Overshoot Bias and the National Ozone Standard

ABSTRACT

The new ozone standard continues to require rounding ozone design values to the nearest 10 ppb. A rationale for the rounding convention has been that the design value for the ozone standard is biased upwards. It has been argued that rounding compensates for this overshoot bias. This paper investigates the degree of overshoot bias in the design values for the old and new standards under various assumptions about measurement error and autocorrelation. It was found that there can be substantial overshoot bias in the design value of older standard, but much less for the new standard's design value. Moreover, the total error in the new design value is on the order of 1-3% for reasonable assumptions about measurement error. This implies that rounding may tend to misclassify nonattainment areas as attainment. The results suggest that the rounding requirement for the new ozone standard may not be justified.