The subject of this paper is of real importance and deserves to be published. A high quality work is done. But, it needs to be revised to have a greater impact on readers.
The core of the paper focuses on the generalisation of the approach of using SBS for sampling spatially aggregated ecological populations. However, as this mechanism is already well documented in the scientific literature, focusing the paper on this issue makes it lose scientific importance and relevance. Also the main scope of Methods in ecology and evolution is to promote new methods, and this paper is not a new method. This is the first time that we demonstrated the mechanism behind the impact of the distance between sampling points on the accuracy of sampling protocols. The paper should focus on this mechanism.
One of the main problem I encountered when reviewing this paper, is the R scripts were not provided. I could not check the used population estimator nor variance estimator (which are not detailed in the paper). I was also not able to review the functions programmed by the authors, and to see which SBS they used (this is not detailed too).

Introduction:
A more precise literature about the comparisons between SBS and SRS/SYS is needed in introduction, highlighting their results. The superiority of SBS when the population is aggregated is already well known, this should appear in your introduction. Also, McGarvey et al 2016 and Magnussen and Fehrmann 2019 are not the only papers on SBS, new estimators development…
L66 – 71 – putting equations in the introduction is awkward. They should be put in the method section. Especially since, only the theory of estimators for the protocol in simple random is defined, not the one for SYS and SBS.

Method :
Since the method you used is almost identical to the one used in Kermorvant et al (2020), you should at least cite it and compare the methods.
You need to define the theory for SYS and SBS population estimators and variance. Inclusion probabilities of sampling units are not the same that SRS’s ones.
You do not present which SBS sampling design you have chosen to use. I guess it was BAS (because it is one of the keywords). Why did you chose it? It is known that it is not the most appropriate and that it can have strong weakness – see Robertson and al (2017 and 2018).
For reproducibility, you must make your R scripts available.
L-188 : Is the index of dispersion bounded ? What are the minimum and maximum values ?

Discussion :
In page 19, you stated: « The SYS and SBS methods did not reach their highest relative precision at the point where the mean distance between the sampling units was equal to the median cluster diameter, as was the state for the virtual populations. », you omitted to discuss about it in the appropriate discussion part.

ROBERTSON, B. L., MCDONALD, T., PRICE, C. J., et al. A modification of balanced acceptance sampling. Statistics & Probability Letters, 2017, vol. 129, p. 107-112.

ROBERTSON, Blair, MCDONALD, Trent, PRICE, Chris, et al. Halton iterative partitioning: spatially balanced sampling via partitioning. Environmental and Ecological Statistics, 2018, vol. 25, no 3, p. 305-323.

KERMORVANT, Claire, COUBE, Sébastien, D’AMICO, Frank, et al. Sequential process to choose efficient sampling design based on partial prior information data and simulations. Spatial Statistics, 2020, vol. 38, p. 100439.