Previous model studies suggested that ambient measurements of key chemical species and ratios of species could be used to assess the sensitivities of ozone formation to reductions in precursor emissions. Threshold values of these indicator species and ratios, delineating the transition between VOC and NOx sensitivity, were proposed. Subsequently, measurement studies have assumed the universality of these threshold criteria and have compared local observed indicators with previously established criteria to assess ozone sensitivities. In this study the concept of indicator species is extended to combinations of observable species that are consistently associated with different site characteristics (e.g., ozone sensitivities and emissions features). The results of SARMAP Air Quality Model (SAQM) simulations in the San Joaquin Valley, California, are used to investigate the applicability of indicators to assess ozone sensitivities and emissions features. The use of three indicator ratios (O3/(NOy-NOx), HCHO/NOy, and H2O2/HNO3) along with SAQM-derived threshold criteria is found to be effective for identifying VOC- or NOx-sensitive regimes. NOy and (NOy-NOx)/NOy are found to be useful in describing emission features and threshold criteria are derived by SAQM prediction. SAQM-derived threshold criteria for assessing ozone sensitivities are found to differ from threshold criteria proposed by previous studies using different models and under different conditions. Such differences suggest that threshold criteria are likely to be dependent on locations and environmental conditions, including emission patterns and rates. Therefore local observed indicator species and ratios can be used to determine ozone sensitivities only if appropriate threshold criteria have been derived for the local conditions. ¿ 1998 American Geophysical Union