OR/15/032 Appendix 5 App methodology steps

Mee K, Duncan M J. 2015. Increasing resilience to natural hazards through crowd-sourcing in St. Vincent and the Grenadines. British Geological Survey Internal Report, OR/15/32.

Step 1: Identify key in-country partners and their needs
Citizen science is most effective when there are clearly defined aims or questions to be resolved (Pocock et al., 2014) and who better to determine these than the key in-country partners. This step is crucial in determining who key partners are, whether an App would be of use to them and the key issues it could resolve (and what it shouldn’t do). The monitoring institution and civil protection/emergency management are likely to be key partners and their roles, governance and disaster risk management framework should be well-understood.

Statements of intended collaboration are also likely to be established at this point, if they do not already exist. During this step, it is also essential to consider how the use and impact of the App will be evaluated, in order to determine baselines for comparison.

Step 2: Understand the natural hazard threat and disaster context
The second stage is to understand the potential natural hazards, their likelihood of occurrence and threat to populations — this assessment is the role of a mandated monitoring organisation (if there is one) but researchers also contribute to this understanding. The key factors to understand are:a) the threat to the country from different natural hazards (volcanic and non-volcanic) and their interaction, b) the previous history of natural disasters and lessons learned, and c) likely future response actions and location of critical infrastructure.

Understanding these factors helps to determine what hazards the country is most at risk from, where communities are most exposed and key aspects of vulnerability.

Step 3: Characterise mobile phone/internet use
Paramount to the success of any citizen science initiative is establishing that (a) a ‘crowd’ exists and (b) there is sufficient access to the necessary technology (see Polock et al.’s [2014] decision framework for citizen science). A monitoring institution and civil protection/emergency managers will already have a good understanding of the communities they serve and will know whether engagement in data collection is likely to be successful or not. Nationally however, it is also useful to establish telecommunications trends such as: how many people have mobile phones/smartphones; what is the dominant platform; how many people have access to the internet; do people use social media and if so, which is most popular? These questions can help to determine whether a user community exists and whether technological changes need to be implemented in the App e.g. develop on another platform. If there are sufficient social media users, additional services can be developed e.g. to trawl sites for relevant information.

Step 4: Recruit a core group of citizen scientists and wider networks The monitoring institution and authorities may know key individuals or groups in different communities, for example community leaders, local disaster groups, schools who would be interested in volunteering.

Motivating participants requires that we provide them with clear reasons as to why their participation is so important (Polock et al., 2014). At Tungurahua volcano, the motivations for volunteer observers (vigías) are an important component of the network’s success — all feel a sense of duty or moral obligation and want to help reduce risk to their family and community (Stone et al., 2014). Vigías stated that the voluntary nature of the role is very important to them.

The motivation for engagement may relate to different socio-economic and cultural factors and will require understanding the country context where the App is being promoted. However, we can take lessons from existing studies. It is important that those recruiting citizen scientists make the purpose of the App clear, emphasise the benefits to those involved (the information they gather will be shared and there will be opportunities to engage with scientists) and guidance on how to safely monitor unrest and other hazards in volcanic environments.

Step 5: Identify other ‘big data’ initiatives in the region There may be ongoing ‘big data’ initiatives or risk/resilience projects in the region that the App could link into. These activities may include existing citizen science initiatives, social media networks related to natural hazards, ground monitoring networks, early warning systems, open GIS platforms and earth observation projects. Since the App has the dual purposes of collecting and sharing information, it is critical to map out these existing resources to avoid duplication and to make connections where possible/necessary.

Step 6: Agree app specifications
A list of app specifications should be discussed and agreed between both the stakeholders (what is needed?) and the developers (what innovations are technically possible?). These might include developing the App on another platform, agreeing what extra data and information should be displayed on the map and establishing linkage to other ‘big data’ initiatives. It is also important to determine any anticipated challenges in developing the App to these agreed specifications.

Step 7: Develop and pilot The App is ‘customised’ following agreement of what extra specifications should be added. As citizen science works best when small scale trials are undertaken to test the approach (Tweddle et al., 2012), a pilot study should be devised in-country with stakeholders and identified users, in order to gauge their response and identify local or social relevance, as well as the interests and motivations of potential participants (see Tweddle et al., 2012). Validation should be considered and tested at this stage. Such reflection will help to inform awareness raising campaigns and whether any demonstrations/training are required.

Step 8: Implement, evaluate and review

Following implementation, evaluation could be through user and stakeholder questionnaires, evaluation of quantity and quality of data collected or by reviewing the impact of the App. Evaluation should be a regular and ongoing research process.