association and South-Agri), with executive roles, and were the main organisers of the recon. The Ministry of Agriculture is the highest ranking governmental farming actor. The ministry interacts with the national civil organization, which interacts with the local civil organization. Four of the 11 actors were associated actors (Nos. 8–11, Table 7). The Iceland Catastrophe Insurance provides insurance against natural disasters for buildings that have fire insurance; the Directorate of Labour deals with issues of unemployment and paid for temporary labourers to farmers and farming actors; the police deal with law enforcement and take charge in matters of Civil Protection; and the municipality manages community services and affairs.

3.6 Actor disaster-related activities

3.6.1 Interview summaries

A narrative summary of the answers to the interviews is presented in Table 8. Not all actors are necessarily associated with all disaster-related objectives, e.g., the insurance companies in this study do not address life safety, resulting in no action towards objective #5.

Table 8 Actors’ activities to achieve objectives
Full size table

For overview purposes, Table 9 shows the results in Table 8 coded as: Y (action taken), 0 (no action taken) and L (limited action). Limited action usually referred to addressing DFM objectives similar to their daily objectives, but not specifically with respect to EFJV. One exception was the municipality, which addressed objectives with respect to EFJV, but from a police perspective, not their own.

Table 9 Actors’ activities to achieve objectives coded
Full size table

3.6.2 Activities in the context of Civil Protection

For the purpose of placing farming actors’ activities in context with the overall response, the activities of the local chief of police are discussed here and the broader aspect of Civil Protection. Disaster-related activities in Iceland are by law coordinated by chiefs of police at the local level and by the National Commissioner of the Icelandic Police at the national level, managed by a Civil Protection and Emergency Management Department (Act on Civil Protection, Lög um almannavarnir 2008). The national level also responds locally, for example, by running a temporary service centre, and providing guidance to residents regarding ash and other volcanic hazards through pamphlets and the Internet (www.almannavarnir.is).

Shortly after the 1999 earthquake swarms in EFJV, scientists notified the National Civil Defence of Iceland (now Civil Protection Department of the National Commissioner of the Icelandic Police since 2003) of these events, which initiated a discussion of a possible eruption. At the time, a significant amount of research had been published on eruptions of Katla, the neighbouring volcano and subsequent flooding, as it was seen as a risk due to past activity (see Fig. 6), but less on the EFJV.

The interview with the local chief of police revealed the following (see Tables 8 and 9). The local police do not perform natural disaster risk analyses, but they sent a letter of request to the Minister of Justice (now Interior) in 2002 to fund a scientific risk analysis for the Katla and Eyjafjallajökull volcanoes. The analysis was published in 2005 (Gudmundsson et al. 2005). It was a flood hazard analysis only; ash hazard was not included. A digitized simulation of floods on topographical maps was produced for the area; however, exposure, vulnerability or risk studies were never completed. The local police did not perform or coordinate any direct mitigation activities, but did encourage others to do so. The 2005 hazard analysis was used to update existing evacuation plans for Katla to include EFJV. The police/Civil Protection coordinated a full-scale evacuation and communication exercise in 2006 for emergency services and citizens (Department of Civil Protection and Emergency Management 2006). In the weeks prior to the eruption, the local chief of police held approximately 10 town hall meetings, including one with presentations from scientists, rescue teams and the National Commissioner of the Icelandic Police. The main focus of learning from the police perspective after the event has been on improving technical communications.

In the early morning of 14 April, some 800 residents were evacuated due to the summit eruption. Most of them were allowed to go back the same day and the rest the day after (Gudmundsson et al. 2010) for safety reasons, depending on how close they lived to the rivers that were in danger of flooding. After the evacuation, the farmers wanted to go back to their farms to tend to their animals. The police allowed temporary visits during the evacuation period after a mobile phone-monitoring system had been set up to call farmers out of the area if need be.

A temporary service centre was set up in the lowland farming area at the south-western corner of EFJV (Heimaland), where those affected could talk to representatives of various services, such as insurance companies, building authorities, medical and farmers’ extension services (Thorkelsson 2012). The manager was a policeman and represented the National Commissioner of the Icelandic Police. His role was to provide advice and information to the affected population (citizens and municipal staff), coordinate projects with the municipalities in recovery and liaise with ministries and agencies, for projects related to the eruption (Department of Civil Protection and Emergency Management 2010). The manager also had considerable knowledge of agriculture, which motivated The Farmer’s Association to work with him (Bjarnason 2014). A staff member of the National Police Commissioner also visited farms during the eruption to give general advice to farmers.

