Modeling Spontaneous Volunteers in Cross-Border Crisis Scenarios

2.1 Spontaneous Volunteers’ Fundamentals

SV have received considerable attention in research and practice in recent years, although the phenomenon of people gathering and working together to help in the aftermath of disasters is not a new one. As early as 1957, Fritz and Mathewson (1957) described how volunteers without affiliation to relief organizations travel to disaster sites to help. Included in their description were also first aid activities after accidents and local neighborhood help as activities of volunteers. The focus here is on groups that emerge after disasters, as described by Dynes (1970) as informal groups with non-regular tasks. Stallings and Quarantelli (1985) further identified five basic factors that favor the emergence and functioning of such communities: (1) the existence of an acute or imminent danger; (2) the perception of this danger by supra-regional organizations and the press, and the consequent attention to those helping; (3) a set of social norms and values shared by the group, which enables a consensus on the goal to achieve; (4) the establishment of new social relationships, including relationships with other organizations, which gives the group a certain legitimacy; and (5) the existence of sufficient resources to face the danger, such as information and skills but also material things (Stallings and Quarantelli 1985). Independent of group motivation, an individual’s willingness to help often results from a personal relationship or identification with the victims (Fritz and Mathewson 1957, 47), in addition to a lack of authority in the initial phase. Another reason for this supporting behavior is the generally urgent need to help and to make a positive contribution in such a hopeless situation (Lowe and Fothergill 2003).

In recent years, special attention has been paid to SV in connection with social media, which has made the phenomenon even more visible after major disasters. In Europe, for example, volunteers used social media to organize themselves and support emergency services during the 2013 floods in the Elbe. The tasks that SV can perform range from the provision and supply of professional emergency forces (e.g. food) to goods and blood donations, and simple tasks, such as filling sandbags or even providing special machinery of local construction companies (Twigg and Mosel 2017). Research on the assignment of SV tasks after an event and SV coordination was undertaken by the research project INKA of the German Red Cross. GIS-supported coordination was addressed by the KUBAS research project (Sackmann et al. 2018; Schorr et al. 2014).

Other recent studies on SV discuss the improvement of the operational involvement and integration of SV in the structures of civil protection and preparation, as well as the response phase. For example, in a particular research project, SV activities were specified and individual tasks with corresponding work requirement profiles were formulated in the form of an activity catalog (Drews et al. 2019). In consequence, civil protection units were able to recognize the potential of the helpers with the idea of using this knowledge in upcoming situations (Drews et al. 2019). The research project Wukas, which dealt with the topic of occupational health and safety for SV, was also built on this framework. In said project, easy-to-understand pocket cards were developed for instructing SV in occupational health and safety measures. Moreover, requirements for personal protective equipment were defined, and these could be carried by emergency response units to equip potential SV (Bier et al. 2022; BUW and Malteser 2021).

During the flood disaster in Central Europe in 2021, studies on SV recorded how the abovementioned concepts were used (Drews et al. 2021) while surveys of SV recorded their motivation to help, the duration of their commitment and the distance between their place of residence and the place of volunteering (Bier et al. 2023).

Further research can be found on technology-supported SV integration. The focus of the research project “KatHelfer” is on volunteers’ cultural backgrounds, and the idea is to integrate volunteers via local community structures by using a smartphone application (Posselt 2023). A special topic in the field of SV with regard to technology is the unique forms of engagement, such as digital volunteers – i.e. people who take on coordinating activities in social media to manage SV, or other tasks such as mapping a disaster area. Institutionalized forms of volunteering in this area are Virtual Operations Support Teams (VOST), which draw on a pool of volunteers. VOSTs are affiliated with disaster relief organizations to contribute by analyzing social media content and they can, therefore, engage with SV (Fathi et al. 2020).

A broader overview on the different forms of civic engagement is provided by the ATLAS Engage research project, which records and categorizes the various forms of engagement. This is future-oriented research aiming to determine how social transformation will alter civic engagement in coming years. The question here is how voluntary engagement in society can be promoted in the future and how different forms of volunteering can be integrated in civil protection (Katastrophenforschungsstelle 2021). Despite the wide range of studies in the area of spontaneous volunteering, the specific context of spontaneous helpers in border regions is not sufficiently addressed in research.

