The v-transformed copula-based simulation technique is a versatile tool employed across fields such as hydrology, environmental science, soil science, finance, and mining. In the context of mining, it is specifically utilized for estimating the panel-wise ore and metal tonnage in ore deposits. Traditional copula-based techniques, however, are grounded in the assumption of strict stationarity, which often falls short in accurately evaluating complex geological structures. This inadequacy stems from the failure of stationary models to account for the transitional behavior prevalent in geological processes, leading to gradational contact points or soft boundaries between ore and waste zones. Addressing this shortcoming, the study presented here devises a non-stationary v-transformed copula model tailored for simulating ore bodies and estimating recoverable reserves. Central to the efficacy of this model is the accurate representation of gradational geological boundaries, which is achieved through the integration of an automatic boundary detection algorithm. The model is put to the test through its application to a renowned open-pit copper deposit in India. The efficacy of this model for reserve estimation is rigorously assessed by comparing grade-ore tonnage and grade-metal tonnage curves from the non-stationary and stationary copula models, and blast hole data. This comparative approach provides a robust means of evaluating the performance of the new model. The findings of the study indicate that the non-stationary copula based simulation model outperforms the stationary model.