Next Article in Journal
A PolSAR Change Detection Index Based on Neighborhood Information for Flood Mapping
Previous Article in Journal
A Simulation Environment for Validation and Verification of Real Time Hyperspectral Processing Algorithms on-Board a UAV
Open AccessArticle
Peer-Review Record

Remotely Sensed Water Limitation in Vegetation: Insights from an Experiment with Unmanned Aerial Vehicles (UAVs)

Remote Sens. 2019, 11(16), 1853; https://doi.org/10.3390/rs11161853
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Reviewer 3: Anonymous
Remote Sens. 2019, 11(16), 1853; https://doi.org/10.3390/rs11161853
Received: 3 July 2019 / Revised: 30 July 2019 / Accepted: 6 August 2019 / Published: 9 August 2019
(This article belongs to the Section Remote Sensing in Agriculture and Vegetation)

Round 1

Reviewer 1 Report

Dear Authors,

Your manuscript entitled “Remotely sensed water limitation in vegetation: insights from an experiment with unmanned aerial vehicles (UAVs)” is interesting, it presents how existing vegetation indicators, commonly used in plant studies, show changes in water content, as they are correlated with this water value. It is interesting to introduce the context of ecophysiology and remote sensing in the introduction, which confirms the belief that this kind of research as the most right and necessary.

Figure 3 - the R value is given once in the graphs, once it is R ^ 2, which is a completely different measure, therefore the values of R squared do not give such high results.

Accept after minor revision.

Sincerely 

Reviewer


Author Response

Reviewer 1

Your manuscript entitled “Remotely sensed water limitation in vegetation: insights from an experiment with unmanned aerial vehicles (UAVs)” is interesting, it presents how existing vegetation indicators, commonly used in plant studies, show changes in water content, as they are correlated with this water value. It is interesting to introduce the context of ecophysiology and remote sensing in the introduction, which confirms the belief that this kind of research as the most right and necessary.

 

Figure 3 - the R value is given once in the graphs, once it is R ^ 2, which is a completely different measure, therefore the values of R squared do not give such high results.

Response: Good catch. We have updated the caption to be clearer, and it now reads:

“a) “Figure 3. Relationship ((R = spearman’s rho statistic (rs)) between NDVI and water content at: 1pm on Day 1; 1pm on Day 2; and all values; b) Relationship (R = spearman’s rho statistic (rs)) between NDRE and Water Content at: 1pm on Day 1; 1pm on Day 2; and all values; and c) relationship (R-squared from linear regression) between plant water content and mid-day water potential”

 


Author Response File: Author Response.docx

Reviewer 2 Report

Generally, this is a good study, but some of the descriptions of the approach and methods for analysis need some attention.

Introduction:

Page 3: The NDVI is parenthetically described and referenced, but the other index mentioned are not described (e.g. NDRI and GNDVI). This is needed somewhere in the intro.

Line 183 - 187: There are mixed US and Metric units here. Please stick to metric with parenthetical US metrics if needed.

If there is a desire to include thermal imaging in the discussion, it needs to be mentioned in the intro, ad htn all remote sensing should be included. The intro assumes that optical sensing is "remote sensing", while it is not the only way to sense plant moisture. Revise the text to acknowledge you are using just optical RS, and drop the mention of thermal in the discussion.

Methods:

The number of pixels averaged in each treatment/control site needs to be included for each altitude. In particular, it needs to be clear how many samples are in each type so the assessment of the statistical analysis results can be made. It seems you have many more pixels (area) in the control than in the treatments, so this should be remedied for the study.

It is unclear why measurements of the full study area are included. The full study area includes bare ground and other irrelevant types. In the graphs, the data showing "study site" is confusing.

Results:

Table 1 is a result, not a part of the methods.

Since the data in figure 2 are averages of several pixels, the points could show the spread of the data. It would help to better understand the relationship.