2.2 SV in a Cross-Border Context

As stated in the introduction, special problems arise in the assignment of SV to border regions. These include that different countries feature different systems of volunteer assistance. However, a common understanding of the meaning of SV in different countries is fundamental to the analysis. Thus, we follow a definition of SV that can be used across borders and includes the characteristics outlined in Section 2.1, such as being unaffiliated and arising spontaneously. A suitable framework covering these aspects can be found in International Organization for Standardization (2017), which defines SV as an “individual who is not affiliated with an existing incident response organization or voluntary organization but who, without extensive preplanning, offers support to the response to, and recovery from, an incident” (ISO 22319:2017-04, p. 1). This definition provides a framework for a mutual understanding.

To outline the relevance as well as to understand the specifics of SV in border regions, a brief look at some studies is presented here.

The study by Bier et al. (2023) takes the flood in the Ahr valley in Germany as an example to highlight the challenges of volunteers in a non cross-border context and outlines their relevance and expected application to borderland scenarios. First, the motivation to help is analyzed, but also the travel behavior of the helpers, as well as the integration into the structures of disaster management and the use of communication infrastructures by the volunteers.

In particular, the article of Bier et al. (2023) refers to the 2021 flood in Europe and focuses on the Ahr valley in Germany. Here, according to the German Federal Ministry of the Interior and Community (BMI 2022), up to 100,000 volunteers were on site. The article picks up on some exciting aspects that may also apply to volunteers in border regions: SV will travel to disaster areas in large numbers during large-scale disasters. It can be assumed that this is not only the case for regionally limited disasters, but also for cross-border disasters. They are motivated to do something good and identify themselves with those affected. To do so, they are also willing to travel great distances, i.e. it is not only the local/regional affected population that offers its help. This is supported by an empirical study of Klein, Wiens, and Schultmann (2022) investigating the social cohesion and the willingness to help each other especially in border regions. The findings of a representative study show that 63.6 % of respondents in the French-German border region are willing to help also in their neighboring country. The motivation of volunteers to cross national borders in order to provide on-site help is also addressed as well as a question on perceived problems. Here, the majority of respondents states the unfamiliarity with crisis response procedures of the neighboring country as main hindering factor. This is also pointed out by Bier et al. (2023) highlighting that the status quo in the involvement of volunteers in disaster management structures remains low, despite years of research and knowledge of the situation among practitioners (in Germany). In the Ahr Valley there has been (almost) no involvement of disaster management professionals and volunteers and the SV have been left to their own efforts and organize themselves accordingly. The desire for integration or contact with the disaster control organizations is definitely present. There is also a potential recognized and demanded by the volunteers themselves to make their help more effective. The involvement of SV not only has advantages in terms of resource allocation, but can also avert hazards for the volunteers. This point is particularly relevant for cross-border scenarios as the SV may not know the emergency numbers in the other country or their insurance may not apply. Disaster areas pose risks that require instruction and equipment. However, as the equipment can be different in other countries, the correct use might not be known to volunteers. In addition, volunteers (as well as) professional forces are affected by psychological stress. Much of the coordination and communication among SV takes place via the Internet and social media, e.g. via Facebook and instant messaging services. Bier et al. (2023) conclude that in addition to individual aspects, these two things are essential to meaningfully integrate the help of SV: Coordination and communication. It must be ensured in advance that the authorities have the competencies and structures to actively involve helpers (across borders). In particular, it is to note that authorities need to be able to deal with different languages to address helpers also from other countries.

These aspects are also of particular importance for border regions. Ansell, Boin and Keller (2010) describe three difficulties in cross-border events. The disaster has an unknown origin (in another country), the course of the disaster is difficult to assess, and the measures taken to deal with it may initially be unknown to the partner organizations on the other side of the border. These problems can make coordination and communication difficult, even though border regions in Europe have already grown very close together. For example, Dahles and van Hees (2007) describe everyday problems in cooperation in civil protection, such as the lack of interoperability of material or in data exchange. Civil protection responsibilities and planning often relate only to individual countries, and cross-border risks are not considered, nor is the planning of SV operations. In the worst case, national borders become visible again, as happened during the pandemic (see next section). The close relationships in civil protection suffered from the border closures and the exchange of relief goods/material was suddenly strictly controlled again. Citizens were also no longer able to cross borders easily everywhere; negative COVID-19 test certificates had to be provided. On the other hand, the pandemic raised a huge wave on local actions of mutual support and brought up innovative solutions to overcome the difficulties.

