A location engine that generates an estimate of the measurement bias of the pressure sensor in a wireless terminal. The estimate is determined from a probability distribution characterized as having a flat top with exponentially decaying sides. The probability distribution is generated by inferring i) a range of elevations within which the wireless terminal is most likely present, wherein the elevations have associated barometric pressures, and ii) a corresponding range of measurement biases based on an inferred range of the barometric pressures. The location engine updates the probability distribution over time. The location engine makes a single estimate of measurement bias available by determining a mean of the updated probability distribution, such as an average of the low and high endpoints that define the flat top. Subsequently, the location engine generates an estimate of the elevation of the wireless terminal by accounting for the estimate of measurement bias.