This manuscript presents a substantial advancement in both methodology and data availability for investigating soil erosion dynamics in fragile ecosystems, with a particular focus on the Loess Plateau. By integrating long-term time-series Landsat and Sentinel imagery with relevant environmental datasets, the authors develop a high-resolution land-cover and soil erosion dataset spanning 25 years, while further shortening the update frequency to a quarterly scale. Such temporal coverage and resolution represent a notable improvement over existing regional product. The reported mapping performance is robust for a large-scale application, with an overall accuracy of 81.44% for land-cover classification and a mean absolute error of 4.5% for soil erosion estimates. Based on an examination of the released dataset, this work is expected to provide a valuable data foundation for future studies on land-surface processes, ecological restoration, and environmental change across the Loess Plateau. Leveraging this self-produced dataset, the authors further analyze the spatial and temporal evolution of land cover and soil erosion, revealing a pronounced overall reduction in erosion intensity over the study period. The long-term and high-frequency observations enable novel insights into the seasonal heterogeneity of erosion driven by precipitation, the role of vegetation dynamics in erosion mitigation, and the influence of topographic factors. In addition, the attempt to assess potential erosion under an optimized vegetation configuration provides practical implications for soil conservation and land management strategies in the region.

The manuscript is clearly written and accompanied by high-quality data visualizations. Overall, this study makes a valuable contribution in terms of both data products and analytical perspectives, and I believe it is suitable for publication after revisions. The detailed comments are as follows:

1. The land-cover dataset developed in this study is directly applied to soil erosion estimation, while the land-cover products referenced in the Introduction and used for accuracy comparison are mostly general-purpose datasets with coarse thematic categories and broad spatial coverage. Given the strong dependence of soil erosion modeling on land-cover characteristics, it would be helpful to clarify whether any land-cover datasets specifically designed for soil erosion assessment have been developed or applied in the Loess Plateau region. If such erosion-oriented land-cover datasets exist, how do their spatial resolution and mapping accuracy compare with the proposed LP-QLC10 dataset?
2. The land-cover mapping in this study is based on eight categories. Please clarify the rationale for selecting these specific categories. In particular, it would be helpful to explain how these categories are relevant to soil erosion research and whether they are designed to effectively represent erosion-related surface characteristics.
3. The imagery used for land-cover mapping is derived from multiple satellite sources (e.g., Landsat and Sentinel) and forms dense temporal sequences. To improve the transparency and usability of the dataset, it is recommended to provide a concise visualization of the image acquisition timeline, such as stacked bar charts or timeline histograms showing the temporal distribution of input images. Displaying acquisition dates in this way would help users better understand data availability, temporal coverage, and seasonal sampling characteristics, thereby facilitating more effective use of the dataset.
4. Given the long-term and high-frequency nature of the proposed land-cover and soil erosion dataset, the manuscript would benefit from a more explicit discussion on how these results could be translated into policy-relevant applications. In particular, it is recommended that the authors clarify how the dataset and the identified erosion patterns could support practical decision-making at the policy level, such as evaluating the effectiveness of existing soil conservation measures, and guiding ecological restoration planning in the Loess Plateau.

General comments:

This manuscript synthesized Sentinal-2 and Landsat series high resolution data and applied an advanced framework, cross-temporal consistency-constraint learning, to map land cover (LC) time series 2000-2024 over China's Loess Plateau region. In this framework, encoder is applied on the time series to obtain consistency-constrained time series (CCTS), which enhanced the temporal coherence and semantic stability of land-cover representations, and first build segmentation model accounting for inter-temporal consistency in feature levels, then trained segmented features under the supervision of labeled datasets, and finally perturbed CCTS accounting for different noise in the original dataset to perform intra-temporal refinement over the classification and produce the final LC time series. After analysis of the land use change over the past two decades, authors applied the LC product to estimate the average annual soil loss per unit area and assess the water-induced soil erosion over the whole Loess Plateau area. Overall, this is solid work introducing a new fine-resolution dataset with comprehensive and in-depth application case study, which matches the scope of ESSD. There are several concerns from my side, which suggest further improvement before the manuscript can be accepted for publication.

Specific comments:

1. As the major scope of ESSD, work shall focus on your data product first. The strategy using CCTS is claimed to bring more consistency, which is the biggest highlight in your data product, but lacks quantitative evaluation in the whole manuscript. I suggest authors quantitatively evaluate the improvement of spatial and temporal consistency in LC mapping with/without building CCTS.
2. Grassland, cropland and barren land all have low PA, which will bring substantial uncertainty to your conclusions. Please consider the possible uncertainties that might change the conclusion of your analysis. 

Technical corrections:

Please indicate in the title that the soil erosion discussed in this manuscript is specifically water-induced erosion.

Line 202: How exactly did you perturb the image to consider intra-temporal consistency? 

Line 286: It seems like the meaningful spatial resolution is only 30m for 2000-2015, then your title claiming the data product to be a 10-m resolution is misleading?

Line 308: Why does alpha have two different values for warm and cold periods separately? Is it because of snow fall contributing almost no weathering or just the weathering rate reduced under low temperature?

Line 309: Is snowfall included or excluded?

Line 317: What's the reason for re-calibrating Chinese soil? It seems to be a linear bias correction formula that due to some systematic bias? I assume soil in China has no difference from other regions when represented by texture (sand, silt, clay, OC and bedrock)

Line 323: Cropland has a strong seasonal signal. But here in your equation you mentioned A is an annual value. Are you using the same NDVI and parameters for different seasons? Please clarify.

Line 391: Under fine resolution such as 10m, the regular 8 LC types might not be enough for mosaic landscape cases, e.g., the green space in built-up regions, such as parks, that can be classified as either forest or cropland. I would suggest authors to discuss this point.

Line 391: Are both validation LC products having the same LC type definition and follow the same criteria to map LC?

Line 399: "By contrast, GLC_FCS30D often misinterprets water bodies as wetland due to surrounding vegetation, and further misclassifies substantial grassland areas as cropland or bare land" Do you have reference to prove this conclusion? Is nearby vegetation the reason for misclassification?

Line 458: "In particular, 8.6% of the region was transformed from bare land to grassland". In table 5, Barren land PA is not high and mostly confounded by grassland. This also jeopardizes your further conclusion, for example, "the transformation from bare land to grassland produces the most erosion reduction" in line 527. Please discuss and provide uncertainty.