The manuscript by Xie et al. “Landsat-based Irrigation Dataset (LANID): 30-m resolution maps of irrigation distribution, frequency, and change for the U.S.” developed new irrigation mapping datasets in the US for both cropland and pasture with high spatial resolution across a relatively long-time span (1997-2017). The irrigation mapping showed high accuracy compared with validation, and other multiyear results revealed interesting regional and local patterns in irrigation changes. This work will be an important contribution to the community. The manuscript is well-written and the presentation is clear.

Below I have some minor comments:

L65. There are several different irrigation products used in this study. For readers who are am not familiar with each of them like me, it is helpful to add a column in the table to briefly summarize the method for producing each dataset. 

L125: What years are those selected reference and validation points?

L264: The authors need to discuss the spatial scale issue when comparing different datasets (points vs. pixel, spatial resolutions). How do different spatial resolutions influence the comparison among different data sources? What is a fair comparison? For example, how to make a fair comparison between irrigation fraction and the binary irrigation map?

L296 and Fig 14b: Unlike other products which were only compared with one year of NASS data, LANID was compared with multiple years of NASS data. I think this may contribute to the higher R2 of LANID in Fig. 14b. In terms of R2, LANID is better, but I don’t understand why it is written that the LANID agreement is weaker than MirAD and GMIA.

This is a well-written paper describing the methodology used to produced CONUS-wide, 30 m resolution maps of annual irrigation status. This paper effectively presents results from analysis showing the time series of irrigated area state-by-state, irrigated change at the county and state level, and maps change spatially over the CONUS. Figures are clear and easy to read.

Abstract is appropriately specific and clearly states the need for this product, general methodological approach, and utility of the produced data. 

-- The link to data seems to be the data published with the RSE manuscript, and appears not to contain maps showing the new LANID_V2 data mapping irrigated hay and pasture. Should this be updated, or another repository offered for the examination of the new data described in this manuscript?

Introduction is well-written and provides a good summary of why irrigation  is important, and the impacts and benefits of irrigation. Literature review is appropriately specific and comprehensive. Table 1 is comprehensive and appropriate.

Methods section is clear and concise, and describes a sound approach  given the challenge of detecting irrigation from Landsat images. The doption of two different approaches for detectionin the humid East and semi-arid West US, while adding complexity, is justified given the low contrast between irrigated and non-irrigated lands in the East.

Map evaluation and comparison designs seems to choose appropriate, previously produced maps for comparison to LANID.

Figure 4 is epecially attractive. Should Figure 4 take into account uncertainty estimates?

Figures 5, 6, and 7 are informative and well done.

Irrigated pasture and hay: where is this data in the repo referred to in the abstract?

Maximum extent, frequency, and formerly irrigated and intermittent irrigated land: interesting findings. Line 259: what is meant by 'energy sorghum'? Table 3 is interesting and informative.

Figures 11 - 13 clearly display the improvement in mapping resolution and accuracy over previous maps.

The comparison of irrigated area maps is clear and offers cogent explanations of why differences in the maps exist, in terms of differences in irrigated lands' definition and classification methodology.

The discussion of uncertainty, limitations, improvements and potential applications is appropriate.