Knowledge Discovery and Data Mining

We solve the "bottleneck" problems for people in the data analysis arena.

Spatial modelling of plant species potential habitats
V. Prokhorv, T. Rogova & A. Saveliev

Floristic and vegetation databases open new possibilities for spatial and dynamic ecosystem modelling in different scales, which may have a predictive value. The model worked out is based on a probabilistic evaluation of suitable habitat conditions for species requirements. It consists of three blocks:

· Potential habitats complying with ecological requirements of species model

· A model of potential species distribution in different landscapes

· A model of potential species distribution in different vegetation patches

For potential habitats modelling the Ciganov system for the broadleaf-coniferous zone was used. Indicator values of eight environmental factors (temperature, soil moisture, light etc.) were used in the building of the first block of the model. A probabilistic evaluation of environmental conditions for sample plots was made. Using the resulting data, the Ciganov system was transformed into a "local" probabilistic indicator system, considering actual natural data. The suitability of environmental conditions to species requirements was found out by comparing probabilistic values of each plot and the plant species. As a result, the suitability degree was established and a continuous spatial model was computed using the kriging method on a 1 km x 1 km grid. The probability of species occurrence in a certain landscape patch was established as species frequency in corresponding landscape type and it was evaluated for each patch on two hierarchical levels. The probability of species occurrence in different vegetation patches was established in a similar way, using vegetation maps. Finally, adding the results of three spatial models together for each grid node we got a continuous spatial distribution model representing species ecological requirements, landscape and phytocoenotic relations. This model was applied to 500 rare vascular plants species of the Tatarstan flora.

Also see: Meet the Team >