3.7 Farmers’ livelihood outcome in 2010

The farming livelihood outcome for 2010 is described, based on the evaluation of farmers’ situation by the recon team (see Table 10). The organizers identified various unanswered questions, issues of concern and priority actions relating to chemical analysis and monitoring, animal sheds, hay and fodder, labour support, rest periods for farmers, animal transport, pastures, soil, crops and the use of fertilizer for farmland rehabilitation. Damage to buildings was not an issue because the Iceland Catastrophe Insurance insured building damage, and fire departments, rescue teams and volunteers, organized by non-farming actors, provided services to wash buildings.

Table 10 Recon teams’ viewpoint after field visits
Full size table

4 Analysis

The farming-EFJV perspective (presented in Sect. 3) was used as a basis to answer a question relevant to each disaster-related objective (in italics below). The answers are presented in bullet form for each objective and summarized in Table 11 as missing or unplanned activities.

Table 11 Missing or unplanned activities per objective
Full size table

For the sake of brevity, only the activities of the direct farming actors (first seven in Table 7) are included, and the activities for all four operational objectives (4, 5, 6 and 7) are viewed together.

4.1 Risk analysis activities (Table 2, Obj. 1)

Would it have been possible to perform a useful risk analysis from a farming perspective prior to the eruption? A risk analysis is based on exposure, hazard and vulnerability analyses. The area exposed to an eruption in EFJV is a small farming community. Information about farming livelihood strategies was available, and numerous actors were involved in the farming processes (Sect. 3.1). Therefore, detailed information about exposure in terms of assets and processes could have easily been collected. Hazard research prior to the eruption indicated that Iceland is an active volcanic island, that an explosive and/or effusive eruption, a summit and/or flank eruption, intermittent eruption, harmful ash levels, lightning, CO2, lava flow, tephra fall, fluorine, sand and dust from volcanic ash, floods with ice melt and waterborne ash could have been expected from an eruption in Iceland, and that Eyjafjallajökull was an active volcano (Table 4). Ash levels from EFJV had not been estimated in the hazard analysis of EFV in 2005, but a 20-cm ash cover (Fig. 6) was expected in the exposed area, given eruptions in other nearby volcanoes. Degassing of magma and subsequent chemical transfer into the atmosphere, onto the ground and into rivers, were known threats. When EFJV erupted in 2010, it did indeed involve effusive and explosive eruptions, flank and summit eruptions, magma discharge of lava flow, water-transported ash and airborne ash, plume, ice melt, glacial river flooding and lightning, and the ash level in the farming areas was measured well below the 20 cm mark (Fig. 7). Therefore, the EFJV 2010 was well within the range and type of an expected volcanic eruption in the region. In regard to vulnerability, research had presented numerous possible effects on natural and physical assets (Table 5). A comparison of Tables 5 and 6 shows that various issues for farmers, such as ash covering pastures, fodder, water, visibility, animal fluoride tolerance, soil conditions and general interruption of agricultural activities (Table 6), had been discussed in previous research (Table 5). Farming in Iceland has changed significantly through the centuries, for example, through technical advances in harvesting and milking, increased societal knowledge and increased education amongst farmers (Jóhannesson and Agnarsson 2005). Changes in production lead to changes in vulnerability; for example, increased technology may decrease (e.g., produce fodder faster) or increase disaster vulnerability (e.g., damage to machinery), calling for periodically updated disaster risk analysis from a farming perspective.

  • Yes. The actual impact from Eyjafjallajökull was within the range of a hazard analysis from a volcanic eruption in the region, so a risk analysis would have been useful.

