Your search found 5 records
1 Pattinson, N. B.; Kuen, R.; Kuen, R. 2022. Artificial intelligence-based biomonitoring of water quality. Colombo, Sri Lanka: International Water Management Institute (IWMI). CGIAR Initiative on Digital Innovation. 32p.
(Location: IWMI HQ Call no: e-copy only Record No: H051644)
(2.97 MB)
The miniSASS was developed as a citizen science tool for monitoring the health of river systems and reflecting the water quality through assessing macroinvertebrates communities. The miniSASS samples the macroinvertebrate community in a river reach and compares the community present to the expected community under ideal natural conditions. The information garnered during a survey relies heavily on the accurate identification of macroinvertebrates by lows killed citizen scientists. This leaves a potential for errors in identification which may impact the accuracy of results and, ultimately, of the river health assessment. In response, we initiated the development of a smartphone application with built-in machine-learning algorithms for the automatic, real-time identification of macroinvertebrates. This report presents our data, methodology, and preliminary results from the automated identification algorithms.
(Location: IWMI HQ Call no: e-copy only Record No: H052345)
(1.32 MB)
The mini stream assessment scoring system (miniSASS) was developed as a citizen science biomonitoring tool for assessing the water quality and health of stream and river systems. A miniSASS survey involves sampling the aquatic macroinvertebrate community in a stream or river reach and using the known sensitivities and tolerances of the taxa present to infer information about the water quality and health of the stream or river. The quality of the outcomes of a miniSASS survey is dependent on good sampling technique and accurate identification of aquatic macroinvertebrates by low-skilled citizen scientists. As such, there is potential for errors in sampling and identification which may impact the accuracy of results. In response, we aimed to 1) develop a smartphone application (miniSASS mobile app with built-in machine learning (ML) algorithm for the automatic, real-time identification of aquatic macroinvertebrates) to assist in miniSASS surveys, 2) modernise and upgrade the miniSASS website to handle new data submissions (including photographs) and improve the user interface (UI), and 3) develop an online miniSASS training course. This report presents the methodology and preliminary results pertaining to these objectives.
(Location: IWMI HQ Call no: IWMI Record No: H052509)
(1.02 MB)
Freshwater systems are disproportionately adversely affected by the ongoing, global environmental crisis. The effective and efficient water resource conservation and management necessary to mitigate the crisis requires monitoring data, especially on water quality. This is recognized by Sustainable Development Goal (SDG) 6, particularly indicator 6.3.2., which requires all UN member states to measure and report the ‘proportion of water bodies with good ambient water quality’. However, gathering sufficient data on water quality is reliant on data collection at spatial and temporal scales that are generally outside the capacity of institutions using conventional methods. Digital technologies, such as wireless sensor networks and remote sensing, have come to the fore as promising avenues to increase the scope of data collection and reporting. Citizen science (which goes by many names, e.g., participatory science or community-based monitoring) has also been earmarked as a powerful mechanism to improve monitoring. However, both avenues have drawbacks and limitations. The synergy between the strengths of modern technologies and citizen science presents an opportunity to use the best features of each to mitigate the shortcomings of the other. This paper briefly synthesizes recent research illustrating how smartphones, sometimes in conjunction with other sensors, present a nexus point method for citizen scientists to engage with and use sophisticated modern technology for water quality monitoring. This paper also presents a brief, non-exhaustive research synthesis of some examples of current technological upgrades or innovations regarding smartphones in citizen science water quality monitoring in developing countries and how these can assist in objective, comprehensive, and improved data collection, management and reporting. While digital innovations are being rapidly developed worldwide, there remains a paucity of scientific and socioeconomic validation of their suitability and usefulness within citizen science. This perhaps contributes to the fact that the uptake and upscaling of smartphone-assisted citizen science continues to underperform compared to its potential within water resource management and SDG reporting. Ultimately, we recommend that more rigorous scientific research efforts be dedicated to exploring the suitability of digital innovations in citizen science in the context of developing countries and SDG reporting.
(Location: IWMI HQ Call no: e-copy only Record No: H052512)
(602 KB)
This report provides an overview of the mini Stream Assessment Scoring System (miniSASS) and South African Scoring System Version 5 (SASS5) as biomonitoring techniques for assessing the ecological condition of streams and rivers based on the identification of aquatic macroinvertebrates. While miniSASS relies on minimally trained citizen scientists to identify macroinvertebrates at the Order-level, SASS5 utilizes expertly accredited practitioners for finer resolution, even up to the family-level. However, the reliance on citizen scientists for miniSASS identification introduces limitations in terms of precision, accuracy, and reliability. To address these limitations, ongoing developments within the CGIAR Initiative on Digital Innovation include the creation of a miniSASS smartphone application, an upgraded website, an interactive online course, and a machine-learning identification algorithm to assist with photo identification. Additionally, a revised dichotomous key has been developed to improve operator identification during miniSASS surveys. Furthermore, the potential for upscaling the machine-learning identification algorithm to assist in identifying the 91 family-level taxa used in SASS5 assessments has been explored. The outcomes of these developments and explorations presented in this paper aim to enhance the overall effectiveness and reliability of both the miniSASS and SASS5 techniques. By leveraging digital innovation and incorporating machine-learning technology, we anticipate the efficiency, accuracy, and accessibility of biomonitoring assessments will significantly improve, ultimately contributing to a better understanding and management of our aquatic ecosystems.
(Location: IWMI HQ Call no: e-copy only Record No: H052516)
(17.6 MB)
The Enviro-Champs initiative was developed as a community driven, citizen science initiative in Mpophomeni township in Kwa-Zulu Natal (KZN), South Africa. Over time, the scope of work done and data collected by the Enviro-Champs has expanded. There is now recognition both locally and globally that the Enviro-Champs initiative shows great promise for national and global upscaling. However, several areas within the initiative remain where it could be improved, especially technologically. GroundTruth, in conjunction with technical and funding support from CGIAR Research Initiative on Digital Innovation and the International Water Management Institute (IWMI), engaged in a project which aimed to i) establish recruitment, training, and education tools to support establishment of a technologically integrated and upgraded Enviro-Champs initiative, ii) develop an outline for a training and education workshop for Enviro-Champs once they are hired, iii) improve data collection and reporting capacity and efficiency with a sustainable system (in collaboration with CGIAR and FormShare), and iv) pilot test technological improvements to the Enviro-Champs initiative within the Mpophomeni Enviro-Champs in conjunction with the South African National Biodiversity Institute (SANBI), and Umgeni Water. The overarching aim was to develop a technologically innovative and upgraded best-practice framework for the Enviro-Champs, from recruitment, through training and data collection, to data management and reporting. The primary outcome was to have a fully functional, digitally improved Enviro-Champs system in Mpophomeni, that could serve as a working template for upscaling the Enviro-Champs initiative elsewhere in Southern Africa or the world. This report reflects the process and outcomes of this project to date.
Powered by DB/Text
WebPublisher, from