Balance properties of the short‐range forecast errors in the ECMWF 4D‐Var ensemble

116

A new method for generating ensemble predictions based on an ensemble of data assimilations has been developed. Using an ensemble of four‐dimensional ensemble‐variational minimizations provides an approach which is close to the Met Office's operational data assimilation system and less computationally expensive than other alternatives. In developing this system, several inflation schemes have been compared. One form of additive inflation, based on analysis increments, was developed and found to be very effective at increasing the overall ensemble spread and correcting systematic biases in the model. However, the analysis increments are not flow‐dependent since they are randomly drawn from a long archive. It was decided to scale back their amplitude to avoid them dominating the overall performance. Of the other inflation schemes considered, it was found that relaxation‐to‐prior‐perturbations was the most effective at maintaining the ensemble spread. However, this scheme also produced perturbations which are too large‐scale and too balanced. The relaxation‐to‐prior‐spread scheme performed well in many respects, but required a relaxation factor greater than one to produce an acceptable spread. Therefore these two schemes were combined in order to mitigate the drawbacks of each. This combination proved successful and was used in final testing of the ensemble against the currently operational ensemble transform Kalman filter (ETKF). The ETKF has its perturbations centred around a high‐resolution ‘deterministic’ analysis. This was seen to be an important benefit, and the new ensemble system also benefited from being recentred around the high‐resolution analysis. This recentred system has slightly lower forecast skill than the ETKF over a variety of variables, due to the fact that the spread of this ensemble is less than the spread of the ETKF ensemble. The deficiency of the spread of the new ensemble system will be addressed in ongoing work.

Section: Comparison Of Rtpp and Rtpsmentioning

confidence: 99%

Inflation and localization tests in the development of an ensemble of 4D‐ensemble variational assimilations

Bowler

Clayton

Jardak

et al. 2017

116

“…The diagonal matrix D defines the spectral background‐error variance densities associated with eigenmodes. The background‐error variance spectral densities are computed from the statistics of the estimated short‐range forecast errors of the ECMWF model using the ensemble approach (Žagar et al ., ). In the applied dataset, the greatest variance, about 45%, is associated with the equatorial Rossby modes.…”

Section: Modeling Aerosols In Maddammentioning

confidence: 97%

Inferring atmospheric dynamics from aerosol observations in 4D‐Var

Zaplotnik

Žagar

Benedetti

et al. 2020

This article explores the potential of aerosol observations to provide wind information in four-dimensional variational data assimilation (4D-Var). It is shown that the relative horizontal gradients, crucial for wind extraction from tracers, are on average greater for the aerosol mixing ratio than for the specific humidity, observations of which are known to provide significant information on the wind field. The potential of aerosols to infer atmospheric dynamics is investigated in the Tropics, where the wind information is most critical. An intermediate-complexity incremental 4D-Var system, the Moist Atmosphere Dynamics Data Assimilation Model (MADDAM), has been developed, with a forecast model that simulates most dominant processes involving moisture, aerosols, and dynamics: nonlinear advection, condensation, and wet deposition. The results of 4D-Var experiments reveal a detrimental impact of saturation-related nonlinearities and aerosol wet deposition on the extraction of wind from aerosol data. Fraternal-twin experiments show about 30%smaller impact of aerosol data on the wind analysis compared with humidity data, mainly due to the greater aerosol observation error and suboptimal background-error covariance model. However, the assimilation of aerosol data together with temperature and humidity observations shows significant added value for wind analyses. K E Y W O R D Saerosols, background-error model, data assimilation, humidity, tracing, wind extraction, 4D-Var

“…The applied spectral background‐error variances for dynamical variables are taken from the estimated short‐range forecast errors of the European Centre for Medium‐Range Weather Forecasts (ECMWF) model using the ensemble approach (Žagar et al. , ). This provides a realistic distribution of the background‐error variances among various equatorial waves.…”

Section: Moist Atmosphere Dynamics and Data Assimilation Modelmentioning

confidence: 99%

An intermediate‐complexity model for four‐dimensional variational data assimilation including moist processes

Zaplotnik

Žagar

Gustafsson

2018

This article presents a new Moist Atmosphere Dynamics Data Assimilation Model (MADDAM), an intermediate-complexity system for four-dimensional variational (4D-Var) data assimilation. The prognostic model equations simulate nonlinear moisture advection, precipitation, and the impact of condensational heating on circulation. The 4D-Var assimilation applies the incremental approach and uses transformed relative humidity as a control variable. In contrast to the model dynamical variables, which are analyzed in multivariate fashion using equatorial wave theory, moisture data are assimilated univariately. MADDAM is applied to study the extraction of wind information from time series of moisture observations in the Tropics, where the lack of wind information is most critical. Results show that wind tracing in the unsaturated atmosphere depends largely on the ability of the assimilation model to resolve spatial gradients in the moisture field, which is determined by the spatial density and accuracy of observations. In the saturated atmosphere, a combined assimilation of moisture and temperature data is shown to improve wind analyses significantly, as the intensity of the condensation process is susceptible to the slightest changes in saturation humidity and thus temperature. Moreover, a perfect-model 4D-Var with moisture observations can extract wind information even in precipitating regions and strongly nonlinear flow, provided sufficient observations of humidity gradients are available. MADDAM is envisaged to serve as a testbed for new developments in 4D-Var assimilation, with a focus on interactions between moist processes and dynamics across many scales. KEYWORDShumidity control variable, moist 4D-Var, moisture observations, tropical data assimilation, wind tracing INTRODUCTIONSimplified numerical prediction models are valuable testbeds for data assimilation research. An idealized framework, with respect to numerical weather prediction (NWP) models, aids understanding and simplifies research. Simplified models allow us to (a) develop and implement new algorithms faster than in the NWP case, and (b) perform numerical experiments in a controlled environment in which various issues, difficult to grasp in a real NWP, can be understood more easily. At the same time, simplified models should still be complex enough to capture the main dynamical and physical aspects of the phenomena of interest, in order to explain the observed features of the circulation and to be of any value for NWP. An important model of this type is based on the rotating nonlinear shallow-water equations (for example, Vallis, 2006), which include both balanced (vorticity-dominated) dynamics and gravity-wave dynamics as well as their interactions. Shallow-water models (SWMs) were applied in a number of data assimilation studies to develop new concepts and to study the value of mass-field and wind-field observations (for example, Žagar et al., 2004a, and references therein). This question is especially important in the Tropics, where wind-field informati...