In 2006, an unprecedented atmospheric confined space accident took place in a sampling shed at the Sullivan Mine in Kimberley, British Columbia. This accident suggests that a risk assessment should be carried out on a regular basis at mine reclamation sites for many years after closure. in this paper, an Atmospheric Fuzzy Risk Assessment (AFRA) tool is described that can assess atmospheric risk given heuristic and measured data at such sites. It can also serve to transfer knowledge about atmospheric hazards in an enclosed structure. The system uses fuzzy logic to input and output information and to perform weighted inferencing. The paper describes the developmental process as well as system verification and validation based on a number of known test and reference waste dumps. AFRA is a heuristic expert system based on fuzzy logic and the first tool that was developed to assess the atmospheric risk of mine waste dumps. The atmospheric risk is estimated by fuzzy Mamdani system given the values of four major elements of risk comprising of: gas generation, gas emission, gas confinement, and human exposure. The ability of AFRA to adapt its risk assessment to different climate conditions is explained. There are many physical, chemical, and environmental factors which fluctuate over time affecting oxygen-depletion in waste dumps. AFRA can help mining engineers and mine managers recognize this type of danger when conducting a confined space inventory at a reclamation site. (C) 2012 Elsevier Ltd. All rights reserved.