4.2 Disaster Risk Mitigation activities (Table 2, Obj. 2)

Could a mitigation analysis prior to the eruption have led to ideas on how to reduce disaster risk? A sound risk analysis would have informed a detailed mitigation analysis and led to realistic ideas for mitigating measures taken by farmers, their families and farming actors in regard to damage, injury and disruption. One example is given here on increased visibility and traffic safety during ash storms. The Land Conservation Authorities systematically cultivates vegetation along roads to limit ash, sand and dust remobilization (e.g., in the Hekla region and the sands along the south coast) to increase visibility whilst driving in sandstorms and reduce the risk of scratching car paint (see daily objectives in Table 7). A stretch along the main road in the EFJV area had been cultivated by Land Conservation prior to the eruption for these reasons, which were reasons unrelated to the risk of an eruption in EFJV. During ash storms from the eruption, Land Conservation staff members saw that ash density blowing over the roads along the cultivated land was less than in other places, facilitating traffic and increasing road safety because it was easier to see the road. If the agency had been aware of the risk, they could have put efforts into cultivating more land next to the main road in the EFJV region to reduce ash mobilization over main roads.

  • Yes. A systematic mitigation analysis could have led to increased visibility and road safety during ash storms and knowledge of vulnerability processes would have been the basis for further analysis.

4.3 Operational Preparedness activities (Table 2, Obj. 3)

Could a useful contingency plan have been written by and for the farming community to guide impact, rescue, relief and/or recovery operations? When prompted by a realistic risk scenario, farming actors, with their knowledge of farming processes, such as lambing, sheltering of animals, the need for feed, etc., would likely have been able to identify content for a contingency plan to guide the response of the farming actors. Examples include: identifying responsible parties for paying for fencing and fertilizing, for temporary pastures, for the transport of animals and for people organizing chemical analysis and reporting; developing standard operating procedures to ensure that information to farmers is appropriate in content, format, timing and delivery; and that farmers are provided with rest periods. The actors may have identified the need for further research on animals’ fluoride tolerance, health effects, the effect of ash on soil and vegetation of fields and pastures, on factors to consider in deciding whether to build additional shelter for animals or to evacuate them, and, if they are evacuated, who will tend to them in the new location—for example, milking cows. Preparing a contingency plan could have led farming actors towards training in establishing coordination mechanisms and needs assessment.

  • Yes, useful contingency plans could have been written by and for the farming community prior to the eruption, if based on a sound risk scenario and accompanied by training in disaster operations.

4.4 Operational activities (Table 2, Obj. 4, 5, 6 and 7)

Could coordination and assessment efforts amongst farming actors have been performed better? Impact, rescue, relief and recovery operations require effective coordination and timely needs assessments. Establishing such processes at the onset of a disaster is a standard operating procedure (UNOCHA 2006). Coordination was initiated by the national government when it formed a committee of the secretaries general of the ministries the day after the eruption. Subsequently, the Secretary General of the Ministry of Agriculture formed a coordination group of farming actors that included the Farmers’ Association. The Farmers’ Association coordinated with South-Agri. The South-Agri offices, situated in the town of Selfoss between the capital, Reykjavik, and the affected area (See Fig. 4), became the centre of coordination for field activities. South-Agri field staff in the affected area also reacted to the situation by helping farmers move animals into shelters. Thus, the farming community created its own disaster coordination mechanism under the Ministry of Agriculture, based on their normal collaborative network. This mechanism was linked to the coordination mechanism of the police/Civil Protection, for example, through cooperation between the Ministry of Agriculture and South-Agri with the manager of the Temporary Service Centre. The Food and Veterinary Authorities also participated in developing an ad hoc coordination mechanism with various other actors. The initial reaction to the eruption was more or less immediate (according to interviews) and mainly in accordance with their daily objectives (see Table 7). However, in regard to assessments, despite a general understanding of the situation, a joint needs assessment was not performed until four weeks into the operation when the situation in the field had become problematic and farming actors were beginning to show fatigue. The eventual needs assessment systematically and collectively gathered and analysed information to allocate resources and assist the affected farming and animal populations and agricultural facilities in a concerted effort.

  • Yes. Whilst initial efforts were immediate, the coordination for farming actors was ad hoc, and a systematic needs assessment was performed four weeks into the event, prompted by awareness of the problematic situation. This is reactive management. Proactive management could have led to a better performance, e.g., an earlier recon mission.

