News

Application of a GNSS ARO Observation Operator for Data Assimilation During a Sequence of Atmospheric Rivers

23-09-2026

A new scientific publication entitled “Assimilating GNSS Airborne Radio Occultation Bending Angles With MPAS–JEDI During a Sequence of Atmospheric Rivers” has been published as a result of international scientific collaboration, with Dr Paweł Hordyniec from the Institute of Geodesy and Geoinformatics at UPWr, Poland among its co-authors. The study presents the application of GNSS Airborne Radio Occultation (ARO) observations for atmospheric data assimilation.

The study integrates a modified ARO observation operator with the Joint Effort for Data assimilation Integration (JEDI) system and the MPAS-A atmospheric model. ARO observations collected during the Atmospheric River Reconnaissance campaigns were used to investigate a sequence of atmospheric river events over the northeastern Pacific.

An important component of the study was the observation operator for airborne radio occultations, previously adapted by Dr Paweł Hordyniec. The operator accounts for the specific asymmetric geometry of ARO measurements and allows for the presence of horizontal gradients in atmospheric parameters. Its implementation within the JEDI framework enabled ARO observations to be incorporated into data assimilation experiments alongside other types of atmospheric observations.

The experiments indicate improvements in the analysed temperature, humidity, and wind fields, as well as beneficial changes in forecasts of integrated water vapour transport (IVT) and precipitation, particularly at longer forecast lead times. The authors also highlight the potential of ARO observations in regions where conventional atmospheric observations are limited.



Spatial integrated vapor transport (IVT) and RMSE analysis at 72 hr lead time for values ≥250. (a, d) Control forecast IVT magnitude (shades) compared against ERA5 (contours). (b, e) Relative RMSE difference (%) between ARO and Control experiments; blue shading indicates error reduction (improvement) in ARO experiment, while red indicates error increase (degradation) relative to Control experiment. The dashed green line indicates the transect latitude. (c, f) Longitudinal IVT profiles comparing the intensity and core structure of ERA5 (black), Control (red), and ARO (blue). Values inside parentheses denote the mean RMSE calculated specifically within the AR boundaries.

The publication provides another example of the application of methods and tools developed by researchers at the Institute in the field of GNSS-based atmospheric remote sensing and the assimilation of observations into numerical weather prediction models.

Hernández Baños, I., Haase, J. S., Hordyniec, P., Cao, B., Liu, Z., Do, P.-N. (2026). Assimilating GNSS Airborne Radio Occultation Bending Angles With MPAS–JEDI During a Sequence of Atmospheric Rivers. Journal of Advances in Modeling Earth Systems, 18(9), DOI: 10.1029/2025MS005432.