Performance of a local ensemble transform Kalman filter for the analysis of atmospheric circulation and distribution of long-lived tracers under idealized conditions

Abstract

A data assimilation system for the analysis of atmospheric circulation and long-lived tracer distributions in the troposphere and stratosphere has been developed and tested by using a local ensemble transform Kalman filter (LETKF), which has been applied to assimilate both meteorological fields and tracer concentrations into a general circulation model and an atmospheric transport model. Assimilated meteorological fields are used for driving the transport model. The performance of the LETKF data assimilation system is assessed under idealized conditions by assuming that the forecast models provide a perfect representation of atmospheric conditions. The LETKF meteorological analysis facilitates the study of atmospheric transport characteristics and provides high-quality tracer transport simulations, reflecting its flow-dependent and physically well balanced analysis. In particular, eddy mixing features are better analyzed by LETKF than by an analysis that employs a conventional assimilation scheme (i.e., nudging technique). The conventional analysis causes excessive tracer eddy dispersions, which were commonly observed in previous studies using three-dimensional variational analyses. Further improvement in tracer analysis can be achieved by assimilating the tracer concentration within the LETKF. The assimilation of tracer concentration effectively reduces the tracer background error caused by initial distribution and surface flux errors. Tracer analysis can also be improved by considering the covariance with wind fields in a background error matrix, in which wind observation directly impacts the tracer states, reducing 20% of the tracer analysis error in the free troposphere. The sensitivity of the tracer analysis to assimilation parameters and model errors is discussed to obtain an optimal data assimilation framework. © 2009 bv the American Geophysical Union.