The manuscript titled "Assessing evapotranspiration dynamics across central Europe in the context of land-atmosphere drivers" evaluates evapotranspiration (ET) products derived from satellite remote sensing, modeling, and reanalysis data in conjunction with in-situ observations from Integrated Carbon Observation System (ICOS) stations in central Europe from 2017 to 2020. The study investigates the effects of varying land cover types, soil moisture (SM), and vapor pressure deficit (VPD) on ET dynamics, including the severe drought of 2018. It uses extended triple collocation methods to assess the accuracy of ET products, revealing notable differences among products under heterogeneous land cover conditions and during drought years. The findings highlight that ET variability is strongly influenced by VPD under non-limiting soil moisture conditions and demonstrate that SEVIRI, ERA5-land, and GLEAM products show superior performance. The research provides insights into the suitability of various ET products for capturing land-atmosphere interactions and drought impacts across diverse land cover types. This manuscript presents valuable insights into evapotranspiration dynamics across central Europe, but significant revisions are needed before it can be considered for publication.

Major comments:

1. Limitations in the Performance Analysis of ET Products

The manuscript reveals significant performance differences among ET products, but it does not fully address how these discrepancies might affect the conclusions drawn, particularly under specific land cover types or climatic conditions. A deeper analysis of how these differences impact the study's practical implications is crucial.

1. In-depth Analysis of Drought Year Effects

The analysis of the 2018 drought is valuable, but the study does not sufficiently explore how different ET products capture the effects of drought. A more detailed comparison of how products perform under drought conditions is needed, including which models and parameterizations are more sensitive to such extremes.

1. Improvement of Remote Sensing and Ground Observation Matching

There is an acknowledged mismatch between point-scale ICOS data and the coarser-resolution remote sensing products. The manuscript could benefit from a discussion on methods to address this issue, such as spatial downscaling or data fusion techniques, to improve the alignment of ground and remote sensing observations.

1. Considering extending the study period

The study period appears to be too short, and the comparison between satellite-based ET and local measurements seems insufficient. The authors might consider extending the study period and incorporating a comparison across different timeframes. This would enable the use of the water balance principle for a more comprehensive evaluation.

Minor comments

1. At lines 34-35: please add explanation (why).
2. At line 34: please add the specific number to describe “the highest potential dependencies (error cross-correlation)”.
3. At lines 36-38, please explain why.
4. Please add the longitude and latitude in Figure 1.
5. Please add a table that describes all the sites, including their longitude, latitude, land cover type, altitude, and other relevant details.
6. At line 150: Please explain why you used the standardized precipitation-evapotranspiration index (SPEI).
7. If the time step is from hourly to 8-day, consider generated them at a shorter time step rather than monthly which loose too much information.
8. At lines 206-225: The spatial resolution differences among these ET products are quite significant. The authors should consider using methods to standardize all the ET products to a common resolution or use products with longer time periods. Otherwise, direct comparisons may not be valid or meaningful.
9. Table 1: please change as three lines table format and added other dataset’s information such as soil moisture and SPET etc.
10. At lines: 228-233 The removal of the seasonal signal may not be necessary, as the study period is too short for such adjustments.
11. At lines 295-296: Please provide the specific values for the highest R² and lowest RMSE and lowest percentage bias, PBIAS here.
12. Figure 4. Please added some statices numbers in every panel.
13. Same as Figure 4

Review Report for Manuscript ID: egusphere-2024-3386

Title: Assessing Evapotranspiration Dynamics Across Central Europe in the Context of Land-Atmosphere Drivers

General Comments

This study provides a comprehensive evaluation of evapotranspiration (ET) products across central Europe using a combination of in-situ, remote sensing, and reanalysis datasets. The authors analyzed the performance of multiple ET datasets under different climatic conditions, particularly focusing on the severe drought. The study effectively addresses a research gap by assessing the agreement and discrepancies between ET products in the context of soil moisture (SM) and vapor pressure deficit (VPD) interactions. The manuscript is relevant for researchers studying land-atmosphere interactions, hydrology, and ecosystem responses to climate extremes. However, several key areas require further clarification and improvement to strengthen the manuscript before publication. The authors should clarify the rationale for dataset selection, improve the discussion on physical interpretability, and provide additional insights into the role of vegetation stress and uncertainty quantification.

