Improving global land cover by playing games.

At present there is no single satellite-derived global land cover product that is accurate enough to provide reliable estimates of forest or cropland area to determine, for example, how much additional land is available to grow biofuels or to tackle problems of food security. This project aims to improve the quality of this land cover information by vastly increasing the amount of in-situ validation data available for calibration and validation of satellite-derived land cover. Land cover is an Essential Climate Variable (ECV), which is required to support the work of the United Nations Framework Convention on Climate Change (UNFCCC) and the Intergovernmental Panel on Climate Change (IPCC). All ECV are technically and economically feasible for systematic observation.

It is these variables for which international exchange is required for both current and historical observations. The Geo-Wiki ( system currently allows users to compare three satellite derived land cover products and validate them using Google Earth. However, there is presently no incentive for anyone to provide this data so the amount of validation through Geo-Wiki has so far been quite limited. The LandSpotting project will take a truly innovative approach by the addition of crowdsourcing through the development of a game. The game will engage users whilst simultaneously collecting a large amount of in-situ land cover information. The development of the game will be informed by the current raft of successful social gaming that is available on the internet and as mobile applications, many of which are geo-spatial in nature.

At the same time, the Geo-Wiki system will be modified to exploit the latest available satellite images and to use the acquired in-situ validation information to create new outputs: a hybrid land cover map, which takes the best information from each individual product to create a single integrated version; a database of validation points that will be freely available to the land cover user community; and a facility that allows users to create a specific targeted validation area, which will then be provided to the crowdsourcing community for validation. These outputs will turn Geo-Wiki into a valuable system for many users of land cover. This project will engage these users via workshops and ensure the future sustainability of the Geo-Wiki system.

The project duration will be 15 months. IIASA is the lead partner in the project responsible for overall coordination. TU Wien will provide expertise in game design and build the prototype game for collecting in-situ crowdsourced land cover validation information for Geo-Wiki. The other two partners are private companies who will be responsible for the development of a land cover server and the interfaces between different web services (EOX IT Services GmbH) and further enhancements to Geo-Wiki to create the new outputs outlined above and interface with the game environment (Fotec GmbH).


In the last decade, three global satellite-derived land cover products have been created: GLC-2000 (Fritz et al., 2003), MODIS (Friedl et al., 2002) and GlobCover (Defourney et al., 2006). These datasets are currently used as inputs to a range of different global, regional and national scale applications. For example, land cover maps provide the basis for the derivation of functional types used in large scale dynamic vegetation modelling. They are also important inputs to resource assessments of forest and agricultural land and global economic land use models. Moreover, identification of additional land availability for growing biofuel or food crops is essential, in particular in recent years where the price of crops has been at unexpectedly high levels.

However, a pixel-by-pixel comparison of these different landcover datasets reveals areas of the world where they do not agree, in some cases by large amounts (See and Fritz, 2006). This means that we do not precisely know how much land is currently forested or under cultivation because the uncertainty in the estimates provided by these products is too high. This has clear implications for determining deforestation rates, how much land is available for agriculture and levels of biodiversity. This also leaves users of these products with a difficult choice, i.e. which is the best product to choose and what effect will this choice have on a particular application?

There are three recent developments, however, that have opened up new and exciting possibilities for improving land cover validation.

The first is Google Earth, which has already been utilized in the creation of GlobCover (Defourny et al., 2009) and for the validation of the European forest cover map (Pekkarinen et al., 2009).

The second is the increasing trend in the willingness of ordinary citizens to act as human sensors and upload in-situ information to cyberspace, termed volunteered geographic information (VGI) (Goodchild, 2007) or crowdsourcing (Howe, 2008). This information can be textual or more usefully photographic, and there are now a multitude of examples of where users submit spatially referenced information to a website, e.g. wikimapia (, Openstreetmap (, MapAction ( and GoogleEarth/Panaramio.

