Assessing the potential of free-tropospheric water vapour isotopologue satellite observations for improving the analyses of convective events

Satellite-based observations of free-tropospheric water vapour isotopologue ratios (HDO / H2O, expressed in form of the delta value delta D) with good global and temporal coverage have become available recently. We investigate the potential of these observations for constraining the uncertainties of the atmospheric analyses fields of specific humidity (q), temperature (T), and delta D and of variables that capture important properties of the atmospheric water cycle, namely the vertical velocity (omega), the latent heating rate (Q2), and the precipitation rate (Prcp). Our focus is on the impact of the delta D observations relative to the impact achieved by the observation of q and T, which are much more easily observed by satellites and are routinely in use for atmospheric analyses. For our investigations we use an Observing System Simulation Experiment; i.e. we simulate the satellite observations of q, T, and delta D with known uncertainties and coverage (e.g. observations are not available for cloudy conditions, i.e. at locations where the atmosphere is vertically unstable). Then we use the simulated observations within a Kalman-filter-based assimilation framework in order to evaluate their potential for improving the quality of atmospheric analyses. The study is made for low latitudes (30 degrees S to 30 degrees N) and for 40 d between mid-July and the end of August 2016. We find that q observations generally have the largest impacts on the analyses' quality and that T observations have stronger impacts overall than delta D observations. We show that there is no significant impact of delta D observations for stable atmospheric conditions; however, for very unstable conditions, the impact of delta D observations is significant and even slightly stronger than the respective impact of T observations. Very unstable conditions are rare but are related to extreme events (e.g. storms, flooding); i.e. the delta D observations significantly impact the analyses' quality of the events that have the largest societal consequences. The fact that no satellite observations are available at the location and time of the unstable conditions indicates a remote impact of delta D observations that are available elsewhere. Concerning real-world applications, we conclude that the situation of delta D satellite observations is very promising but that further improving the model's linkage between convective processes and the larger-scale delta D fields might be needed for optimizing the assimilation impact of real-world delta D observations.