Research of the Arctic Soils Using an Artificial Neural Network

Abstract

Desert-Arctic soils -balasamy (W-C1), are found in the most northerly position in the Arctic. These soils are characterized by a light granulometric composition and are formed in the areas recently released from glaciers, and develop under a crust of blue-green algae. Arctic soils (AO-AY-BC-C) are common on loamy and gravellyloamy soils (Severnaya Zemlya, Novaya Zemlya, Franz Josef Land, North of the Taimyr Peninsula). They are characterized by wedge-shaped horizons, and are formed in the form of polygons with a diameter of 0.5-1.0 m under moss-shrub vegetation. Carbonate pelozems (WSA-SSA) are found on deluvial deposits of carbonate rocks on loamy-gravelly soils. The vegetation cover is represented by lichens and rare specimens of flowering plants. In the Arctic tundra, on the most drained areas on loamy and gravelly-loamy soils, humified weak-clay (gley) soils (AO-A-CRMg-C(D)) are common. In terms of morphology and chemistry, these soils are similar to Arctic soils, but differ from them in the large development of wedge-shaped horizons. In this work, the composition of Arctic soils was studied using a neural network.