Major Comments

Justification for Selected ET Products, the authors compare ET estimates from various remote sensing and modeling products (MODIS, SEVIRI, GLEAM, ERA5-land, GLDAS). However, it would be beneficial to explicitly justify the selection of these specific products over other alternatives such as FLUXCOM, ETMonitor, EB-ET. MODIS is kind of ET more relying on optical data. GLEAM is based on microwave data. Please check other thermal ET product. One could be EB ET. Chen et al. 2021, Remote sensing of global daily evapotranspiration based on a surface energy balance method and reanalysis data. Journal of Geophysical Research: Atmospheres, 126(16): e2020JD032873.

Additionally, the manuscript should discuss the potential biases associated with the retrieval algorithms used in each dataset and how these may affect ET estimates under different climatic conditions.

Physical Interpretability and Model Dependencies, the study provides robust statistical comparisons but lacks a deeper discussion on the physical implications of the observed differences. For example, why do some products perform better at evergreen needle-leaved sites compared to agricultural sites? How do land cover heterogeneity and seasonal changes influence model uncertainties?

Since GLEAM incorporates reanalysis and satellite-based observations, its correlation with other datasets like ERA5-land and GLDAS-2 might be inflated. Have the authors accounted for interdependencies between products in their error analysis?

The role of vegetation stress and physiological controls (e.g., stomatal closure) in driving ET reductions during drought should be better discussed, perhaps using additional to support this point.

Evaluation of Uncertainty and Error Cross-Correlation (ECC), The extended triple collocation (ETC) analysis is a valuable approach, but some ECC values are quite high, particularly at agricultural sites. The manuscript should explicitly discuss how ECC influences the reliability of the results and whether certain datasets may be inherently dependent.

Clarity of Figures and Statistical Significance, the scatter plots and time series comparisons are informative, but additional clarity is needed in figures showing product inter-comparisons (e.g., Figures 4, 5, 6). Including a statistical significance test for differences between ET products would enhance the rigor of the results.

Minor Comments

Grammar and Style: Some sentences are long and complex, making them difficult to follow. Consider simplifying and improving readability. For example: The ICOS network has undertaken a large effort to ensure high-quality LE measurements, which are comparable among different ICOS stations.”. Suggested revision: “The ICOS network has made significant efforts to ensure consistent high-quality LE measurements across stations.”. Line 555 Grammar mistake, This is, products were most consistent with each other at stations with less complex land cover conditions and changes throughout the seasons (the evergreen needle-leaved stations DE-Ruw and FI-Let).

Line 49, rephrase the sentence ‘Since precipitation (P) ‘and evaporation are the two key components of the global water cycle’ (Miralles et al., 2011), another important proxy for analyzing water stress and its effects on ecosystems is evapotranspiration (ET).’ .

Line80, optical, thermal, infrared, or microwave observations are used to derive ET based on surface energy balance, physical and empirical models (Bayat et al., 2021, 2024; Rahmati et al., 2020; Zhang et al., 2016). The cited reference does not include thermal observation based ET from surface energy balance method.

Terminology Consistency: The terms "ET estimation," "ET retrieval," and "ET modeling" are used interchangeably. It would be beneficial to define them more precisely and use consistent terminology throughout the manuscript.

Temporal Aggregation Effects: Some ET product has a lower temporal resolution than other datasets. Have the authors checked whether this affects the observed discrepancies, if upscaled to 15 days, even monthly temporal resolution?

Line 558, The authors wrote that: The remote sensing products, SEVIRI, MODIS, and GLEAM, performed equivalently well or even better than the in-situ measured (ICOS), I don`t understand why the remote sensing products can be better than in-situ measured data? This is confusing readers. How can a satellite ET product be better than measurement?

Line 514, ET is more controlled by atmospheric demand rather than atmospheric supply, I can understand when the atmosphere is warming, it will need more vapor evaporated from ground. This could be a kind of atmospheric demand, but do not understand what is atmospheric supply? What kind of supply from atmosphere can influence ET? Are you saying precipitation? Please rephrase this sentence to make it clear.

Line 523, Further, results show that VPD and SM are negatively coupled during extreme events as reported also by (Zhou et al., 2019)-à by Zhou et al. 2019. Same as reported by (De Santis et al., 2022).

Section 4, there are many other global ET product, which are not discussed. Please check and compare them.