General comments:



This ms addresses an important question, which is whether the next generation of gravity missions (NGGM/MAGIC) will deliver sufficient accuracy and temporal resolution to enable precipitation estimation via the SM2RAIN algorithm. The ms contributes to applying SM2RAIN for the first time, using TWS as input rather than soil moisture. However, several methodological choices require better justification before potential publication in HESS. The use of ERA5L as both the input proxy and the reference benchmark creates insufficient independence between the forcing and evaluation, which is not adequately acknowledged. The synthetic error framework needs stronger physical grounding, and the discussion of regional performance disparities (Alps, Baltic coast, Scandinavia) remains shallow. Furthermore, the ms could be improved by considewreing a deeper mechanistic analysis of failure modes, clearer separation of the validation and synthetic experiment frameworks, and more rigorous attention to equifinality in parameter calibration.



Specific comments:



1.    Around line 20: The study uses ERA5L soil moisture layers as a TWS proxy and ERA5L precipitation as the benchmark for validation. This impacts the independence of the evaluation framework, as mentioned above, a limitation that is not adequately discussed in the framework, where the algorithm is essentially optimized and tested against the same underlying data model. The authors must explicitly acknowledge this limitation and discuss how it may inflate performance metrics. An independent validation using, e.g., GPCC or E-OBS precipitation data, could be considered, even if only for a subset of stations or a sub-period.

2.    Around line 85: The third stated objective refers to "adding Gaussian error on simulated precipitation estimates," but the Gaussian errors are actually added to the TWS input, not to the precipitation estimates. Please, revise this to avoid confusion about the experimental design.

3.    Around line 115: The TWS estimate (Eq. 1) is computed exclusively from the four soil moisture layers of ERA5L and does not account for snow water equivalent, groundwater, surface water, or canopy interception, all of which are components of the full GRACE-measured TWS signal. The authors should either redefine their variable explicitly as "soil water storage" (SWS) or justify why this simplification is appropriate for the European domain and the 2003–2012 period. This distinction has direct implications for interpreting how faithfully ERA5L-TWS proxies GRACE-based TWS.

4.    Around lines 135: The SM2RAIN parameters (a, b, Z*) are calibrated point-by-point against ERA5L precipitation. No discussion is provided about parameter sensitivity, equifinality, or the spatial coherence of calibrated parameters across Europe. A supplementary figure showing the spatial distribution of calibrated parameters would greatly strengthen the methodology and allow readers to assess whether unrealistic parameter values emerge in certain regions (e.g., the Alps, Nordic regions).

5.    Around line 155: The synthetic errors (1.9, 4.2, 19, 42 mm) are introduced as additive Gaussian noise drawn from the NGGM mission error table (Daras et al., 2023). However, real mission errors are not necessarily Gaussian in space and time since they may be correlated, anisotropic, and orographically structured. The authors should discuss the limitations of the Gaussian error assumption and whether correlated or spatially structured noise could materially change the conclusions.

6.    Around line 120: The bilinear interpolation from ERA5L’s native 0.1° resolution to 100 km is described briefly but without justification. How does this resampling affect the TWS signal and the precipitation estimates, particularly in areas with complex topography such as the Alps or the Scandinavian mountains? The authors should verify that no significant smoothing bias is introduced by this step. Furthermore, a realistic GRACE-TWS signal would require a spherical harmonic representation up to a degree and order corresponding to NGGM.

7.    Around line 175–185: The ms identifies systematic underperformance in the Alpine region of northern Italy, in Estonia and Latvia, and along the Norwegian coast, but does not provide a mechanistic explanation. Are these failures related to orographic precipitation, snow processes, coastal sea-level effects on TWS, or inherent limitations of SM2RAIN? Please consider exploring this issue, and a dedicated paragraph examining the physical drivers of regional failure would improve the scientific depth of the results section.

8.    Around line 200: In Section 3.1, ERA5L precipitation serves as the reference; in Section 3.2, the SM2RAIN-derived output from Section 3.1 is used as the reference for the synthetic experiments. This might imply that any systematic errors in the Section 3.1 output are inherited and compounded in Section 3.2. The authors should discuss how errors in the first step propagate into the second and explicitly state this design choice.

9.    Section 3.2 appears twice, first for "Performance of NGGM synthetic TWS for precipitation estimation" (p. 7) and again for "Performance of NGGM\MAGIC mission with respect to GRACE-FO" (p. 10). This should be corrected throughout.

10.    Around line 275: The authors define R > 0.7 as the threshold for "satisfactory" performance, but this threshold is neither justified in the text nor referenced to any established community standard. Given that the application context is precipitation estimation for hydrological use, the justification should be tied to downstream impacts (e.g., flood forecasting, drought detection), or a sensitivity analysis should show how conclusions change under alternative thresholds such as 0.6 or 0.8.

11.    Around line 280: The statement that the GRACE-FO error is 25 mm and its temporal resolution is 30 days is used as the baseline for comparison. Please add a citation or derivation that supports these specific values in the main text. The source and spatial scale at which this 25 mm figure applies must be explicitly cited and contextualized, as the GRACE-FO error itself might be resolution-dependent.

12.    Section 4, Discussion and Conclusion: This section reads as a summary of the results without engaging with the literature on gravity-based hydrological applications or competing approaches. The authors should discuss how their findings offer any complementary advantage over SM2RAIN-SM, which is briefly mentioned but deserves more elaboration.