Correlated observation errors in data assimilation

TLDR

Data assimilation combines observations and prior model forecasts to generate NWP initial conditions, with observation weighting determined by error estimates; remote‑sensing errors are often correlated but these correlations are usually ignored. The study aims to describe three approaches for treating observation error correlations. Using an idealized dataset, the authors compare the information content of each simplified assumption with that obtained under the correct correlation specification. Ignoring correlations leads to a significant loss of information, whereas retaining an approximate correlation yields clear benefits. © 2007 John Wiley & Sons, Ltd.

Abstract

Abstract Data assimilation provides techniques for combining observations and prior model forecasts to create initial conditions for numerical weather prediction (NWP). The relative weighting assigned to each observation in the analysis is determined by its associated error. Remote sensing data usually has correlated errors, but the correlations are typically ignored in NWP. Here, we describe three approaches to the treatment of observation error correlations. For an idealized data set, the information content under each simplified assumption is compared with that under correct correlation specification. Treating the errors as uncorrelated results in a significant loss of information. However, retention of an approximated correlation gives clear benefits. Copyright © 2007 John Wiley & Sons, Ltd.