The viability of trajectory analysis for diagnosing dynamical and chemical influences on ozone concentrations in the UTLS. Issue 11 (5th June 2017)

• Record Type: Journal Article
• Authors: Bergman, J. W.; Pfister, L.; Kinnison, D. E.; Hintsa, E. J.; Thornberry, T. D.
• Is Part Of: Journal of geophysical research. Volume 122:Issue 11(2017)
• Page Start: 6025
• Page End: 6042
• Publication Date: 2017-06-05
• Subjects: chemistry and dynamics -- upper troposphere/lower stratosphere -- ozone -- Lagrangian model -- aircraft observations; Atmospheric physics -- Periodicals; Geophysics -- Periodicals; 551.5
• DOI: 10.1002/2017JD026487
• Languages: English
• ISSNs: 2169-897X

Abstract

Abstract: To evaluate the utility of trajectory analysis in the tropical upper troposphere/lower stratosphere, Lagrangian predictions of ozone mixing ratio are compared to observations from the Airborne Tropical TRopopause EXperiment. Model predictions are based on backward trajectories that are initiated along flight tracks. Ozone mixing ratios from analysis data interpolated onto "source locations" (at trajectory termini) provide initial conditions for chemical production models that are integrated forward in time along parcel trajectories. Model sensitivities are derived from ensembles of predictions using two sets of dynamical forcing fields, four sets of source ozone mixing ratios, three trajectory formulations (adiabatic, diabatic, and kinematic), and two chemical production models. Direct comparisons of analysis ozone mixing ratios to observations have large random errors that are reduced by averaging over 75 min (~800 km) long flight tracks. These averaged values have systematic errors that motivate a similarly systematic adjustment to source ozone mixing ratios. Sensitivity experiments reveal a prediction error minimum in parameter space and, thus, a consistent diagnostic picture: The best predictions utilize the source ozone adjustment and a chemical production model derived from Whole Atmosphere Community Climate Model (a chemistry‐climate model) chemistry. There seems to be slight advantages to using ERA‐Interim winds compared to Modern‐Era Retrospective Analysis for Research and Applications and to using kinematic trajectories compared to diabatic; however, both diabatic and kinematic formulations are clearly preferable to adiabatic trajectories. For these predictions, correlations with observations typically decrease as model error is reduced and, thus, fail as a model comparison metric. Plain Language Summary: To evaluate the utility of trajectory analysis in the tropical upper troposphere/lower stratosphere, predictions of ozone mixing ratio are compared to observations from the Airborne Tropical TRopopause EXperiment. Model predictions are based on backward trajectories that are initiated along flight tracks. Ozone mixing ratios from analysis data interpolated onto "source locations" (at trajectory termini) provide initial conditions for chemical production models that are integrated forward in time along parcel trajectories. Model sensitivities are derived from ensembles of predictions using two sets of dynamical forcing fields, four sets of source ozone mixing ratios, three trajectory formulations, and two chemical production models. Direct comparisons of analysis ozone mixing ratios to observations have large random errors that are reduced by averaging over 75 min (~800 km) long flight tracks. These averaged values have systematic errors that motivate a similarly systematic adjustment to source ozone mixing ratios. Sensitivity experiments reveal a prediction error minimum in parameter space and, thus, a consistent diagnostic picture: The best predictions utilize the source ozone adjustment and chemical production derived from National Center for Atmospheric Researchs Whole Atmosphere Community Climate Model. For these predictions, correlations with observations typically decrease as model error is reduced and, thus, fail as a model comparison metric. Key Points: Lagrangian predictions of ozone mixing ratios are viable diagnostic tools provided that ensemble averaging and sensitivity testing are employed Ozone mixing ratios from global models and reanalysis data have removable systematic errors Sensitivity calculations identify the most realistic region of model parameter space and reveal subtle dynamical relationships