Can an ensemble transform Kalman filter predict the reduction in forecast-error variance produced by targeted observations?

The ensemble transform Kalman filter (ET KF) is currently used at the Natio nal Centers for Environmental Prediction (NCEP) to identify deployments of aircraft-borne dropwindsondes that are likely to significantly improve 1-3 day forecasts of winter storms over the continental United States. It is un ique among existing targeted observing strategies in that it attempts to pr edict the reduction in forecast-error variance associated with each deploym ent of targeted observations. To achieve this, the ET KF predicts the varia nce of 'signals' for each feasible deployment, where a signal represents th e difference between two forecasts, initialized with and without the target ed observations. For linear forecast-error evolution, the signal variance i s equal to the reduction in forecast-error variance, provided that observat ion- and background-error covariances are accurately specified and identica l to those produced by the operational data-assimilation scheme. However, b ackground-error covariances assumed by the ET KF are both imperfect and dif ferent from the imperfect error covariances used in NCEP's 3D-Var data-assi milation scheme, and hence their signal statistics are likely to differ. In spite of these differences, a linear relationship of positive gradient i s found to exist between the ET KF signal variance and the sample variance of NCEP signal realizations at both the targeted analysis and forecast veri fication times, for 30 forecasts from the 2000 Winter Storm Reconnaissance Program. This relationship enables the NCEP signal variance to be predicted by the ET Ki via a statistical resealing that corrects the ET KF's current over-prediction of signal variance magnitude. A monotonically increasing r elationship is also found to exist between the NCEP signal variance and the reduction in NCEP forecast-error variance. The ET KF signal variance predi ctions can be used to make quantitative estimates of the forecast-error-var iance reducing effect of targeted observations, Potential benefits include (i) making rapid decisions on when and where to deploy targeted observation s, (ii) warning operational data quality-control schemes against the reject ion of observational data if the signal variance is large, and (iii) estima ting the likelihood of economic benefit due to any future deployment of obs ervations.