Review of “The impact of Aeolus observations on wind and rainfall predictions”

I. General Assessment

This manuscript presents a comprehensive evaluation of the impact of Aeolus Doppler wind lidar observations (using the reprocessed Baseline 16 dataset) on global wind and precipitation forecasts. The authors have utilized the ECMWF Integrated Forecasting System over an impressively long evaluation period of more than three years. This long-term perspective is, in my view, the standout feature of the study, as it provides a level of statistical confidence often missing from shorter-term campaign evaluations.

The overall quality of the work is high. The experimental design is robust, and the improvements in wind forecasts—especially in the upper troposphere—are presented clearly and align well with our current understanding of the Aeolus mission's capabilities. While the paper is logically structured and the figures are informative, I believe the discussion regarding precipitation could be strengthened. My comments primarily focus on the interpretation of these results and some technical inconsistencies that need to be ironed out. I recommend publication after the following minor revisions are addressed.

II. Major Comments

1. Physical Mechanisms vs. Correlation: A recurring theme is the attribution of precipitation skill improvements to a better representation of large-scale dynamics (e.g., jet streams or Rossby waves). While this is a plausible physical explanation, the analysis remains essentially correlational. Since the causal chain isn't explicitly demonstrated through diagnostics, I suggest adopting a slightly more cautious tone, noting where physical causality is inferred based on spatial/temporal overlap.

2. Significance and Relevance: The improvements in SEEPS and FSS, while statistically significant, are quite small in absolute terms (often <1%). I recommend a more nuanced distinction between statistical significance and practical meteorological relevance. Additionally, acknowledging the scale-dependence of FSS would help ground the results.

3. The 12-hour Accumulation Window: Given that Aeolus provides "snapshots" in time, I wonder if the 12-hour precipitation accumulation window used for verification might be masking more immediate impacts on the convective cycle. Did the authors consider shorter windows (e.g., 6h) to see if the impact is more pronounced closer to the satellite overpass times?

4. Southern Hemisphere Signal: The pronounced wintertime signal in the Southern Hemisphere is a compelling finding but feels under-explored. Expanding the discussion to consider factors like baroclinic activity or the relative scarcity of other observations in that region would significantly enhance the scientific value of the paper.

5. Reference Dataset Transparency: Using ERA5 is standard, but it is worth explicitly stating for transparency that ERA5 is not a fully independent reference, as it shares the same underlying model physics as the forecasts being tested.

III. Minor and Technical Points

1. Interpretation of SEEPS (Figure 8): In Section 2.2.1, you define the change in SEEPS as -(EXP - CTRL) so that positive values indicate improvement. I appreciate this effort to make the results more intuitive. However, since SEEPS is traditionally negatively oriented, this "reversed" convention must be explicitly restated in the caption of Figure 8 to prevent confusion.

2. The "Tropical Disconnect": The fact that substantial wind improvements in the tropics don't translate into much precipitation skill is fascinating. Could this be a limitation of the verification products, or perhaps a result of how parameterized convection handles the corrected wind field? A brief comment on this would be very welcome.

3. "Entire Troposphere" Claim: In the abstract and conclusions, you mention improvements "across the entire troposphere." However, Figure 2 shows that in the lower troposphere at mid-latitudes, the impact is often negligible. I suggest a more precise phrasing.

4. Precipitation Regimes: Using "dry" vs "wet" as a proxy for lidar conditions is a useful approximation, but please clarify that "dry" surface conditions do not always equate to cloud-free profiles.

5. Terminology & Typos:

• Please unify the terms "forecast day" and "lead time".
• Page 8 typo: “the impact of Aeolus is may be more pronounced.”
• Figure 7 / Section 3.3: The abbreviation for “Observed Heavy, Forecast Dry” jumps between OHDF and OHFD. Please unify this.
• Section 3.2: Rephrase “above all cloud layers” to something like “above most deep cloud layers.”
• Figure 6 Caption: Please ensure it explicitly states whether values are absolute or relative.

The paper "The impact of Aeolus observations on wind and rainfall predictions" by Borne et al. is a comprehensive investigation of the impact of Aeolus satellite wind observations on wind and rainfall predictions by the global ECMWF forecasting system. Main findings are that wind forecasts are particularly improved in the tropics and in the Southern Hemisphere. Rain forecasts are particularly improved in the extended winter season at midlatitudes of the respective hemisphere. Main reason is that during this season precipitation is strongly correlated with the wind systems. Different from this, rainfall in the tropics is more randomly linked to convective events which are more difficult to predict.

Overall the paper is very well written and fits into the scope of WCD. The paper is therefore recommended for publication in WCD after addressing my minor comments as detailed below.

My main comment is that the authors should elaborate a bit more about the suitability of ERA5 as a reference for the winds.

Further, some additional discussion would help the reader to digest that improvements in the percent-range may appear small, but are quite relevant for weather foreasting.

SPECIFIC COMMENTS:

(1) l.82: Please state more clearly that both the Aeolus Rayleigh and Mie products were used for the Aeolus OSE.

(2) l.116: Here you should elaborate a bit on the suitability of ERA5 as a reference. Particularly for the tropics where the geostropic constraint breaks down it has been shown that ERA5 zonal-mean zonal wind seems to agree better with Aeolus than other reanalyses, and also the signatures of tropical waves seem to agree well between ERA5 and Aeolus (Ern et al., ACP, 2023). When comparing with tropical radiosonde stations ERA5 and MERRA2 seem to perform similarly well.

Further, a comparison of geostrophic winds derived from radio occultations and ERA5 shows good agreement (Nimac et al., AMT, 2025).

Ern, M., Diallo, M. A., Khordakova, D., Krisch, I., Preusse, P., Reitebuch, O., Ungermann, J., and Riese, M.:

The quasi-biennial oscillation (QBO) and global-scale tropical waves in Aeolus wind observations, radiosonde data, and reanalyses,

Atmos. Chem. Phys., 23, 9549-9583, https://doi.org/10.5194/acp-23-9549-2023, 2023. 

Nimac, I., Danzer, J., and Kirchengast, G.:

The added value and potential of long-term radio occultation data for climatological wind field monitoring,

Atmos. Meas. Tech., 18, 265-286, https://doi.org/10.5194/amt-18-265-2025, 2025.

(3) l.202: Please add reference Martin et al., QJRMS, 2023a to this list.

(4) l.247: Please mention here that the coverage by radiosondes in the Southern Hemisphere is quite sparse. For illustration you should add the reference Durre et al. (2018).

Durre, I., Yin, X., Vose, R. S., Applequist, S., and Arnfield, J.:

Enhancing the data coverage in the Integrated Global Radiosonde Archive,

J. Atmos. Ocean. Techn., 35, 1753-1770, https://doi.org/10.1175/JTECH-D-17-0223.1, 2018.

TECHNICAL COMMENTS:

l.44: Wind  -> wind

l.58: please resolve abbreviation DWL

l.367: more ful  -> more skillful 

l.427: data availability statement for Aeolus is missing

l.493: delete: "A journal of the atmospheric sciences, applied meteorology and physical oceanography"