The third important development is social gaming, which is a phenomenon that has taken off in the last five years. Games that are integrated within a social networking site such as Facebook illustrate the power to reach and continually engage a large number of individuals. The number of active Facebook users is estimated to be greater than 400 million, where 100 million are accessing Facebook from mobile devices (Facebook, 2010). Within Facebook, there are many games that encourage interaction with friends and provide a reward system to incentivise play, e.g. Farmville. Likewise there are many gaming applications for the mobile phone that store a vast amount of geographic information, e.g. Foursquare.

These three developments together provide new opportunities for validation of satellite derived land cover products and can fill the current gap in available in-situ data.

Degree of Innovation

There are several innovative elements to this project. The first is the addition of true crowdsourcing to Geo-Wiki to provide in-situ validation information for land cover. Crowdsourcing is an area that is relatively new but which is now rapidly taking off. There are websites where users validate specific points on the earth and upload photos (e.g. Degrees of Confluence and Geograph) but there is no current example of where crowdsourcing has been used for the in-situ validation of land cover products.

The second innovation is the merging of EO-derived data and the world of social gaming. This is possible because of the interdisciplinary nature of the of the project, which brings together researchers from the remote sensing community, game designers as well as industrial partners specialized in building the architecture and services for EO data. Most games use very simple, artificial backgrounds and landscapes with little attempt to represent the real world as observed remotely. The injection of EO-derived data to the gaming world adds new realism to the game environment that is currently lacking. It is also recognised that authenticity is a vital ingredient to selling products in today’s marketplace (Gilmore and Pine II, 2007). This realism is made possible by the recent democratisation of EO-derived data and the frequency at which EO-derived datasets are now available, e.g. a new up-to-date MODIS image is available on a daily basis.

Another innovation is the development of a hybrid land cover product, which has not been attempted before. The idea of hybridisation is not new, i.e. principles of data fusion and ensemble modelling are used to gain additional information or a more accurate estimate from multiple, redundant data sources in a range of different fields. Given that more than one land cover product exists, exploiting the synergistic effect of hybridisation is a natural progression for improving land cover information.


The project duration will be 15 months and is organized into six Work Packages as follows:

  1. Project Management: IIASA is the lead partner responsible for overall coordination of the project. The other partners are the Vienna Technical University, who will provide the game expertise (WP 4), EOX IT Services GmbH, who will be responsible for the development of a land cover server and the interfaces between different web services (WP3) and Fotec GmbH, who are tasked with providing further enhancements to Geo-Wiki (WP3). Each individual partner will also be responsible for local management of their project staff and resources.
  2. Requirements Analysis: All project partners will be involved in determining the requirements of the new Geo-Wiki system using the standard ISO Reference Model for Open Distributed Processing. This methodology is organized into five viewpoints (enterprise, information, service, technology and engineering), which will be used as a starting point for the requirements analysis.
  3. Server and Geo-Wiki Design & Implementation: The first part of this Work Package will design and deliver the server consisting of a product server; the processing engine and the land cover database (led by EOX IT Services GmbH). The second part of the Work Package (led by Fotec) will design and deliver the Geo-Wiki System consisting of an upgraded Geo-Wiki client with the new functionality defined in WP2 and the Google Earth API.
  4. Game Environment Design & Implementation: This Work Package deals with the development of the game, which will be led by Vienna Technical University. The starting point will be a review of existing social, mobile and online video games that currently engage a large number of users. The successful elements of these games will be used to inform the game design process. The approach taken will follow that outlined by the game designer Schell (2008). A prototype of the game will then be created within the Geo-Wiki environment.
  5. Service Validation and Acceptance: The server, the modified Geo-Wiki system and the game will all be validated against the original requirements set out in WP2. Where relevant, system acceptance tests will be based on ‘good practice’ from the ECSS (European Committee for Space Standardization) software engineering standards. The success of the game will be measured through the amount of land cover validation that is collected as well as the spatial coverage.  Moreover, the quality of the validation points will be scrutinized by comparing them to the standards of the CEOS Cal/Val sub-working group.
  6. Outreach and Exploitation: This Work Package will be lead by IIASA and will involve implementing the dissemination strategy for the project.


