Probabilistic evaluation of emergency levels associated with the stability of a mining tailings dam

ABSTRACT This study introduces a probabilistic framework to evaluate emergency levels associated with tailings dam stability, using Monte Carlo simulations combined with three limit equilibrium methods (Spencer, Sarma, and Morgenstern-Price). For each scenario, including the current operating condition and thresholds defined by the Brazilian standard (FS = 1.50, 1.30, and 1.10), strength parameters were treated as random variables, generating FS distributions from which failure probability (Pf) and reliability index (β) were derived. Results show that small reductions in FS result in nonlinear increases in and significant declines in. While the current condition indicates high reliability (β ≈ 6.0, Pf < 10-9), scenarios with FS = 1.30 and FS = 1.10 show substantial risk escalation, with Pf rising to 10-4 and >10%, respectively. Findings demonstrate that deterministic FS thresholds, as currently adopted in Brazilian regulations, may misrepresent structural safety. Our approach provides basis for emergency classification in dam safety management.