A new hybrid land cover dataset for Russia: a methodology for integrating statistics, remote sensing and in situ information

Despite being recognized as a key baseline dataset for many applications, especially those relating to biogeochemical cycles, land cover products in their current form are limiting. Typically they lack the thematic detail necessary for driving the models that depend upon them. This study has demonstrated the ability to produce a highly detailed (both spatially and thematically) land cover/land use dataset over Russia – by combining existing datasets into a hybrid information system. The resulting dataset contains detailed subclasses of land cover and attributes necessary for biogeochemical modeling. In lieu of suitable validation data, a confidence map was produced creating six classes of confidence in the agreement between the various remote sensing and statistical datasets. In specific regions, a significant difference between the remote sensing products and the official statistics was observed. For example, in the northwest of Russia the statistics appear to be underreporting the amount of forest land which has likely been increasing in recent decades because of encroachment of forests on abandoned marginal agricultural land.