An iPad Tower Defense Game

"Defend your city from dangerous enemies!"

Available on the App Store here.



"Show your tagging skills!"

Play it on Facebook here.

A Facebook Strategy Game

"Conquer the world in this exciting strategy game!"

Play it on Facebook here.

Pixel Art World

"Create with people from all around the world a gigantic painting!"

Available soon!


Tobias Sturn, Dietmar Pangerl, Linda See, Steffen Fritz, Michael Wimmer
Landspotting: A Serious iPad Game for Improving Global Land Cover
In GI-Forum. July 2013.
[Paper] [Slides]

Tobias Sturn, Michael Wimmer, Peter Purgathofer,
Steffen Fritz

Landspotting - Games for Improving Global Land Cover
In Foundations of Digital Games, pages 117-125. May 2013.

Dietmar Pangerl
A Serious iPad Game for Improving Global Land Cover

[Poster] [Thesis]

Florian Felberbauer
Games with Purpose - Improving 3D Model and Land Cover Data using Crowdsourcing
[Poster] [Thesis]

Steffen Fritz
, Peter Purgathofer, F. Kayali, M. Fellner, Michael Wimmer, Tobias Sturn, Josef Schuh, G. Triebnig, S. Krause, F. Schindler, M. Kollegger, Christoph Perger, M. Duerauer, W. Haberl, Linda See, Ian McCallum
Landspotting: Social gaming to collect vast amounts of data for satellite validation

Christoph Perger, Ian McCallum, Franziska Albrecht, Linda See, Steffen Fritz, David Thau
Improved Cropland Mapping in Ethiopia
In GI-Forum. July 2013.

Steffen Fritz, Ian McCallum, Franziska Albrecht, Linda See, Christoph Perger, Mathias Karner, Dmitry Shchepashchenko, Marijn van der Velde
Geo-Wiki Competitions


The international Landspotting project team includes the following companies / institutes:


International Institute for Applied Systems Analysis (IIASA), Schlossplatz 1, A-2361 Laxenburg, Austria

IIASA is an international, scientific, non-governmental institute conducting policy-oriented research addressing problems such as climate change, which have to be seen in a global context and can be resolved only by international cooperative action and problems of common concern to many countries that need to be solved at the national level. Examples of the latter type include energy security, demographic change and sustainable development. IIASA is funded by scientific institutions in countries all over the world, which grants research unconstrained by political or national self-interest, even though some funding comes from external sources, such as EU-funded projects.


  • Fotec - Forschungs- und Technologietransfer GmbH is the research subsidiary of the University of Applied Sciences Wiener Neustadt (FHWN). The link to the university and worldwide co-operation with universities and research institutions make Fotec the interface between teaching, science and industry.
  • EOX IT Systems GmbH is a young, innovative company, founded in the year 2008 by people having the strong will to realize the market perspectives present in the field of spatial observation information technology. At present, six managerial and engineering staff are working at EOX.  The main area of EOX activities concerns the development and advancement of e-Environment and geo-spatial information infrastructures with special emphasis placed on satellite Earth Observation systems and next generation applications.
  • Two groups from the Faculty of Informatics at the Vienna University of Technology participate in the project:
    The Institute of Computer Graphics and Algorithms performs extensive fundamental and applied research in computer graphics. Areas research include modeling and rendering for computer graphics, scientific visualization, virtual environments, color, and technology for computer games. Besides research projects, the institute specializes in consulting and technology transfer as well as computer graphics related education on both undergraduate and graduate level.
    The Human Computer Interaction Group at the Institute of Technology Design and Assessment is concerned with research and teaching with interdisciplinary questions in informatics, integrating theories and methods of social sciences and design sciences into computer science. Areas of research include socio-technical system design using creative and participatory methods; design and development of multimodal cooperative environments and tangible user interfaces; design and usage of virtual spaces for different domains and stakeholders, especially in art, architecture, health care, new media, electronic learning and digital games; methods of evaluation and analysis.