Sarcheshmeh copper deposit is located in the Uromiyeh-Dokhtar volcanic belt, 50 km south of Rafsanjan and 160 km southwest of Kerman. Estimating rock mass uniaxial compressive strength (UCS) and other geothechnical parameters are vital to determining optimum and stable mine slopes, designing new exploratory boreholes and blast holes, mine planning, optimizing production scheduling and even designing crusher feed in Sarcheshmeh copper mine. To get an estimation of uniaxial compressive strength (UCS) throughout mine 3D extent it is necessary to use available point load test data from all old and new boreholes. The point load strength test is one of the cheapest and fastest methods used in estimating UCS. In the first stage of this study, the UCS values are estimated using point load measurements along with other qualitative geological and geotechnical parameters including rock type, alteration type and intensity at Sarcheshmeh copper mine. Due to having both saturated and unsaturated UCS and point load values, the data sets were grouped first and their linear and non-linear correlation and regression lines between different groups were obtained. These regression lines were then used to convert the saturated values to unsaturated ones because of the majority of the available point load measurements were obtained with unsaturated condition. The final result of this stage resulted in having 67 rock units having UCS values in Sarcheshmeh copper mine. In the next step the UCS values were estimated at the center of a three dimensional blocky solid model with the cell size of 12.5*12.5*6.25 meter throuout ore body extent. The three-dimensional distribution of UCS in Sarcheshmeh copper mine was obtained using different algorithms including statistical-structural techniques (nearest neighbor, inverse distance to a power), geostatistical (ordinary kriging, indicator kriging) and neural network methods. Following estimation of UCS at block centers using different methods, the obtained results and performance of each method were compared and validated through employing the 21 set asided borehole data. The comparison of results obtained by above mentioned methods was based on matching histograms of measured against predicted values, the well known scatter plot of measured against predicted values (the so called QQ plot) and the root mean square error (RMSE) of the difference between measured and predicted values from 21 boreholes in four different levels including 2200, 2250, 2300 and 2350 levels. The comparative results showed that the most appropriate estimation methods with least RMSE value is the nearest neighbor method and inverse distance to a power of 3 respectively.
