A Geo-Wiki campaign on capturing data on cropland extent took place in September 2016, collecting data at 36,000 unique locations around the world

Supported by the ERC CrowdLand Project (No. 617754) and the FP7 SIGMA project (No. 603719) the Cropland Campaign ran for a three-week period and involved more than 80 volunteers. The idea was to collect cropland extent in 300 x 300 m pixels in Geo-Wiki, which were each subdivided into 25 sub-pixels:

IIASA © IIASA

Cropland Campaign: The Geo-Wiki interface for collecting cropland information based on image interpretation. (a) is the sub-grid of pixels that users must classify; (b) is the Submit button that users must press once they have completed their interpretation; (c) allows the user to change the background imagery; (d) shows the ‘View in Google Earth’ button, which users can press to be shown the location in Google Earth so that that they can view historical imagery; and (e) shows the NDVI profiles that can be viewed when the user clicks on a location.

In this way, the cropland percentage could be calculated at each location. The sampling strategy to create the 36,000 sample units is described in detail in Laso Bayas et al. (2017) and Waldner et al. (2019). There were two methods used to ensure quality of the data. The first was the use of control data points, which consists of 1,793 sample locations validated by students trained in satellite image interpretation while the second consists of 60 expert validations for additional evaluation of the quality of the contributions. These control data are available when downloading the full data set. The first control data set was using during the campaign, where users were awarded points if they agreed with the controls or lost points if they disagreed. This was also intended as a way to help train the volunteers. Users were also provided with many examples to aid in classification as shown below:

IIASA © IIASA

Cropland Campaign: Definition and examples of cropland (yellow shading) and areas of non-cropland 

Cropland data collected during the cropland validation campaign, 2016 © IIASA

Cropland data collected during the cropland validation campaign, 2016

The second method to ensure quality was the requirement that each unique location be validated at least three times by different participants. In this way, consensus could be examined and locations with high disagreement could be revisited by experts. The spatial distribution of cropland collected from the campaign is shown on the left.

The data from the Cropland campaign are described in detail in a paper published in Nature’s Scientific Data (Laso Bayas et al., 2017); the data set can be accessed through the Pangaea data repository. Such a data set can be used for training classification algorithms or for independent validation of maps of cropland extent.   

Geo-Wiki

Waldner, F., Schucknecht, A., Lesiv, M., Gallego, J., See, L., Pérez-Hoyos, A. d'Andrimont, R., Thomas de Maet, T., Laso Bayas, J.C., Fritz, S., Leo, O., Kerdiles, H., Díez, M., Van Tricht, K., Gilliams, S., Shelestov, A., Lavreniuk, M., Simões, M., Ferraz, R., Bellón, B., Bégué, A., Hazeu, G., Stonacek, V., Kolomaznik, J., Jan Misure, J., Verón, S.R., de Abelleyra, D., Plotnikov, D., Mingyong, L., Singha, M., Patil, P., Zhang, M., Defourny, P. (2019): Conflation of expert and crowd reference data to validate global binary thematic maps. Remote Sensing of Environment, 221, 235-246, https://doi.org/10.1016/j.rse.2018.10.039

Laso Bayas, J.C., Lesiv, M., Waldner, F., Schucknecht, A., Duerauer, M., See, L., Fritz, S., Fraisl, D., Moorthy, I., McCallum, I., Perger, C., Danylo, O., Defourny, P., Gallego, J., Gilliams, S., Akhtar, I.H., Baishya, S.J., Baruah, M., Bungnamei, K., Campos, A. et al. (2017): A global reference database of crowdsourced cropland data collected using the Geo-Wiki platform. Scientific Data. 4: 170136, https://doi.org/10.1038/sdata.2017.136

Data Set available from Pangaea as follows:

https://doi.pangaea.de/10.1594/PANGAEA.873912

Publications

Waldner, F., Schucknecht, A., Lesiv, M. , Gallego, J., See, L. , Pérez-Hoyos, A., d'Andrimont, R., de Maet, T., Laso Bayas, J.C. , Fritz, S. , Leo, O., Kerdiles, H., Díez, M., Van Tricht, K., Gilliams, S., Shelestov, A., Lavreniuk, M., Simões, M., Ferraz, R., Bellón, B., Bégué, A., Hazeu, G., Stonacek, V., Kolomaznik, J., Misurec, J., Verón, S.R., de Abelleyra, D., Plotnikov, D., Mingyong, L., Singha, M., Patil, P., Zhang, M., & Defourny, P. (2019). Conflation of expert and crowd reference data to validate global binary thematic maps. Remote Sensing of Environment 221 235-246. 10.1016/j.rse.2018.10.039.

Laso Bayas, J.C. , Lesiv, M. , Waldner, F., Schucknecht, A., Duerauer, M., See, L. , Fritz, S. , Fraisl, D. , Moorthy, I., McCallum, I. , Perger, C., Danylo, O., Defourny, P., Gallego, J., Gilliams, S., Akhtar, I.H., Baishya, S.J., Baruah, M., Bungnamei, K., Campos, A., Changkakati, T., Cipriani, A., Das, K., Das, K., Das, I., Davis, K.F., Hazarika, P., Johnson, B.A., Malek, Z., Molinari, M.E., Panging, K., Pawe, C.K., Pérez-Hoyos, A., Sahariah, P.K., Sahariah, D., Saikia, A., Saikia, M., Schlesinger, P., Seidacaru, E., Singha, K., & Wilson, J.W. (2017). A global reference database of crowdsourced cropland data collected using the Geo-Wiki platform. Scientific Data 4 e170136. 10.1038/sdata.2017.136.

Fritz, S. , See, L. , Perger, C., McCallum, I. , Schill, C., Schepaschenko, D. , Duerauer, M., Karner, M., Dresel, C., Laso-Bayas, J.-C., Lesiv, M. , Moorthy, I., Salk, C.F., Danylo, O., Sturn, T., Albrecht, F., You, L., Kraxner, F., & Obersteiner, M. (2017). A global dataset of crowdsourced land cover and land use reference data. Scientific Data 4 p. 170075. 10.1038/sdata.2017.75.