Improvement of numerical weather prediction model analysis during fog conditions through the assimilation of ground-based microwave radiometer observations: a 1D-Var study

Abstract

Abstract. This paper investigates the potential benefit of ground-based microwave radiometers (MWRs) to improve the initial state (analysis) of current numerical weather prediction (NWP) systems during fog conditions. To this end, temperature, humidity and liquid water path (LWP) retrievals have been performed by directly assimilating brightness temperatures using a one-dimensional variational technique (1D-Var). This study focuses on a fog-dedicated field-experiment performed over winter 2016–2017 in France. In situ measurements from a 120 m tower and radiosoundings are used to assess the improvement brought by the 1D-Var analysis to the background. A sensitivity study demonstrates the importance of the cross-correlations between temperature and specific humidity in the background-error-covariance matrix as well as the bias correction applied on MWR raw measurements. With the optimal 1D-Var configuration, root-mean-square errors smaller than 1.5 K (respectively 0.8 K) for temperature and 1 g kg−1 (respectively 0.5 g kg−1) for humidity are obtained up to 6 km altitude (respectively within the fog layer up to 250 m). A thin radiative fog case study has shown that the assimilation of MWR observations was able to correct large temperature errors of the AROME (Application of Research to Operations at MEsoscale) model as well as vertical and temporal errors observed in the fog life cycle. A statistical evaluation through the whole period has demonstrated that the largest impact when assimilating MWR observations is obtained on the temperature and LWP fields, while it is neutral to slightly positive for the specific humidity. Most of the temperature improvement is observed during false alarms when the AROME forecasts tend to significantly overestimate the temperature cooling. During missed fog profiles, 1D-Var analyses were found to increase the atmospheric stability within the first 100 m above the surface compared to the initial background profile. Concerning the LWP, the RMSE with respect to MWR statistical regressions is decreased from 101 g m−2 in the background to 27 g m−2 in the 1D-Var analysis. These encouraging results led to the deployment of eight MWRs during the international SOFOG3D (SOuth FOGs 3D experiment for fog processes study) experiment conducted by Météo-France.

Improving thermodynamic profile retrievals from microwave radiometers by including radio acoustic sounding system (RASS) observations

“…In this study, the focus is on the combination of the MWR and RASS observations in the retrievals to improve the accuracy of the temperature profiles in the lowest 3 km compared to physical retrieval approaches that do not include the information from RASS measurements. Some studies have used analyses from numerical weather prediction (NWP) models as an additional constraint in these variational retrievals (e.g., Hewison, 2007;Cimini et al, 2006Cimini et al, , 2011Martinet et al, 2020); however, we have elected not to include model data in this study because we wanted to evaluate the impact of the RASS profiles on the retrievals from a purely observational perspective.…”

W-band radar observations for fog forecast improvement: an analysis of model and forward operator errors

“…Low cloud whose base is descending is very likely to be observed before an instance of cloud base lowering (CBL) fog. A Vaisala CL-31 ceilometer (Martucci et al, 2010) was used to measure the cloud base height. This uses a pulse lidar to sense the cloud base and is capable of sensing up to three layers simultaneously with a range from 0 to 7.6 km.…”

Towards the use of conservative thermodynamic variables in data assimilation: preliminary results using ground-based microwave radiometer measurements

“…In 1D-Var systems, the integrated liquid water content, Liquid Water Path (LW P ) can be included in the control vector x as initially proposed by Deblonde and English (2003) and more recently used by Martinet et al (2020). A first experimental set-up has been defined where the minimization is performed with the control vector being (T, q v , LW P ).…”

“…This is particularly the case during fog conditions with much shorter vertical correlation length-scales at the lowest levels and large positive cross-correlations between temperature and specific humidity (Ménétrier and Montmerle, 2011). In this context, Martinet et al (2020) have demonstrated that humidity retrievals could be significantly degraded if sub-optimal background error covariances are used during the minimization. New ensemble approaches allow a better approximation of background error covariance matrices but rely on the capability of the ensemble data assimilation to correctly represent model errors, which might not always be the case during fog conditions.…”