Performance of a Portable FT-NIR MEMS Spectrometer to Predict Soil Features
Round 1
Reviewer 1 Report
The work is of the greatest relevance in soil science. It brings new knowledge in the expeditious determination of soil physical-chemical parameters, with high-resolution field equipment.
Ambitious work in the number of samples treated (182), but modest in the number of variables analyzed. It would be interesting to mention the potential of the equipment in the detection of other variables in the soil, such as soil contamination.
The paper is well written and structured, that denotes a well planned and elaborated research work, with a correct treatment and discussion of results.
The least positive aspect of the work is the quality of the figures, which should be improved, in particular:
- the axes of the graphs in figures 3 and 4 are illegible,
- the aerial views of figure 1 need to be better identified and, the sampling locations deserve a better definition.
Author Response
Thanks for your positive comments and for the requests of improvements, which are clear and relevant. The soil samples used for this study were analysed with traditional laboratory methods only for the variables used for this work, therefore we cannot test the prediction of other variables with this dataset. However, to improve the paper according to your comments, we introduce the concept of the suitability of these spectrometers to investigate other parameters, like contaminants, in the final part of the discussion, reporting two very recent publications about the use of spectroscopy to investigate microplastic in soil.
We have tried to improve the figures to make them more readible, increasing the font size, and the figure 1. In paricular, for Fig.1 we improved the aerial image quality and the size of the sampling dots. We also decide to avoid different colours for the areas, making the image more "cleaning".
Thanks again for your review
Reviewer 2 Report
The work is interesting, it deals with a topic of relevance for soil sciences, in particular for precision agriculture, considering innovative instrumentation. However, there are parts to be reviewed.
In my opinion the numbers can be reported without the decimal separator "," which can mislead the reader.
All figures must be replaced with images of better resolution and definition.
It would be appropriate to insert milestones to justify the sentences in lines 26-32.
Check the sum of the samples at lines 120-122; 80+65+27+11= 183 and not 182.
Check lines 125-126.
According to the authors, can the difference between the two white references affect the results of the acquired data? Justify this issue in the text.
Why for NS you chose to make 3 replicates while for AF 5? Clarify this as well.
Do you think NS can be used to determine other soil characteristics, such as contaminants?
Author Response
Thank you for your comments, very clear and relevant. Following, you can see the replies to your questions:
In my opinion the numbers can be reported without the decimal separator "," which can mislead the reader.
REPLY: Ok, no problem to delete the thousands separator “,”
All figures must be replaced with images of better resolution and definition.
REPLY: We have improved the fig.1 using only white colour and increasing the size of the dots, and fig.3 incresing the font size.
It would be appropriate to insert milestones to justify the sentences in lines 26-32.
REPLY: We included three important publications aimed on soil spatial variability and precision farming, to support our assumptions.
Check the sum of the samples at lines 120-122; 80+65+27+11= 183 and not 182.
REPLY: Yes, thank you to highlight this mistake. Pignola site had 26 samples. We modify the text.
Check lines 125-126.
REPLY: Thank you again, we removed the two lines remained from a previous sentence.
According to the authors, can the difference between the two white references affect the results of the acquired data? Justify this issue in the text.
REPLY: we added in the chapter 2.2 the sentences: “Although the NS used a different WR and no ISS standardization method [30], it was not relevant for the aim of this study. Indeed, this work wanted to test the efficiency of NS to predict soil features, but did not aim to create a standardized spectral library, also because NS has different spectral range and resolution of full Vis-NIR spectrometers.”
In the future, we will try to standardize the methodology to make the NS spectra comparable with the spectra collected with other NS spectrometers and Full Vis-NIR spectrometer.
Why for NS you chose to make 3 replicates while for AF 5? Clarify this as well.
REPLY: Because the instruments were used by different people of the paper research group. The people who used Fieldspec strictly followed the Ben-Dor et al. protocols and white standard, because such Fieldspec spectra will be used for a wider library. The people who used Neospectra wanted to use the sensor as easier as possible and without particular standardization, because the aim of the work was to test the performance of an easier and lower cost method. We thought that 3 replicates of spectra was a good compromise between speed and good representativeness of the sample. We have explained this choice in the chapter 2.2: “The decision of 3 replicates of spectral scanning was driven by a compromise between fair representativeness of the soil samples and the speed of the analysis.”
Do you think NS can be used to determine other soil characteristics, such as contaminants?
REPLY: Probably yes, in particular organic compounds or contaminants linked to clay or organic matter. The main problem for such kind of research is to find a right number of samples (analysed by traditional methods) to test the use of NS spectrometer. I hope to have the possibility to make such kind of study. In the discussion chapter, I added a sentence regarding this possibility, also reporting two very recent publication on the study of microplastics in soil by spectroscopy.
Thanks again for your professional review!
