The Accuracy of Trajectories during ANATEX Calculated Using Dynamic Model Analyses Versus Rawinsonde Observations

Abstract

Aircraft tracer measurements all made within 300 km of the release sites during, the Across North America Experiment (ANATEX) provided 30 separate trials to evaluate the error of back-trajectory calculations. The trajectory calculations used dynamic-model-output meteorological data from the NOAA prognostic Nested Grid Model (NGM) and from comparable gridded meteorological fields from 4-day rawinsonde or 2-day rawinsonde observations. Over all trials, no significant difference was discernable in trajectory accuracy using the different meteorological input data; absolute trajectory error ranged from 20% to 30% of the travel distance. When the trajectories were grouped into similar categories, the NGM data provided the but results when smaller-scale flow features were present; ones that could not be resolved in the rawinsonde network. Four-day rawinsonde data provided the best results when fronts or low pressure systems were present in the vicinity of the tracer release. Slower transport in more homogeneous zonal flow regimes resulted in some of the smallest errors of around 15% of the travel distance for all the methods. In 70% of the events, the 4-day and 2-day rawinsonde calculations were comparable, suggesting that linear temporal interpolation of the flow field is reasonable for most situations. The primary conclusion is that the NGM-generated meteorological data can be an adequate substitute for rawinsonde data in trajectory calculations.