4.5 Systematic learning activities (Table 2, Obj. 8)

Would systematic learning from farming actor activities associated with the 2010 eruption be useful? Lack of farming-disaster risk analysis (4.1), systematic mitigation analysis (4.2) and disaster training and planning (4.3) led to ad hoc, reactive and delayed measures (4.4). These facts provide reasons for farming actors to separately and jointly review their pre-eruption and real-time EFJV-related activities to learn lessons in case of future and possibly larger eruptions. Lack of pre-disaster planning leads to more time spent on operational planning after the onset of the disaster, causing delays in assistance reaching the affected population. Indeed, the lack of plans resulted in discussions amongst farming actors on what to do, on clarifying actors’ responsibilities, such as determining responsibility for organizing field sampling and reporting schemes for chemical monitoring, determining logistical support for animal evacuation, estimating operational cost and deciding who shall pay for what, determining coordination mechanisms and resolving other issues normally addressed in contingency plans to ensure effective and efficient response operations, taking place after the eruption started. Learning processes should also address gaining clarity on the roles and responsibilities of ministries before, during and after disasters. The Ministry of Agriculture sought clarification on their role from those legally responsible for disaster coordination, i.e., the Minister of the Interior and National Commissioner of the Icelandic Police, but no clarification was received. At the time of the interview, it was still unclear who bore the responsibility for initiating a systematic learning process.

  • Yes. A systematic learning process could reveal the current knowledge of farming actors in regards to risk analysis, mitigation and preparedness, could increase pre-planning to shorten operational planning and hasten field response and could clarify roles and responsibilities in the farming community.

5 Results

The reason for inadequate information getting to farmers during the 2010 includes lack of pre-eruption activities by farming actors regarding an eruption in EFJV. Four of the seven farming actors performed no risk analysis at all, whilst three performed general activities related to ash risks. Five farming actors took no mitigation measures, and two participated in mitigation-related activities for general volcanic threats. No farming actor initiated any activity towards designing a coordination mechanism or writing a contingency plan for farming actors as a group. Risk analysis, mitigation analysis and contingency planning could have prepared the actors to collect, compile and timely share relevant information.

Inadequate information for farmers can also be linked to the delay in joint needs assessment. The reason for the delay can be attributed to lack of training in disaster management, leaving farming actors to gradually realize problems amongst farmers as the problems grew bigger and therefore manage reactively.

The lack of pre-event activities is not due to insufficient pre-eruption information. On the contrary, substantial information was available to perform a realistic risk analysis for EFJV from a farming perspective. The hazard was well known, the exposure area was small with well-understood assets and processes, and relevant vulnerability studies existed that could have been used as a basis of further vulnerability analysis. The resulting risk scenario would have resulted in a higher intensity scenario than actually occurred; therefore, any planning for the risk scenario would have been within the scope of the actual event.

The study shows a lack of post-event learning. Four farming actors had not initiated a process to learn from their experiences (one actor believed its institute did not need to do so). Two have initiated, but not completed, a learning process. One has changed response procedures, based on the experience, but none has initiated a learning process for the farming community in general. The experience did give some actors insight into what could have been improved. For example, during an interview one actor stated that, in retrospect, the recon should have been performed sooner and more often, but this has not been enough to initiate a learning process within the learning community.

6 Discussion on perspective

The impact of the EFJV eruption was not a disaster by the UN’s definition of disaster (UNISDR 2009), as it did not cause disruption at the community level to the extent that it exceeded society’s ability to function. The serious disruption was on the farm level. There were less than 20 farms in the main affected areas, and fewer than 10 of these were in the most affected area (see Fig. 4a). There was general agreement amongst the interviewees that the coping level of actors and farmers was due to fortunate circumstances: a small, homogeneous, sparsely populated affected area, where the farmers and institutional farming staff knew each other very well, and unfavourable winds were of short duration. Nevertheless, numerous groups were involved, demonstrating the diversity of perspective in the disaster-related activities. For example, scientists, police and farming actors have different roles and responsibilities and demonstrated different perspectives on what constitutes disaster-related activities.

