SAEON scientists win United Nations data science award at COP23

Tweet

SAEON scientists have won an award from the United Nations Data for Climate Action Challenge contest for a prototype near-real time monitoring tool that reports on the state and changes in vegetation based on satellite observations.

The challenge

Run by the U.N. Global Pulse and sponsored by Western Digital and the Skoll Foundation, the contest tasked contestants with developing research papers or data visualisations that address problems tied to climate change and help achieve the U.N. Sustainable Development Goals (SDGs), with emphasis on SDG13: Climate Action.

Professor Adam Wilson from the State University of New York at Buffalo (centre) represented the winning team at the awards ceremony on November 12  at the Sustainable Innovation Forum, a companion event to the annual United Nations Climate Change Conference (COP23), which took place in Bonn, Germany, from November 6-17. The change detection tool (EMMA) has been developed for Fynbos vegetation (a global biodiversity hotspot), but has potential to be adapted to many other ecosystems (Photo: Jasper Slingsby)

U.N. Global Pulse partnered with companies such as Orange, Microsoft, PlanetLabs and others, to provide contestants with private sector data sources and tools to address problems tied to climate change.

“Big Data and analytics is a game changer for policy-making. There is just one problem: Much of the data is privately owned and we have to find ways to share it,” said Robert Orr, Special Advisor to the U.N. Secretary General on Climate Change.

U.N. Global Pulse Director, Robert Kirkpatrick, said he hoped the competition would pave the way for a global data philanthropy movement, whereby the private sector makes their data available for the public good.

SAEON’s winning submission

SAEON’s team submitted a research paper that included prototype web and smartphone applications for an ecosystem monitoring tool called the “Ecosystem Monitoring for Management Application (EMMA)” (see http://www.ecologi.st/post/emma/ for more detail). The tool uses MODIS NDVI data from NASA’s Terra and Aqua satellites to detect abnormal changes in the greenness of Fynbos vegetation in near-real time.

Tools that provide spatially explicit biodiversity and environmental data, monitored over the long term, but made available in near-real time, are the critical basis for ecosystem management and stewardship. Existing tools, like Global Forest Watch, a partnership convened by the World Resources Institute, have shown the value of being able to track change in near-real time, providing insight into global change impacts on forest ecosystems and providing deforestation alerts in near-real time.

Unfortunately, near-real time vegetation change detection tools are largely limited to forests, with few tools for non-forest ecosystems such as the shrublands, woodlands and grasslands that dominate South Africa and much of the rest of the globe. Detecting abnormal change in non-forest ecosystems is highly challenging. The state of the vegetation varies dramatically due to natural disturbances, long-term trends or cyclical functions, such as those relating to fire, postfire recovery or seasonality.

By comparing observed MODIS NDVI data with the expectation based on a statistical model that accounts for season, environmental conditions and time since fire, EMMA is capable of distinguishing between unnatural aberrations - such as declines resulting from drought, development or invasive species - and natural variation.

EMMA highlights MODIS pixels on the ground (~250 by 250m squares), which can then be examined remotely using high frequency, high-resolution imagery from Planet Labs (3m pixels recorded almost every day). Should this imagery fail to provide a clear diagnosis of the impact, the team also developed a prototype smartphone application (VeldWatch) to aid field inspections. VeldWatch allows field rangers, landowners and citizen scientists to navigate to locations highlighted by EMMA to investigate impacts, but also allows them to map any other locations where they have spotted threats to ecosystems such as invasive plants, landslides or illegal clearing of vegetation (see http://www.ecologi.st/post/veldwatch/ for more information).

SAEON’s team was led by Dr Jasper Slingsby of the Fynbos Node, and included Dr Glenn Moncrieff of Ixio Analytics and a SAEON Research Associate, and collaborator Professor Adam Wilson from the State University of New York at Buffalo.

The winners

The contest whittled down 450 entries from 63 countries to 97 finalists, awarding six prizes. These included three thematic prizes for research papers, two prizes for data visualisations and one grand winner.

SAEON’s team was led by Dr Jasper Slingsby of the Fynbos Node (left), and included Dr Glenn Moncrieff of Ixio Analytics and a SAEON Research Associate (right), and collaborator Professor Adam Wilson.

SAEON’s submission won the Thematic Award for Climate Mitigation, addressing SDG13: Climate Action and SDG15: Life on Land, among others (1:02 on the video http://www.datamakespossible.com/2017/11/06/data-innovation-generating-climate-solutions/).

Other winners included the Georgia Institute for Technology (Thematic Award for Climate Adaptation), Yale University (Thematic Award for Climate & Other Sustainable Development Goals) and Eriktronic Engineering (for their data visualisation).

The Grand Prize was awarded to a collaboration between the Mexican National Institute for Ecology and Climate Change (INECC) and the University of California at Berkeley for research that seeks to support greater electromobility in the Metropolitan Area of Mexico City. They studied mobility and social patterns to propose optimal routes for electromobility policies and the development of necessary infrastructure, such as power stations and transport hubs of multimodal transport. They also evaluated the reduction of emissions of CO₂ and PM2.5 fine particles associated with three types of vehicle electrification policies for: a) taxis, b) buses and c) light vehicles (read more).

Where to from here?

Preliminary testing of EMMA and VeldWatch is providing very positive results. The system has the potential to be a valuable asset for a range of ecosystem management applications - from highlighting illegal clearing of indigenous vegetation or identifying spread of invasive alien species, to monitoring invasive plant control operations or detecting drought impacts.

From a research perspective, the system will accumulate a record of verified land-cover change and can also be used for monitoring and understanding ecosystem carbon and water dynamics and their responses to global change impacts.

While this initial phase is focused on Fynbos ecosystems, there is much potential to adapt the system and expand it to include additional ecosystems in South Africa and elsewhere.

EMMA and VeldWatch are still in the development and testing phase, with various components currently being piloted by a small group of test users. Moving beyond the proof-of-concept phase to an operational system will require additional funding for staff to maintain the system and strengthen partnerships with collaborators and potential users.