Interactive comment on “ A coupled remote sensing and the Surface Energy Balance with Topography Algorithm ( SEBTA ) to estimate actual evapotranspiration under complex terrain ”

4877-line 7 What you mean by “limited” in “limited temporal and spatial scales”? 4877-line 9 What do you mean by complex terrain? Where the complexity of terrain addressed in the formulations given in the subsequent sections of the paper? 4877-line 14 Does this paper gives “Full account” for the dynamic impacts of complex terrain and changing land cover with some varying kinetic parameters over time? I guess it is better to use another appropriate word instead of “full” as still there are still unknown issues about effects of terrain on ET, especially in regard to advection and atmospheric turbulence. 4877-line 15 The phrase “in concert” replaced with more appropriate word.

one station is used to compare the flux estimated at different dates to observations.It's not enough, especially as to test a new algorithm including a new information on the elevation.It would be necessary to compare at least two stations at different elevations or with different slope or aspect values.Sometimes, the paper does not give enough information on the data used and the accuracy linked.For some figures or in tables, it is important to add the standard deviation or the confidence interval.Globally the paper is interesting with a lot of references, but it must be corrected in adding information at specific points listed below.It can be published if modifications are made to improve the text.

Specific comments:
Title: Maybe the title can be modified because the expression 'complex terrain' can include a lot of different surfaces and not necessary the relief.Reference Nagler (in reference list it misses 'r 'to Nagler p4907 2005b) P4880 line 14: the sentence ' the residual models are the best . . .' is a little bit exaggerated, because SVAT models using remote sensing data with assimilation methods can be also very efficient.Otherwise, residuals models don't represent the soil moisture evolution which is important for water management.

P4880
Line 25: As SEBTA is not yet described in detail, it's embarrassing to write that it's the best at this place!I find that the most important point is the improvement with the topography information.The two other points concerning roughness estimation and automatic calculation for separating wet and dry pixels have been yet performed in other studies with similar models.

P4883
Line 10: the G formation is a critical point (very questionable here) because several papers have shown bad results with such formula (add references and discussion on this point).The coefficients were defined for some surfaces, (it's a very empirical approach, not validated over various surfaces).
-Have you measurements to chose G=0.5 Rn for water surfaces?C1650 P4884 Line1: It can be difficult to find wet and dry pixels according to the dates and the region studied.The spatial resolution of images used is also important to take into account.
( How many stations?I suppose that you have air temperature measurements on these stations which could be also used for validation. be careful: daily evaporation is not negligible for some surfaces (bare soil. . . ) Line 13: soil moisture is not only driven by precipitation (irrigation. . .can modify also the soil moisture) P4879 line 6: reference Kogan (remove et al) Ref Jian (add and Islam) C1649

Figure 5 :
Figure 5: some days present precipitation events, that means clouds (have you estimated surface flux for these dates?) how do you explain that you have bad estimations when wheat is harvested?You start with 48 dates but on the graph fig 6 some dates miss why.The coefficients can be arguable.(fig 6 is not necessary, Values can be given in the text or in fig 5 caption)

Table 1
does not give a lot of information, see refJetse et al, SurV Geophys 2008The estimation of z0 in the standard version of SEBAL can be arguable, but other authors have used this model in estimating z0 with more accurate methods.I would not add SEBTA in this table 1 as it is not yet presented.Moreover, the argument of its use with any time period seems to me wrong since as for SEBAL or S-SEBI, it is based on the use of remote sensing data acquired in the optical range, therefore only for clear days.The temporal scales are the same than other models.

Table 3
table with the regions classified according to the main classes for elevation and landuse classes would be welcome to follow the analysis.P4898.The conclusion could be more nuanced.(discussionontheseason,theimagenumber,the rain days. . . ) C1653P4898 lines23-28: put in a table all these values given in the text and add a column in table4 with standard deviation P4900 line 14: the introduction of LULC data (give the information before in the text).This approach is not new.Other models use also these data (give refs)Line 12: not really shown because there is not validation on stations with different elevation factors.The paper only show results on simulations and discussed the impact of topography on simulation results.These variations were not validated.P4903 line 1: 48 images were used but when there are rainy days or clouds, there is no discussion to fill the gap.In the table 3, only 15 dates are presented (add some justifications)Validation for flux based only for one station (representative ?) so the conclusion must be reviewed and nuanced.What is your consistency index?Meaning?Line 14: 'indispensable is strong! the model proposed here, can be useful and can be compared to other models...Interactive comment on Hydrol.Earth Syst.Sci.Discuss., 7, 4875, 2010.
A C1654