Advances in ecological modelling of soil properties by self-organizing feature maps of natural environment of Lower Silesia (Poland)

Andrzej Stankiewicz, Piotr Kosiba

Abstract


The paper provides the use of self-organizing feature maps for determination of soil properties in its initial stage of development formed of massive rocks and how SOFM can be used for the study of environmental objects. The study area was Lower Silesia (Poland) overgrown with common, unique and protected vegetation of lichens, bryophytes and vascular plants. The parent rock of the studied soils consists of Miocene volcanites from the middle part of the Sudety Margin Fault. Soil samples were collected from 20 sites. The soil reaction (pH) and concentrations of Cd, Co, Cu, Fe, Mn, Mo, Ni, Pb, S, Ti, Zn in surface soils were analyzed. Statistical analysis was carried out by one-way ANOVA. The SOFM was used to demonstrate the non-linear ordination and visualization of soil properties. The SOFM showed the influence of parent rock on soil chemical properties generated by it. SOFM appeared to be effective and proper/fit for phenomena and processes taking place in natural environment and is useful in ecology and ought to be taken into account as a possible tool of estimation of various plants and their biotopes. The model can be useful as alternative techniques in modelling the ecological complex data, and provide a novel framework for the discovery and forecasting of ecosystem structure and behaviours in response to environmental changes.

Keywords


self-organizing feature maps; neural network; ecological modelling; soil composition; parent rock; basalt

Full Text:

PDF