Earth Observation for Land and Water Challenge: A Hackathon Leveraging the DUNIA Platform
By David Ongo with contributions from the ICPAC Climate Change Technical Working Group
The groundbreaking land & water hackathon challenge organized by IGAD Climate Predictions and Applications Centre (ICPAC), EO Africa DUNIA Cloud computing platform, and Kenyatta University brought together students, early career developers, and professionals drawn from different fields to leverage Earth Observation (EO) data to address critical land and water management challenges in East Africa.
Initially, the hackathon targeted participants from universities around Nairobi; in the end, it attracted participants from the ICPAC member countries. This edition targeted Kenyan universities. The face-to-face event was held at the Kenyatta University Spatial and Environmental Planning GIS Lab from the 1st to the 5th, of July 2024.
The pool of talented participants was drawn from Kenyatta University, University of Nairobi, Jomo Kenyatta University of Agriculture and Technology, Meru University, and Technical University of Kenya. Early career participants were drawn from various startup companies in Kenya. Ten teams of three were formed after training on the DUNIA cloud computing platform, GMES & Africa services under the IGAD Climate Predictions and Applications Centre (ICPAC)—the hackathon aimed to foster innovative solutions through collaboration and cutting-edge technologies. Participants utilized the platform, a powerful tool for EO data analysis, to tackle various themes related to sustainable land and water management. The event highlighted the importance of integrating advanced data technologies to enhance the region’s environmental resilience and resource management.
The teams were tasked to provide solutions to a pool of challenges including;
❖ Hydrology and Water Resources
❖ Landscapes and Forest Ecosystems
❖ Agriculture and Food Security
❖ Protected Areas
Throughout the hackathon, teams engaged in intensive collaborative work, leveraging EO data to develop actionable insights and solutions for their themes. The DUNIA platform allowed participants to analyze satellite data effectively, extracting valuable information for sustainable land and water management practices.
Outcomes and Impact
The hackathon culminated in the presentation of innovative solutions developed by the participants. These solutions showcased the potential of EO data in transforming land and water management practices across East Africa. The event fostered creativity and technical skills and emphasized the importance of data-driven decision-making in addressing environmental challenges.
One notable project from the hackathon is Team 2’s “Forest Change in the Aberdare Forest and Its Impact on Air Quality.” This project uses Sentinel-2 and Sentinel-5P data to analyze forest cover changes and their relationship with nitrogen dioxide (NO2) levels from 2018 to 2024. The team employed tools like the Dunia platform and Google Earth Engine for data processing and visualization. Their findings indicate that increased forest cover decreased atmospheric NO2 levels, while forest fires contributed to NO2 pollution. The project underscores the importance of forest conservation and fire management for air quality improvement.
The Aberdare Forest, a vital ecological region in Kenya, has undergone significant changes from 2018 to 2024. One of the teams analyzed these changes and their impact on air quality, focusing on nitrogen dioxide (NO2) levels. They found an increase in forest cover from 2018 to 2021 and a decrease from 2021 to 2024, mainly due to forest fires in 2022, 2023, and 2024. Increased forest cover correlated with decreased NO2 levels, while decreased forest cover led to increased NO2 levels. Their key recommendations include;
· Strengthening forest conservation efforts
· Implementing comprehensive fire management and prevention strategies
· Engaging local communities in conservation initiatives.
Another significant project from the hackathon team was the “Maize and Cassava Field Crop Classification to Address Food Security.” This project, which harnesses the power of remote sensing, machine learning, and geospatial technologies, has the potential to revolutionize crop yield predictions. By accurately identifying and categorizing maize and cassava fields, the project aims to optimize resource allocation and monitor crop health, thereby enhancing agricultural productivity and contributing to sustainable development and economic growth.
The project focuses on Busia County in Western Kenya, a key agricultural area with a tropical rainforest climate, producing crops like maize, millet, beans, cassava, and sugarcane. Utilizing Google Earth Engine due to technical issues with the DUNIA platform, the team visualized masked images to analyze agricultural fields, calculating various indices such as NDVI, LAI, MSAVI, GCI, IRECI, and S2REP to detect trends in crop health and growth over time. These indices helped create histograms to visualize changes in vegetation and field conditions.
The team uploaded the training data for cassava and maize, performed an accuracy assessment of the classification model, and visualized the classified image. With an accuracy rate of 83.96%, the model provides valuable predictive capabilities, enabling farmers and policymakers to optimize resource allocation, manage risks, and make informed decisions to improve agricultural productivity. This strategic approach ensures consistent production, bolsters food security, and encourages sustainable farming practices. The team’s work exemplifies the innovative use of EO data to tackle food security challenges in Busia County.
Another significant project by the teams is the analysis of “Lake Baringo’s Water Extent Changes.” This project focuses on monitoring the temporal changes in the water extent of Lake Baringo using Sentinel-2 satellite imagery. By leveraging the Dunia platform and Google Earth Engine, the team processed and visualized data to assess the impact of climatic variations and human activities on the lake’s water levels. The water area showed fluctuations correlating with seasonal rainfall patterns and human activities. The team used rainfall data from the Kenya Open Data portal to predict future water levels, highlighting potential flood risks or water scarcity. Their key recommendations include monitoring climatic variations and human activities affecting water levels and implementing sustainable water management practices to mitigate adverse impacts.
The success of this hackathon underscores the need for continued investment in EO technologies and collaborative initiatives. By harnessing the power of EO data, stakeholders can develop resilient systems to combat climate change, ensure sustainable resource management, and improve food security in the region.
For more details on the concept and structure of the hackathon, you can access the concept note here.
Acknowledgments
The event was made possible through the dedicated efforts of the GMES & Africa IGAD Climate Predictions and Applications Centre Consortium, EO Africa DUNIA, Kenyatta University, and all participating startups, students, and developers. Their commitment to innovation and sustainability is paving the way for a more resilient and secure future for East Africa.
The Earth Observation for Land and Water Challenge hackathon has set a new benchmark for collaborative innovation in the region. By leveraging advanced technologies and fostering a spirit of teamwork, the event has demonstrated the immense potential of EO data in solving some of the most pressing environmental challenges of our time. As the participants return to their respective fields, the knowledge and solutions developed during this hackathon will undoubtedly contribute to the sustainable management of land and water resources.