Seite - 14 - in options, Band winter 2017/2018

www.iiasa.ac.atoptions ◼ winter 2017/1814 Right time, right place In 1998, Steffen Fritz was working at the Joint Research Centre in Ispra, Italy. His job was to process satellite images of the Earth, creating maps of global land cover. These data were an important input to environmental models including the IIASA Global Biosphere Management Model (GLOBIOM), which is used to understand competition for land between agriculture, forestry, and bioenergy. But the satellite data had many gaps, and sometimes the resolution was too poor to confidently identify land cover types, differentiate settled areas from abandoned ones, or forest from cropland. This was a big problem for researchers working on large-scale systemic models. Fritz says, “You can have the best tools and the best global models, but in order to make sure you’re on the right track, it’s very important to train and validate the model on historical data. And there was a lot of uncertainty in the data.” It was around this time that Google Earth was launched. Where data from the NASA Landsat satellite covered the Earth at 30-meter resolution, Google Earth images—from satellite imagery, aerial photography, and Geographic Information Systems data—zoomed into a miraculous 50 centimeters, allowing people in many places to identify their house, or even their car in the images. “It was a major jump, you could see structures where you couldn’t see anything before. It was the first time you could really validate maps and understand what you were seeing,” says Fritz. But higher resolution also meant greater data volumes. How could researchers possibly sort through the hundreds of thousands of images it would take to check against the satellite land-cover data on a global scale? “We realized that to get something useful out of all these data, we needed a crowd. That was the only way,” he says. In 2007 Fritz moved to IIASA to pursue a crowdsourcing project. Working with Fritz and IIASA researchers Ian McCallum and Dmitry Shchepashchenko, Christoph Perger, then a student at the University of Applied Sciences, Wiener Neustadt, designed the first iteration of the crowdsourcing platform Geo-Wiki. It was a simple interface where people could compare existing land-cover maps with high- resolution imagery, checking the data against multiple sources to determine where it was correct and where it could be improved. If you build it, will they come? It was a good idea. But when Geo-Wiki was launched, the researchers were disappointed. By the end of 2009, the site had only 109 registered users. Linda See joined the team in 2010. She says, “I first came to IIASA when Steffen’s group was building Geo-Wiki. It was a great concept, but nobody was participating yet. We started to think about how we could get more people involved. That’s when we started with competitions and campaigns.” In order to engage and motivate participants, the team experimented with several strategies. They started running campaigns for limited periods of time, with a clear goal. Competitions allowed participants to compete for prizes such as electronic devices, small cash awards, or even coauthorship on a paper. More recently the team has experimented with gamification, developing simple games that can be played on a mobile phone or tablet, and micropayments for each quality contribution. A two-way street The strategies worked—the number of registered Geo-Wiki users had grown to nearly 15,000 by September 2017. But keeping people engaged has remained one of the major challenges for the team. Inian Moorthy, who manages the EU-funded LandSense project at IIASA, says, “There’s the initial challenge of awareness—making sure people find out about your project. But after that, sustaining participation once someone actually joins, this is something that we are continually grappling with. Whomever we get into the project, how can we keep them involved and engaged and contributing?” Bilous, who contributed to a recent field size campaign as well as an earlier global forest map, works in forestry research at a much more detailed scale of specific forest ecosystems. He was inspired to get involved by the potential of contributing to something larger. “Our planet is large and beautiful, even researchers do not always imagine a diversity of landscapes on the Earth’s surface. In Geo- Wiki…the scale of the project is impressive. I really want the Geo- Wiki team to succeed in their project, so I’m trying to help,” he says. Cipriani adds, “With Geo-Wiki I can make a difference for the benefit of the society, helping understand landscape use and improving the interpretation of satellite imagery to better detect changes occurring on the Earth’s surface.” At the same time, volunteers gain a better understanding of the scientific process, as well as experience and knowledge of remote sensing data—an aspect that has attracted many teachers and students to the project. Fritz says, “There is a massive misconception of the accuracy of remotely sensed data. Many times people think because it’s coming from space, it must be correct. But there is a lot of uncertainty, which you see when you look more closely.” The project also gives another tangible benefit back to communities: providing free access to all the data they produce. The scientists post all the processed data on the Geo-Wiki platform, and have published several data sets in the new Nature journal Scientific Data. From citizen science to co-production As the IIASA citizen science team has grown, they have increasingly branched out to different fields, and engaged even more closely with the stakeholders and communities they work with. Wei Liu came to IIASA in 2012 to work in the IIASA Risk and Resilience Program. He joined forces with the Geo-Wiki group to apply citizen science to a flood resilience project in Nepal. In Registered Geo-Wiki users 2009-2017 0 2,000 4,000 6,000 8,000 10,000 12,000 14,000 16,000 2009 2010 2011 2012 2013 2014 2015 2016 2017 Year