Seasonal cycles of O3 in the marine boundary layer: Observation and model simulation comparisons

We present a two-step approach for quantitatively comparing modeled and measured seasonal cycles of O-3: (1) fitting sine functions to monthly averaged measurements and model results (i.e., deriving a Fourier series expansion of these results) and (2) comparing the phase and amplitude of the statistically significant terms between the models and measurements. Two and only two sine terms are sufficient to quantify the O-3 seasonal cycle in the marine boundary layer (MBL) in both the measurements and the model results. In addition to the expected fundamental (one sine cycle per year), a second harmonic term (i.e., two sine cycles per year) is identified as a ubiquitous feature of O-3 in the MBL. Three chemistry climate models (Community Atmosphere Model with chemistry, GFDL-CM3, and GISS-E2-R) approximately reproduce many features of the measured seasonal cycles at MBL surface sites throughout the globe, with some notable quantitative disagreements, but give divergent results that do not agree with O-3 sonde measurements above the MBL. This disagreement and divergence of results between models indicate that the treatment of the MBL dynamics in the chemistry-climate models is not adequate to reproduce the isolation of the MBL indicated by the observations. Within the MBL the models more accurately reproduce the second harmonic term than the fundamental term. We attribute the second harmonic term to the second harmonic of opposite phase in the photolysis rate of O-3, while the fundamental term evidently has many influences. The parameters derived from the Fourier series expansion of the measurements are quantitative metrics that can serve as the basis for future model-measurement comparisons.