At this point it can be concluded that SV will also appear in border regions and offer their help. Due to different or non-existing concepts for the integration of volunteers in this areas, special difficulties may arise when volunteers seek a connection to the civil protection organizations. In border regions, due to the involvement of two or more countries, a larger number of actors are active in civil protection and need to coordinate in their response.

Communication with volunteers is key for their integration into disaster response. However, the demand in border regions is even higher. Joint coordinated communication with volunteers regarding points of contact and tasks need to be pre-planned, taking into account linguistic difficulties. Residents close to the border may be able communicate in the border region, but not necessarily SV coming from further away. Disaster response organizations need to think about in advance how to engage with volunteers via online social media groups. This is backed up by Klein, Wiens, and Schultmann (2022). The article deals with the special needs and reservations of SV in border regions. Potential volunteers were asked about obstacles to their cross-border engagement, and the interviewees primarily cite language and the unfamiliarity of disaster management organizations in border regions as reasons for not becoming active as SV.

For disaster management authorities there is a need to think about in advance how to organizationally involve SV across borders. In the area of information dissemination, online groups in the social media play an important role where authorities should play an active role. There is a need to coordinate these structures across borders.

This article attempts to quantify the different approaches to coordination of SV in border regions (from none to a fully integrated cross-border coordination approach) via an agent-based model, as empirical data is lacking. The importance of coordination in cross-border settings is examined, taking into account particular aspects such as infrastructural conditions and external influences such as border closures in the event of a pandemic.

Summarizing, both studies support the argument that volunteers will be crossing borders to offer their help. However, the authors of this paper did not find sufficient articles on the cross-border utilization of SV in Europe, which is why the approach of modeling their engagement with an agent-based simulation was chosen for this article (see Section 3). Therefore, an existing coordination framework for SV is extended to the specific questions on the effects of missed help requests from citizens as well as decreasing volunteers’ The following Section 2.3 examines the COVID-19 pandemic as relevant cross-border scenario.

2.3 Relevance of the COVID-19 Pandemic to Spontaneous Volunteering

The COVID-19 pandemic is an example of a cross-border crisis. Weber and Wille (2020) outline a chronology of events related to border regions in Europe. On March 15, 2020, the first border in the EU (Poland-Germany) with a neighboring country was closed, the movement of people was restricted, and border controls were introduced. The following day, Germany decided to establish temporary border controls for Austria, Switzerland, France, Luxembourg, and Denmark. In the following days, these decisions were implemented, and restrictions on land and air travel, as well as border crossing points, were specified. These measures were aimed at reducing the movement of people, thus preventing the spread of the virus. With these measures, national borders regained significance in the EU, which, up until then, were hardly or not at all visible or noticeable in many places because of the Schengen Agreement. Weber and Wille (2020) state that these were spatial measures on the national borders of each country and, therefore, special conditions in border regions were not considered. The needs of citizens in border regions remained unconsidered in crisis management measures, since decisions were made by political bodies located far away from such border regions and were utterly focused on the own territory.

The closure of borders also had an impact on day-to-day emergency responses. For example, it led to the suspension of cross-border rescue services between Denmark and Germany (Peyrony, Rubio, and Viaggi 2021, 8–9). The EU reacted quickly (in mid-March 2020) and issued recommendations for measures that border crossing should be possible for border commuters as well as for the establishment of so-called green lanes, which would ensure the necessary movement of goods and enable the crossing of emergency management personnel. Similarly, it was communicated that health service personnel and protective material should be allowed to cross the border (European Commission 2020a, 2020b, 2020c). Over the following weeks, the challenges of border regions also reached the political decision-makers, e.g. the “Cross-Border Task Force Corona,” an association of political decision-makers from North Rhine-Westphalia and the Netherlands, which met on a regular basis, was established. Furthermore, an Interreg project called Pandemric was launched in this border region, which is specifically dedicated to crisis management in the border region. One of the measures of the project was the compilation of the currently valid rules in the border triangle of Belgium, Germany, and the Netherlands for citizens and responsible authorities (2021).