Scientists had monitored the EFJV volcano since 1991, reported the activity to the Civil Protection after it increased in 1999, presented a detailed analysis of flood-related hazards due to EFJV and Katla Volcano in 2005, where ash was not considered, published a hazard assessment in 2008 of a 20 cm ashfall around EFJV and increased their monitoring of EFJV as the activity increased in the months prior to the eruption. A scientific field assessment includes a detailed assessment of ashfall, e.g., analysing the thickness of ash versus the intensity or degree of impact, the frequency, duration and magnitude of the ashfall event, along with the characteristics of ash, such as grain size, mineralogy and content of soluble acidic salts associated with the ash (Wilson et al. 2011a, b). The results of such activities include information like that presented in Fig. 7. No risk analysis was performed, other than mapping of a flood zone and locating constructed elements in the flood zone.

The police, after being informed of the hazard by the scientific community, initiated funding for further hazard analysis relevant to life safety from flooding due to eruption under icecaps, and organized evacuation exercises in the event of flooding. Icelandic law makes pre- and real-time disaster coordination the responsibility of the police. They were actively engaged in coordinating various actors, mainly emergency management agencies, for eight years prior to the eruption. They focused on life-saving activities. In March 2013, three years after the eruption, the local and national police and the local Civil Protection Committee published the first formal response plan specifically for an eruption in the EFJV (Police Chief Hvolsvelli et al. 2013). Its content focuses on population control, evacuation, mass care, traffic control, rescue and flood control. The only farming actor is the National Veterinarian, who works within the Food and Veterinarian authorities. His role during an eruption, according to the plan, is to “work according to his own plan”. The new plan clearly shows that the police perspective does not include the perspective of farming actors.

Despite a history of volcanic impact on farms in Iceland, the farming actors were not informed of the increasing risk. They did not initiate any type of pre-disaster activities and have not (at the time of the interviews) engaged in systematic learning as a community. The farming actors are not considered part of the Civil Protection umbrella since they were not invited to participate in preparedness activities with emergency agencies and are not part of the latest police response plan. However, farming actors were highly active during the response and are important response actors for farmers. The Ministry of Agriculture, the Farmers’ Association and South-Agri participated to some degree in activities regarding all four operational objectives, where the life-saving objective (Objective 5 in Table 2) relates here to the lives of the livestock. Failure to identify relevant stakeholders (actors and those affected) has been documented as one of the possible reasons for systems failure (Lyytinen 1988, Pouloudi and Whitely 1996). The question becomes who is responsible for identifying and activating stakeholders to avoid systems failure.

The point argued here is that the level of specialization amongst livelihood professionals in general requires them do their own risk and mitigation analyses and to prepare their own disaster-related activities and coordination mechanisms. The consequences of ashfall on farming are not well understood by non-specialists. Brunsdon and Park (2009) reach a similar conclusion from the perspective of designing and managing infrastructure when they suggest that consequences to a community due to infrastructure failure are not well appreciated by other utility providers, civil defence agencies, businesses and communities.

7 Conclusions

The reason for inadequate information to farmers can be attributed to lack of pre-eruption activities amongst farming actors, stemming from lack of clarity on who bears the responsibility for identifying actors to assist farmers from a livelihood perspective and initiating their pre-disaster activities.

Livelihood actors provide specialized knowledge on their specific fields and in times of disaster can offer specialized solutions for their industry that are beyond the perspective of emergency services. However, being highly specialized in their field is not enough when it comes to responding effectively and efficiently to disasters. First, specialists need to be updated on any developments regarding natural hazards. Second, they need to be trained in disaster-related activities and standard operating procedures, such as developing disaster risk scenarios from a livelihood perspective, building coordination mechanisms, initiating timely needs assessment and sharing specialized information. The failure to identify farming actors prior to 2010 resulted in lost opportunities to analyse farming-disaster risk and mitigation options, organize coordination mechanisms for farmers and write contingency plans. Livelihood actors should take the initiative in analysing disaster-related legal frameworks from their own perspective to identify and address gaps, such as the provision of information important in their field.

This study provides numerous ideas for further research. For example, additional questions can be asked on the information in Sect. 3. For instance, how to develop a risk-planning scenario from a farming perspective, and what recon questions provide the most useful information. However, the main topic of future research is seeking clarity on the roles within legal frameworks for disaster-related activities, in order to determine necessary changes to such frameworks, to ensure effective and efficient pre-event, real-time and post-event activities amongst livelihood actors.