Torralba J., Ruiz L. Á., Georgiadis C., Patias P., Gómez-Conejo R., Verde N., Tassapoulou M., Bezares Sanfelip F., Grommy E., Aleksandrowicz S., Krätzschmar E., Krupiński M., Carbonell-Rivera J. P. (2021), Methodological proposal for the identification of marginal lands with remote sensing-derived products and ancillary data

The concept of marginal land (ML) is dynamic and depends on various factors related to the environment, climate, scale,culture, and economic sector. The current methods for identifying ML are diverse, they employ multiple parameters andvariables derived from land use and land cover, and mostly reflect specific management purposes. A methodologicalapproach for the identification of marginal lands using remote sensing and ancillary data products and validated on samplesfrom four European countries (i.e., Germany, Spain, Greece, and Poland) is presented in this paper. The methodologyproposed combines land use and land cover data sets as excluding indicators (forest, croplands, protected areas,impervious areas, land-use change, water bodies, and permanent snow areas) and environmental constraints informationas marginality indicators: (i) physical soil properties, in terms of slope gradient, erosion, soil depth, soil texture, percentageof coarse soil texture fragments, etc.; (ii) climatic factors e.g. aridity index; (iii) chemical soil properties, including soil pH,cation exchange capacity, contaminants, and toxicity, among others. This provides a common vision of marginality thatintegrates a multidisciplinary approach. To determine the ML, we first analyzed the excluding indicators used to delimit theareas with defined land use. Then, thresholds were determined for each marginality indicator through which the landproductivity progressively decreases. Finally, the marginality indicator layers were combined in Google Earth Engine. Theresult was categorized into 3 levels of productivity of ML: high productivity, low productivity, and potentially unsuitable land.The results obtained indicate that the percentage of marginal land per country is 11.64% in Germany, 19.96% in Spain,18.76% in Greece, and 7.18% in Poland. The overall accuracies obtained per country were 60.61% for Germany, 88.87%for Spain, 71.52% for Greece, and 90.97% for Poland.

Torralba J., Ruiz L. Á., Georgiadis C., Patias P., Gómez-Conejo R., Verde N., Tassapoulou M., Bezares Sanfelip F., Grommy E., Aleksandrowicz S., Krätzschmar E., Krupiński M., Carbonell-Rivera J. P. (2021), Methodological proposal for the identification of marginal lands with remote sensing-derived products and ancillary data
doi:10.4995/CiGeo2021.2021.12729