(I am not an expert in statistical analysis, but would question if the tests conducted are appropriate for averaged data, or if they include the fact that the measurements are an average of many pixels).


Discussion:

Generally, this is good. The paragraph on thermal imagery is not needed. The study does not include thermal. More appropriate would be to mention microwave - passive or SAR. These are very relevant to the questions of plant and soil moisture.




Author Response

Reviewer 2

Generally, this is a good study, but some of the descriptions of the approach and methods for analysis need some attention.

 

Introduction:

Page 3: The NDVI is parenthetically described and referenced, but the other index mentioned are not described (e.g. NDRI and GNDVI). This is needed somewhere in the intro.

Response:  We have considerably shortened this section according to the requests by other reviewers.

 

Line 183 - 187: There are mixed US and Metric units here. Please stick to metric with parenthetical US metrics if needed.

Response: change completed.

 

If there is a desire to include thermal imaging in the discussion, it needs to be mentioned in the intro, ad htn all remote sensing should be included. The intro assumes that optical sensing is "remote sensing", while it is not the only way to sense plant moisture. Revise the text to acknowledge you are using just optical RS, and drop the mention of thermal in the discussion.

Response: change completed. The word “thermal” has been removed from the introductory paragraph, but retained in the literature paragraph (line 167) as this sentence alludes to what has been done elsewhere.

 

Methods:

The number of pixels averaged in each treatment/control site needs to be included for each altitude. In particular, it needs to be clear how many samples are in each type so the assessment of the statistical analysis results can be made. It seems you have many more pixels (area) in the control than in the treatments, so this should be remedied for the study.

It is unclear why measurements of the full study area are included. The full study area includes bare ground and other irrelevant types. In the graphs, the data showing "study site" is confusing.

Response: We have added Supplementary Table 1 to show the number of pixels at the optimal GSD (altitude). Unequal groups are common and potentially at  risk of violating the homogeneity of variance assumption—the assumption that the population variance of every group is equal, however most non-parametric tests such as the one used here, the Wilcox test can account for differing sample size and are robust when normal assumptions cannot be met. Additionally, the study area average is included on this figure to: 1) provide a reference and 2) because the study area is used later in the paper in the scaling portion of the work.

 

Results:

Table 1 is a result, not a part of the methods.

Response: agreed. Table 1 has been moved to the Results section and now is Table 2. New text has been included:

“Leaf water content samples for all treatments (Tx1, Tx2, Tx3, Tx4, C and W) and pre-dawn and midday leaf water potential for treatments Tx3, Tx4, C and W are shown in Table 2.”

 

Since the data in figure 2 are averages of several pixels, the points could show the spread of the data. It would help to better understand the relationship.

Response: This is a good point, but we would like to keep the figure as is. We show the box blot graph (i.e. the spread of the data, not just averages) in Figure 4. The data in Figure 2 was used to determine best flying altitude, not to provide a statistical inference.

 

(I am not an expert in statistical analysis, but would question if the tests conducted are appropriate for averaged data, or if they include the fact that the measurements are an average of many pixels).

Response:

The Wilcox test is a statistical comparison of the average of two dependent samples and a non parametric version of a t-test that assesses the difference between observations.

 

Discussion:

Generally, this is good. The paragraph on thermal imagery is not needed. The study does not include thermal. More appropriate would be to mention microwave - passive or SAR. These are very relevant to the questions of plant and soil moisture.

Response: We did not include microwave sensors because while important for plant and soil moisture, they are not sensors available on UAVs. In any event, we removed the paragraph on thermal imagery in the Discussion as advised.

