Reviewer comments are in Green writing. Reviewer 2 2.1 General Comments: This study has explored the applicability of the TSEB model for Evapotranspiration (ET) over a mountainous area in Morocco. Although authors have put efforts to collect the in situ data sets for this study but, they are unable to design and implement this study in a right way. Thank you for admitting that the gathering of in situ data represented a large amount of work. We obviously disagree with the others reviewer’s general comments. Agricultural landscape in south Mediterranean areas are complex because of the limited size of the fields, the numerous varieties of crops and the variability of the agricultural practices which results in a large variability of crop cycles and the heterogeneity of the crop hydric status owing to precipitation and, above all irrigation. To our knowledge, there is no accepted and unique methodology to map evapotranspiration over these complex terrains. Considering available remote sensing products from the sensors into orbit, the question of the best tradeoff between complex high-resolution based technic and simpler approach with coarse resolution data arises. To contribute to answer to this question: (1) an experimental setup has been designed. It is mainly composed of a scintillometer for integrated convective fluxes measurements and two eddy covariance tower for heterogeneity sampling; (2) a modeling tool based on an extensively used energy budget model (TSEB) feeded by either LANDSAT high resolution input or MODIS coarse resolution products was run on the study area and evaluated thanks to the in situ data base. Thanks for your detailed reply. I would like to clarify whenever talking about “Landsat” one has to mention the specific sensor name (i.e. MSS, TM, ETM+, OLI) and similar is the case for “MODIS” there are two sensors onboard Terra and Aqua. Throughout the manuscript, authors have not mentioned whether they have used data from MODIS Terra, MODIS Aqua or from both? “…..(TSEB) feeded by either LANDSAT high resolution input or MODIS coarse resolution products” does authors think there is no difference on the TSEB model output when they are feeding data from Landsat or MODIS? Don’t they think there will be a difference in the output when satellite data is from different resolution? Could the reviewer give us some details on what would be the right design and implementation of this study? Well, if you are asking me for the “right design and implementation of the study” I would recommend you to perform a literature review. A REVIEWER is supposed to be an early reader who evaluates a study before it is available for general public, do authors assume that a reader should suggest a design and implementation of the study? It is the authors’ responsibility to make whole manuscript concise. Please show your respect to your peers. 2.2 The study is full of superficial statements Could the reviewer please detail what are the statements that he considers superficial? For authors reference following are some superficial statements from the manuscript - Page 3 lines 111 and 112, “The scale of our study is thus twofold: field and multi-field.” - Whole section 2.2 and 2.3, can authors point out what’s their contributions in these sections? - Page 9 line 282, Note that all Landsat thermal data provided by the USGS were resampled to a resolution of 30 m. why data was resampled to 30 m? - Page 9 lines 284 and 285, However, surface temperature cannot be directly derived from thermal measurements. Cannot understand what authors are trying to explain! - Page 9 line 294 to 296, 𝜀𝑣 and 𝜀𝑠 are the vegetation and the soil emissivity being 0.99 and 0.96 respectively in our. 𝑁𝐷𝑉𝐼𝑆 and 𝑁𝐷𝑉𝐼𝑣 are the minimum and the maximum values of NDVI associated with bare soil and dense vegetation, respectively. The values 0.17 and 0.99 are used here. Why these specific values were selected?

In addition, if he doesn’t mind, we would also be interested in his definition of a superficial review. According to the Cambridge Dictionary one definition of “superficial” is “not complete and involving only the most obvious things” and a “superficial review” will be meaning the same! Hope you have got it. 2.3 write up is not attractive for a reader Considering the write up, we agree with the reviewer that section of the paper may be improved and according to the reviewer comment, parts of the manuscript have been rewritten (see colored version of the new version of the manuscript). We hope that the reviewer will be able to understand better the content of the study with these rewritings (see the question at point 2.10). I have not received the colored version of the manuscript, so, unable to find the changes. 2.4 I have been through the manuscript and think that there are no novel aspects in this study OK. Here what we consider as novel in this paper: - The study area composed of small fields with a strong variability of crops, practices and hydric status as already discussed above. - The application of the scintillometry technic to complex terrain. We agree that it was not sufficiently described in the previous version of the manuscript. A new paragraph has been added discussing the height equivalent used to compute the sensible heat flux (see Line 142-151). Also, the derivation of sensible heat flux (H) from the LAS is based on the Monin Obukhov Similarity Theory (MOST), which requires horizontal homogeneity. Therefore, the challenge herein is to investigate whether this theory still applies under conditions of horizontal heterogeneity and variation in topography along the LAS transect.
- The multi-scale evaluation of the TSEB model thanks to the scintillometer and the eddy covariance in such complex landscape - The comparison of different aggregation technics to feed the TSEB model. Maybe I am wrong, but according to my understanding the novelty of a study lies in the methodology of the study. Can authors pinpoint what part of the methodology they have innovated? Its not an innovation to select a heterogenous study area, collect in situ data or compare different techniques, which is merely the case in this study. 2.5 Methodology is not solid Ok see response to 2.1 and 2.2 2.6 and overall the manuscript is not well organized. OK. The paper is organized in a very classical way: introduction, materials and method, results and discussion, conclusion. Could the reviewer tell us what is wrong with this schedule? we think that the new version is well organized. 2.7 Therefore, I would not recommend its publication in Remote Sensing. OK. See response above. We feel that this reviewer's recommendation is exaggerated if we compare with the other 3 reviewers recommendation. 2.8 1- the title is very long and hard to understand OK. Agreed. The title has been shortened according to the reviewer’s comment. The new title is: “Multi-scale evaluation of the TSEB model over a complex agricultural landscape in Morocco”. 2.9 2- lines 31 to 42 are not related to abstract. Abstract needs to be comprehensive OK. The lines 31 to 42 in the old version of the manuscript are as follows: “The complexity is associated with the type of the vegetation canopy as well as the changes in topography. For validating purpose, a large aperture scintillometer (LAS) was set up over a heterogeneous transect of about 1.4 km to measure sensible (H) and latent heat (LE) fluxes. Additionally, two towers of eddy covariance (EC) systems were installed along the LAS transect. In the vicinity of these towers, measurements of soil and vegetation as well as the meteorological parameters were continuously collected. First, the model was tested at the local scale against the EC measurements using multi-scale remote sensing (MODIS and LANDSAT) inputs at the satellite overpasses. While the input variables (surface temperature, albedo and emissivity) derived from LANDSAT-7/8, were aggregated based on the EC footprint, TSEB was run with the same variables at 1 km derived from MODIS. In overall, the use of the LANDSAT data showed better agreement between estimated and measured latent heat fluxes than with MODIS data.” We don’t understand the reviewer comment as the section pointed out by the reviewer described (1) the specificity of the study region which is one of the reasons that makes the study novel and (2) the methodology. According to the reviewer’s comments and also in response to reviewer 1, the abstract has been shortened. 2.10 3- Why the NDVI product from MODIS was selected? I am unable to understand what type of data was exactly used from Landsat 7 and 8? As the objective of the paper is to assess the accuracy of evapotranspiration maps over complex agricultural landscape, high resolution and medium-scale resolution approaches are compared and the best trade-off between complexity and accuracy is chosen. Accordingly, the same products are used for MODIS (daily) and LANDSAT (16 days): NDVI and land surface temperature.