Spontaneous volunteering has evolved at the individual level during the pandemic. Volunteers sewed masks for relatives or even complete strangers, or performed weekly grocery shopping for elderly people who no longer left the house because of the risk of infection. In many cases, these offers were coordinated via social media in the context of other SV activities. A particular research project has developed special SV guidelines for the pandemic, which outlined measures to protect volunteers against infection. Special activities such as shopping or walking dogs were examined from the point of view of occupational safety and health. Hazards were identified, and appropriate measures for protection were specified. Despite having been presented in a way that was understandable to individuals, these guidelines were issued to organizations that coordinated volunteer assistance (BUW and Malteser 2021).

3.1 Model Design

The pandemic is outlined in Section 2.3 as a borderland-relevant scenario in which SV may appear to provide their help. However, during the COVID-19 pandemic, borders were closed to prevent the virus from spreading. This intervention measure affected border regions in particular, as family members living in other countries were not able to support their vulnerable relatives, who became dependent on external helpers. This situation did not diminish the willingness to help, and SV were seen locally buying groceries for the elderly or vulnerable people. This study expands the framework by analyzing the pandemic as described in Section 2.3, with border closures in which volunteers can only provide local help due to measures such as travel restrictions and contact reduction. Therefore, this paper introduces an additional scenario – Scenario 0 – , in which volunteers are only allowed to help within the area in which they are located.

Moreover, two different types of well-known problems occurring with spontaneous help have been analyzed for border regions. On the one hand, requests from the population may be missed if coordination structures are not well established, and, on the other hand, SV may leave the disaster area if they are not assigned a task that proves that they are needed.

The simulation runs in discrete time steps, called “ticks”, in which time-dependent variables are updated. It is modeled along the structure of the Gaia methodology (Wooldridge, Jennings, and Kinny 2000) and implemented using the Repast Simphony toolkit (North et al. 2013).

3.2 Simulation Experiment

To run the simulation, we initialized 2000 citizens randomly in the border region so that 1000 citizens were located in each country. Their needs varied from 0 to 14. It reduces by 1 every 24 ticks, and the critical threshold for starting a request procedure is a value of 10. Initially, population needs were normally distributed, with a mean of 2 and standard deviation of 4. Per country, there are 5 groups on social media that can cover one area each. Additionally, the number of professionals is chosen for each country, so that they can deal with the five areas simultaneously.

To study the two additional coordination failures, we include the probability of failure p ∈ [0, 100 %] in the second analysis. If p is below a critical threshold, the requests from the population are missed. The probability of requests being missed grows when fewer coordination structures are established. In addition, we include a probability q ∈ [0, 100 %], which represents SV motivation. Motivation is high for all volunteers at the beginning but decreases over time when there is no task assigned to them. If this value too small (i.e. below a critical threshold), SV leave the disaster region and their work power is no longer available.

The Repast Simphony framework allows integrated parameter tracking over time. For the analysis of coordination effects compared between the different scenarios, we track the number of undersupplied citizens during the simulation duration (i.e. at tick 240).

3.3 Results

As presented in previous work (see Schulte et al. 2020), for each simulation run 1000 citizens were located in the region (500 citizens per country). The performance of volunteers’ coordination is measured by their effectiveness in the distribution of resources, i.e. by counting the resulting number of undersupplied citizens after the duration of the simulation. The following results show the mean number and variance of undersupplied citizens over three simulation runs. Figure 2 displays the four scenarios and shows that Scenario 1 without coordination has the worst performance (i.e. highest mean of undersupplied citizens) while Scenario 4 with common coordination in both dimensions features the best performance (i.e. lowest mean of undersupplied citizens).

Figure 2:

Mean undersupplied citizens scenarios 1–4.