 

Reviewer 3

Major concerns:

 

The comparison between NDVI values from the UAV camera and Planet data could be incorrect. I inferred that NDVI values were spatially averaged from the UAV data and then compared to the Planet NDVI values (if this is not the case, then please make it clear in the manuscript). Because NDVI does not scale linearly with reflectance, this type of averaging does not work. For example, consider two pixels with red reflectances [0.1,0.2] and NIR reflectances [0.4,0.25]. If their reflectances are averaged, red = 0.15 and NIR = 0.325. Averaging pixel NDVI values produces an average NDVI of 0.35, but calculating NDVI from the averaged reflectance value produces an NDVI of 0.37. To correctly compare NDVI across spatial resolutions, you should first resample reflectance of the UAV camera data to match the Planet data spatial resolution, and then calculate NDVI.

Response: The reviewer is correct that we needed to clarify this portion of the methods section, and the concerns captured in the above statement were valid. We did however resample each band to match the 3m Planet resolution before calculating NDVI so the values reflected in the manuscript and the figures are correct. We clarified our procedure in the methods as follows:

“We resampled each of the radiometrically calibrated bands from the UAV flight on July 26 2018 at 11:10am (the closest temporal match to the native resolution of PlanetScope Imagery) to 3 m to match the native pixel resolution of PlanetScope imagery. Using the resampled UAV band values we then calculated NDVI. The PlanetScope Ortho Scenes Surface Reflectance product is a 16-bit GeoTIFF with reflectance values scaled by 10,000. We divided the pixel values by 10,000 to compare the reflectance bands and NDVI index of the PlanetScope Imagery with those collected from the UAV [54]. PlanetScope Imagery has 4 bands: blue (455-515 nm), green (500-590 nm), red (590-670 nm), and NIR (780-860 nm).”

 

L339-341: “The resampled UAV NDVI pixel values contained more spectral information, suggesting higher radiometric resolution for the MicaSense RedEdge narrow band sensor in comparison to PlantScope’s sensor.” It’s not clear that more variation in values indicates higher radiometric resolution. It’s also not clear whether “narrow band” is supposed to indicate a difference in spectral resolution. L432 appears to confuse radiometric resolution (number of possible DN/radiance values in a measurement) with spectral resolution. If spectral resolution is really what you’re getting at, I don’t understand why differences in spectral resolution would cause more variation in NDVI values, rather than just a simple bias. A better explanation is needed.

Response:

In response to the reviewer’s comments we altered the text to clarify our point about the differences in the spectral resolution of these sensors. The next text is below. We have added a figure to this response, but did not add it to the paper.

“A narrow band sensor such as the RedEdge can provide more detailed information by capturing a more precise measurement of specific wavelengths. Therefore, the spectral sensitivity of sensor was preserved despite being the spatial resolution being resampled.”

 

Furthermore, we would expect that there would be more variation in NDVI values for the Micasense sensors due to the narrower range in spectral values for each band which discerns more variations in surface reflectance whereas the larger spectral ranges encompassed by Planet’s bands and averaged over a larger range of reflectance values leading less overall variation detected and resulting in a narrower distribution of NDVI values. We have done some additional analysis to support this hypothesis (see figure below)

 

There wasn’t an explanation for why indices vary with spatial resolution (i.e. the first day experiment).

Response: This was an omission. We have added the text:

“Spectral indices vary with spatial resolution for a number of reasons, including the fact that lighting changes throughout the day, as well as leaf moisture.”

 

 L244-245: Reflectance panel wasn’t mentioned earlier. Or is this the white part of the black and white targets? What material was used? Was the irradiance sensor used for correction? UAV NDVI values do seem high for this ecosystem, so it’s important that the calibration procedure is clear.

Response: This information has been added to the Methods:

“Physical radiometric targets were imaged prior to flight for radiometric calibration.”

 

In addition to thermal infrared measurement of stress, solar induced fluorescence should also be mentioned e.g. Zarco-Tejada et al., 2012

Response: In response to a comment from reviewer 2, we have removed the section on thermal sensing. We have added discussion of fluorescence in the final Discussion paragraph:

 

“New work also highlights the ability to measure solar induced fluorescence via narrowband remote sensing. About 1% of solar energy captured by plants is re-emitted by chlorophyll as fluorescence. The amount of fluorescence emission is a direct indicator of the photosynthetic activity of a plant and also provides indirect information about the plant stress [87]. Remote sensing of solar induced fluorescence requires very narrow spectral bands (~0.05 nm) around 690nm and 760nm [80,88].”

