Automated updating of medium-scale soil maps

摘要

An approach towards an automated updating of medium-scale soil maps via imitation of traditional mapping technologies is suggested. It is based on formulation of the rules of mapping in the form of classification trees for separating different soil cover patterns and on creation of the maps of soil-forming factors with the use of satellite data. Algorithms for mapping alluvial soils (Fluvisols), eroded (abraded), and anthropogenically transformed soils are presented. This approach was tested for the southern (Trans-Oka) part of Moscow oblast. The model for an automated soil mapping was realized using ILWIS software. The polygons of alluvial soils were mapped with a higher accuracy via the automated separation of floodplains according to the digital terrain model. The total area of alluvial soils shown on the medium-scale soil map decreased from 373 to 340 km(2). Calculations of slope angles according to digital terrain models allowed us to localize soil cover patterns with participation of eroded soils with a higher accuracy; their area decreased insignificantly: from 791 to 781 km(2). Anthropogenically transformed soils of building areas were mapped for the territory of Moscow oblast on the basis of satellite data for the first time. Their areas were delineated taking into account land use types and comprised 551 km(2), i.e., 15.4% of the total area (3570 km(2)) of the Trans-Oka part of Moscow oblast.