We now compare these results with further specifications made in this study. We simulated 2000 citizens in the region (1000 per country). Furthermore, we take the mean number of undersupplied citizens over a number of 10 runs to obtain further insights into the mechanisms of the simulations.

Figure 3 features Scenario 0, the new scenario picturing a pandemic in which volunteers can only help in their local area. We see that this yields the highest number of undersupplied persons due to the least flexibility in the movement of helpers.

Figure 3:

Mean undersupplied citizens comparing scenario 0 to scenarios 1–4.

After comparing the results with the simulations with 1000 citizens, we observe that the number of undersupplied people increases by doubling the number of affected citizens while keeping the number of potential helpers constant. The highest increase can be observed in Scenario 1, where the complete lack of coordination outlines the importance of coordination, especially for scenarios with high numbers of disaster-affected citizens. This scenario also resulted in the highest variance among the different simulation runs.

If we compare the number of undersupplied citizens per scenario, we can calculate the performance for 1000 simulated citizens in the order of scenario 1, scenario 2, scenario 3, and scenario 4. For 2000 simulated citizens, the performance of scenario 3 is better than that of scenario 2. A detailed look into the simulation data of Schulte et al. (2020) reveals that the worse outcome of scenario 2 is the consequence of an outlier (high variance). As the means were only taken over three scenarios, the outlier simulation run of 25 undersupplied persons had a strong impact on the result (the other two simulation runs had a number of 8 and 9 undersupplied persons, respectively). However, taking the means over 10 runs uncovers a slightly different order of scenario performances: scenario 1, scenario 3, scenario 2, scenario 4, as outlined in the following figure. In this case, the outlier in scenario 4 does not play a significant role because of the higher number of simulations runs.

Figure 4 shows the slope of undersupply rates with regard to the five scenarios, which reaches its highest point between scenarios 1 and 2. Thus, it highlights that the highest potential for improvement lies between these scenarios. Since the difference between scenarios 1 and 2 lies in the establishment of coordination across the border, that should be the first aspect to be improved.

Figure 4:

Mean undersupplied citizens per scenario.

To conclude, we can highlight that, especially in large numbers of affected persons with a constant number of volunteers, the greatest potential lies in the optimization of coordination, but SV flexibility (i.e. no travel restrictions between areas) is needed to obtain an optimal overall result. However, the flexibility aspect is small – in a pandemic, it might be a better strategy to encourage people to help in their area to prevent large virus spreading. Coordinating help and allowing for movement to a neighboring area in case volunteers are unequally distributed are key in such a situation.

We will now focus on the analysis of the two additional coordination problems in this study. First, requests for help are missed in an uncoordinated scenario. Therefore, we analyzed scenario 5 and assumed a probability of p = 50 % of failing requests. This leads to roughly twice as many undersupplied requests of citizens as in scenario 5, without failing requests. As the resources are almost equally distributed in the considered setting, there is no substantial difference if requests from Country A or from Country B fail (compare the similar numbers of undersupplied citizens in Figure 5). If only half of the submitted requests come through, the number of undersupplied citizens is doubled, which is an indicator of a well-derived agent-based model.

Figure 5:

Mean undersupplied citizens with p = 50 %.

Finally, we have analyzed the second hindering factor of coordination, i.e. the decreasing number of SV due to decreased motivation when they are not assigning tasks in a timely manner. We have analyzed scenario 5 and assumed a probability q = 50 % of SV who waited for at least 10 ticks to get a task assigned before leaving the area. We observe that this factor of decreasing workforce clearly has a higher impact, resulting in an even higher number of undersupplied citizens. It should be noted that the decreased motivation of helpers in Country A had a lower impact than in Country B (see in Figure 6 the lower number of undersupplied citizens when the volunteers in Country A leave compared to leaving volunteers of Country B). This results from the fuzzy distribution of the areas in which SV help. Affected citizens who are located in the overlap of the two areas have the potential to request SV from both areas via social media. Thus, the effect of disappearing helpers was smaller for the country with a higher overlap of areas. A similar effect is seen in the previous analysis, i.e. if requests are missed due to discoordination (e.g. Figure 5).

Figure 6:

Mean undersupplied citizens with q = 50 %.