 

Minor comments:

 

L5: formatting, “and”

Response: change made.

 

L48: formatting, extra space before “and”

Response:  change made.

 

L50: “manifest as diurnal and seasonal changes” What kind of changes?

Response:  We have changed the text to the following:

“Such water limitation can result in a diversity of plant responses such as reductions in carbon fixation, growth and even reproduction.”

 

L67: “It is common” I think you should limit this to “Using broadband multispectral remote sensing it is common”, because otherwise you’re excluding other types of remote sensing that have used for vegetation applications.

Response: change made.

 

L72: By “greenness” do you mean color? KT transform greenness?

Response: “greenness” changed to “green color”. Line 77.

 

L116: “estimating” is probably too strong here, since there was no validation. This canopy water content was purely a spectral measure, and while changes were described in response to drought, it’s not clear this measure correlates with actual canopy water content.

Response: Good point: “estimating” changed to “evaluating” line 125

 

I think Section 1.2 can be safely reduced, since NDVI (and to a more limited extent, NDRE) should be very familiar to this audience.

Response: Agreed. We have cut this section down considerably.

 

L206: “were” should be “was” to agree with “content”, or change to “water content measurements” to agree with “were”

Response: changed to “water content measurements”

 

L207: “contents” and “samples” don’t agree

Response:  changed to “content”.

 

Check with journal to make sure that date formatting is okay. Using month/day is mostly specific to the US and Canada.

Response: All instances of “7/day” changed to “July day”.

 

Table 1: Without going into the text and lining up the dates, it’s not clear how “Day 1” and “Day 2” align with the cut dates and with the date imagery was flown. This is later resolved in the following section, but you might consider adding date information to the table caption or to the table itself.

Response:  Date added to Table 1.

 

Field sampling mentioned measurement of leaf water content, but later on plant water content is referred to. Make sure terminology is consistent and defined.

Response: We have changed this to “leaf water content” and “leaf water potential” when we are discussing this experiment. Instances where “plant water stress” or “plant water status” are used in the more general sense remain.

 

L309-310: Leaf and plant water content are used in these two sentences. Are they the same thing?

Response: We have changed this to “leaf water content” and “leaf water potential” when we are discussing this experiment. Instances where “plant water stress” or “plant water status” are used in the more general sense remain.

 

L323: Can you add any more detail beyond “likely a mistake”?

Response: Good point. This was poorly worded. We think this was an anomalous data point due to the fact that we sampled immediately following cutting. But in the spirit of being transparent with all of the data we collected we have included it and then flagged it.  We’ve changed the text to read:

“[Note: the data point at -5 MPa and 52% water content is an anomaly and likely due to the timing of the treatment (immediately following cutting)]”

 

L422: Can you really determine optimality based on this one experiment?

Response: Agreed. “optimal” changed to “useful”.

 

L427: “In this paper we show that the derived NDVI values from UAVs and PlanetScope CubeSat imagery across treatments are consistent” Is this really what Figure 5b shows?

Response: The original statement is a bit misleading. We have changed it to: “In this paper we show that the derived NDVI values from UAVs and PlanetScope CubeSat imagery across treatments are largely consistent with important differences, particularly in their sensitivity to subtle changes to water content (control and water).”


