Impact Assessment And Stochastic Modeling Of Morphometric Parameters Of Thermokarst Hazard For Unpaved Roads

Abstract

Active construction of new roads and other linear structures requires new techniques for the natural hazard assessment. These techniques can involve both stochastic modeling and remote sensing data (RSD). First, the dynamics of thermokarst appearance along an unpaved road (winter road) was analyzed. Then a probabilistic model of the thermokarst morphological pattern was developed for the area in the vicinity of a linear structure, a road in particular. The model operates with initial assumptions based on the physical parameters of thermokarst development and includes relations for estimating the distribution of morphometric dimensions of thermokarst depressions (ponds). The model was empirically tested for the study area, which represented a site with an unpaved road located in West Siberia region. To verify the model, we calculated the correlation coefficient values for the length of the focus projections on the linear structure and the perpendicular axis and compared the empirical distribution of the projections with the theoretical lognormal distribution using the Pearson’s criterion. The proposed model assumptions appeared to be valid for the study area, which makes it possible to proceed to the problem of probabilistic impact risk assessment to a linear structure by foci of human-induced thermokarst.