Smartphones for citizen science water quality monitoring in developing regions

Among many other efforts, high spatial and temporal resolution water quality monitoring data are required to help mitigating the ongoing global freshwater crisis. Citizen science is said to have a high potential to contribute valuable water quality monitoring data, while at the same time offering a range of qualitative benefits such as generation of a social fabric, environmental education, and improved relationships between citizenry and authorities. The potential of citizen science is elevated by integration with technology, especially smartphones, which allow for easy data capture and information sharing among a range of other powerful features.

As part of the CGIAR Initiative on Digital Innovation, we aimed to test some of the most prominent smartphone applications (apps) to investigate their scalability to developing regions for monitoring the Sustainable Development Goal (SDG) target 6.3.2 indicator water quality parameters or other key water quality metrics. We used southern Africa as a case study, since it characterises numerous key challenges to citizen science water quality monitoring using smartphones in developing regions. We evaluated five smartphone apps that are plug-and-play, assessing both their quantitative accuracy as well as their qualitative suitability to a southern African context. We found that the Hydrocolor and EyeOnWater apps showed theoretical promise but would not be useful for typical citizen science monitoring of streams, river, and dams from the banks of those water bodies given their requirement for deep water. The MQuant® StripScan App was not useful given that the reference cards required to use the app could not be sourced and that the app did not function to read the Mquant® nitrate test strips. The Nutrient App showed some promise but has ceased being supported, illustrating the critical importance of designing and developing tools with sustainable financing and maintenance in mind (as well as the need for funders to support key tools so they remain freely accessible) so that the great efforts that go into research and development are not ultimately wasted. The Aquality app was fairly user friendly, intuitive, and accessible for free via the Play Store and the Apple App Store. The development and support team were helpful and responsive, with ongoing research and development regarding the app showing good potential for upscaled functionality and implementation in the future. However, we found that there were significant qualitative and quantitative issues with the app that should be investigated further and addressed to ensure the app is suitable for global use, especially in the context of developing regions. These included that i) the material requirements for the app (i.e., a reference card mailed directly from Deltares in the Netherlands and Hach© nitrate test strips) were difficult to source in South Africa (which is likely to be the case in many other countries), ii) each test carried a significant financial cost (USD 2 – 5 per test), and iii) the guidelines of the app, especially concerning the appropriate lighting requirements, left considerable room for novice user error or lack of standardisation. Concerns over the standardisation and accuracy of the data were supported by the fact that we found little to no correlation between the estimated nitrate concentrations in surface water samples calculated by the app compared to accredited laboratory measurements of the same water samples.

Overall, we suggest that though our data were limited, they provide evidence that the data generated by real citizen scientists using such app around the world need to be carefully validated before they can be trusted. The reality is that doubt over standardisation and validity of results could prove a serious barrier to use of the data for management and policy interventions. More broadly, we identified a need for digital innovations within citizen science to remain human-centric and not become extractive, treating citizen scientists simply as data collection units. The massive potential for a nexus between digital innovation and citizen science will only be realised if we specifically cater towards the human part of the equation and if researchers and developers remain mindful of technicist assumptions about the utility, understandability, and validity of innovative technologies in different contexts.

Citation