Author Response File: Author Response.docx

Reviewer 3 Report

Major concerns:

The comparison between NDVI values from the UAV camera and Planet data could be incorrect. I inferred that NDVI values were spatially averaged from the UAV data and then compared to the Planet NDVI values (if this is not the case, then please make it clear in the manuscript). Because NDVI does not scale linearly with reflectance, this type of averaging does not work. For example, consider two pixels with red reflectances [0.1,0.2] and NIR reflectances [0.4,0.25]. If their reflectances are averaged, red = 0.15 and NIR = 0.325. Averaging pixel NDVI values produces an average NDVI of 0.35, but calculating NDVI from the averaged reflectance value produces an NDVI of 0.37. To correctly compare NDVI across spatial resolutions, you should first resample reflectance of the UAV camera data to match the Planet data spatial resolution, and then calculate NDVI.

 

L339-341: “The resampled UAV NDVI pixel values contained more spectral information, suggesting higher radiometric resolution for the MicaSense RedEdge narrow band sensor in comparison to PlantScope’s sensor.” It’s not clear that more variation in values indicates higher radiometric resolution. It’s also not clear whether “narrow band” is supposed to indicate a difference in spectral resolution. L432 appears to confuse radiometric resolution (number of possible DN/radiance values in a measurement) with spectral resolution. If spectral resolution is really what you’re getting at, I don’t understand why differences in spectral resolution would cause more variation in NDVI values, rather than just a simple bias. A better explanation is needed.

 

There wasn’t an explanation for why indices vary with spatial resolution (i.e. the first day experiment).

 

L244-245: Reflectance panel wasn’t mentioned earlier. Or is this the white part of the black and white targets? What material was used? Was the irradiance sensor used for correction?  UAV NDVI values do seem high for this ecosystem, so it’s important that the calibration procedure is clear.

 

In addition to thermal infrared measurement of stress, solar induced fluorescence should also be mentioned e.g. Zarco-Tejada et al., 2012

 

Minor comments:

L5: formatting, “and”


L48: formatting, extra space before “and”


L50: “manifest as diurnal and seasonal changes” What kind of changes?


L67: “It is common” I think you should limit this to “Using broadband multispectral remote sensing it is common”, because otherwise you’re excluding other types of remote sensing that have used for vegetation applications.


L72: By “greenness” do you mean color? KT transform greenness?


L116: “estimating” is probably too strong here, since there was no validation. This canopy water content was purely a spectral measure, and while changes were described in response to drought, it’s not clear this measure correlates with actual canopy water content.


I think Section 1.2 can be safely reduced, since NDVI (and to a more limited extent, NDRE) should be very familiar to this audience.


L206: “were” should be “was” to agree with “content”, or change to “water content measurements” to agree with “were”


L207: “contents” and “samples” don’t agree


Check with journal to make sure that date formatting is okay. Using month/day is mostly specific to the US and Canada.


Table 1: Without going into the text and lining up the dates, it’s not clear how “Day 1” and “Day 2” align with the cut dates and with the date imagery was flown. This is later resolved in the following section, but you might consider adding date information to the table caption or to the table itself.


Field sampling mentioned measurement of leaf water content, but later on plant water content is referred to. Make sure terminology is consistent and defined.


L309-310: Leaf and plant water content are used in these two sentences. Are they the same thing?


L323: Can you add any more detail beyond “likely a mistake”?


L422: Can you really determine optimality based on this one experiment?


L427: “In this paper we show that the derived NDVI values from UAVs and PlanetScope CubeSat imagery across treatments are consistent” Is this really what Figure 5b shows?

 

Author Response

Reviewer 3

Major concerns:

 

The comparison between NDVI values from the UAV camera and Planet data could be incorrect. I inferred that NDVI values were spatially averaged from the UAV data and then compared to the Planet NDVI values (if this is not the case, then please make it clear in the manuscript). Because NDVI does not scale linearly with reflectance, this type of averaging does not work. For example, consider two pixels with red reflectances [0.1,0.2] and NIR reflectances [0.4,0.25]. If their reflectances are averaged, red = 0.15 and NIR = 0.325. Averaging pixel NDVI values produces an average NDVI of 0.35, but calculating NDVI from the averaged reflectance value produces an NDVI of 0.37. To correctly compare NDVI across spatial resolutions, you should first resample reflectance of the UAV camera data to match the Planet data spatial resolution, and then calculate NDVI.

