Geodesic measurements of mining area deformations indicate that their description fails to be regular, as opposed to what the predictions based on the relationships of the geometric-integral theory suggest. The Knothe theory, most commonly applied in that case, considers such parameters as the exploitation coefficient a and the angle of the main influences range tgβ, describing the geomechanical properties of the medium, as well as the mining conditions. The study shows that the values of the parameters a = 0.8 and tgβ = 2.0, most commonly adopted for the prediction of surface deformation, are not entirely adequate in describing each and every mining situation in the analysed rock mass. Therefore, the paper aims to propose methodology for determining the value of exploitation coefficient a, which allows to predict the values of surface subsidence caused by underground coal mining with roof caving, depending on geological and mining conditions. The characteristics of the analysed areas show that the following factors affect surface subsidence: thickness of overburden, type of overburden strata, type of Carboniferous strata, rock mass disturbance and depth of exploitation. These factors may allow to determine the exploitation coefficient a, used in the Knothe theory for surface deformation prediction.
An analysis of the impact of mining with caving on the surface shows that a type of rock mass strata seems to be one of the critical factors affecting the process. Correlating the values of mining-induced surface deformation with the rock mass structure and the state of its disturbance is of crucial importance. Therefore, if other mining conditions are left unaffected, then those factors exert the key influence on a course and distribution of subsidence and rock mass deformation. A proper description of rock mass type and properties also seems rational for a proper determination of prediction parameters, especially in the case of a multi-seam coal mining, and/or the exploitation carried out at considerable depths. A general outcome of the study discussed in this paper is the development of the methodology and model practices for determining the rock mass type and, as a result, for selecting the optimal values of parameters for predicting the values of surface subsidence in relation to particular geological and mining conditions. The study proves that the type of rock mass may be described by such factors as the influence of overburden strata, the influence of Carboniferous layers, the disturbance of rock mass and the depth of exploitation.