Response: The reviewer is correct that we needed to clarify this portion of the methods section, and the concerns captured in the above statement were valid. We did however resample each band to match the 3m Planet resolution before calculating NDVI so the values reflected in the manuscript and the figures are correct. We clarified our procedure in the methods as follows:

“We resampled each of the radiometrically calibrated bands from the UAV flight on July 26 2018 at 11:10am (the closest temporal match to the native resolution of PlanetScope Imagery) to 3 m to match the native pixel resolution of PlanetScope imagery. Using the resampled UAV band values we then calculated NDVI. The PlanetScope Ortho Scenes Surface Reflectance product is a 16-bit GeoTIFF with reflectance values scaled by 10,000. We divided the pixel values by 10,000 to compare the reflectance bands and NDVI index of the PlanetScope Imagery with those collected from the UAV [54]. PlanetScope Imagery has 4 bands: blue (455-515 nm), green (500-590 nm), red (590-670 nm), and NIR (780-860 nm).”

 

L339-341: “The resampled UAV NDVI pixel values contained more spectral information, suggesting higher radiometric resolution for the MicaSense RedEdge narrow band sensor in comparison to PlantScope’s sensor.” It’s not clear that more variation in values indicates higher radiometric resolution. It’s also not clear whether “narrow band” is supposed to indicate a difference in spectral resolution. L432 appears to confuse radiometric resolution (number of possible DN/radiance values in a measurement) with spectral resolution. If spectral resolution is really what you’re getting at, I don’t understand why differences in spectral resolution would cause more variation in NDVI values, rather than just a simple bias. A better explanation is needed.

Response:

In response to the reviewer’s comments we altered the text to clarify our point about the differences in the spectral resolution of these sensors. The next text is below. We have added a figure to this response, but did not add it to the paper.

“A narrow band sensor such as the RedEdge can provide more detailed information by capturing a more precise measurement of specific wavelengths. Therefore, the spectral sensitivity of sensor was preserved despite being the spatial resolution being resampled.”

 

Furthermore, we would expect that there would be more variation in NDVI values for the Micasense sensors due to the narrower range in spectral values for each band which discerns more variations in surface reflectance whereas the larger spectral ranges encompassed by Planet’s bands and averaged over a larger range of reflectance values leading less overall variation detected and resulting in a narrower distribution of NDVI values. We have done some additional analysis to support this hypothesis (see figure below)

 

There wasn’t an explanation for why indices vary with spatial resolution (i.e. the first day experiment).

Response: This was an omission. We have added the text:

“Spectral indices vary with spatial resolution for a number of reasons, including the fact that lighting changes throughout the day, as well as leaf moisture.”

 

 L244-245: Reflectance panel wasn’t mentioned earlier. Or is this the white part of the black and white targets? What material was used? Was the irradiance sensor used for correction? UAV NDVI values do seem high for this ecosystem, so it’s important that the calibration procedure is clear.

Response: This information has been added to the Methods:

“Physical radiometric targets were imaged prior to flight for radiometric calibration.”

 

In addition to thermal infrared measurement of stress, solar induced fluorescence should also be mentioned e.g. Zarco-Tejada et al., 2012

Response: In response to a comment from reviewer 2, we have removed the section on thermal sensing. We have added discussion of fluorescence in the final Discussion paragraph:

 

“New work also highlights the ability to measure solar induced fluorescence via narrowband remote sensing. About 1% of solar energy captured by plants is re-emitted by chlorophyll as fluorescence. The amount of fluorescence emission is a direct indicator of the photosynthetic activity of a plant and also provides indirect information about the plant stress [87]. Remote sensing of solar induced fluorescence requires very narrow spectral bands (~0.05 nm) around 690nm and 760nm [80,88].”

 

Minor comments:

 

L5: formatting, “and”

Response: change made.

 

L48: formatting, extra space before “and”

Response:  change made.

 

L50: “manifest as diurnal and seasonal changes” What kind of changes?

Response:  We have changed the text to the following:

“Such water limitation can result in a diversity of plant responses such as reductions in carbon fixation, growth and even reproduction.”

 

L67: “It is common” I think you should limit this to “Using broadband multispectral remote sensing it is common”, because otherwise you’re excluding other types of remote sensing that have used for vegetation applications.

Response: change made.

 

L72: By “greenness” do you mean color? KT transform greenness?

Response: “greenness” changed to “green color”. Line 77.

 

L116: “estimating” is probably too strong here, since there was no validation. This canopy water content was purely a spectral measure, and while changes were described in response to drought, it’s not clear this measure correlates with actual canopy water content.

Response: Good point: “estimating” changed to “evaluating” line 125

 

I think Section 1.2 can be safely reduced, since NDVI (and to a more limited extent, NDRE) should be very familiar to this audience.

Response: Agreed. We have cut this section down considerably.

 

L206: “were” should be “was” to agree with “content”, or change to “water content measurements” to agree with “were”

Response: changed to “water content measurements”

 

L207: “contents” and “samples” don’t agree

Response:  changed to “content”.

 

Check with journal to make sure that date formatting is okay. Using month/day is mostly specific to the US and Canada.

Response: All instances of “7/day” changed to “July day”.

 

Table 1: Without going into the text and lining up the dates, it’s not clear how “Day 1” and “Day 2” align with the cut dates and with the date imagery was flown. This is later resolved in the following section, but you might consider adding date information to the table caption or to the table itself.

Response:  Date added to Table 1.

 

Field sampling mentioned measurement of leaf water content, but later on plant water content is referred to. Make sure terminology is consistent and defined.

Response: We have changed this to “leaf water content” and “leaf water potential” when we are discussing this experiment. Instances where “plant water stress” or “plant water status” are used in the more general sense remain.

 

L309-310: Leaf and plant water content are used in these two sentences. Are they the same thing?

Response: We have changed this to “leaf water content” and “leaf water potential” when we are discussing this experiment. Instances where “plant water stress” or “plant water status” are used in the more general sense remain.

 

L323: Can you add any more detail beyond “likely a mistake”?

Response: Good point. This was poorly worded. We think this was an anomalous data point due to the fact that we sampled immediately following cutting. But in the spirit of being transparent with all of the data we collected we have included it and then flagged it.  We’ve changed the text to read:

“[Note: the data point at -5 MPa and 52% water content is an anomaly and likely due to the timing of the treatment (immediately following cutting)]”

 

L422: Can you really determine optimality based on this one experiment?

Response: Agreed. “optimal” changed to “useful”.

 

L427: “In this paper we show that the derived NDVI values from UAVs and PlanetScope CubeSat imagery across treatments are consistent” Is this really what Figure 5b shows?

Response: The original statement is a bit misleading. We have changed it to: “In this paper we show that the derived NDVI values from UAVs and PlanetScope CubeSat imagery across treatments are largely consistent with important differences, particularly in their sensitivity to subtle changes to water content (control and water).”


Author Response File: Author Response.docx

Round 2

Reviewer 3 Report

The following sentence appears to be repeated twice: “PlanetScope Imagery has 4 bands: blue (455-515 nm), green (500-590 nm), red (590-670 nm), and NIR (780-860 nm).”

“We resampled each of the radiometrically calibrated bands from the UAV flight on July 26 2018 at 11:10am (the closest temporal match to PlanetScope Imagery), to 3 m to match the native pixel resolution of PlanetScope imagery.” Later on, the resampled and PlanetScope imagery are both referred to as 3.7 m, not 3 m. Which spatial resolution was used?

“Although there was a positive correlation between NDVI derived from the UAV and PlanetScope, the UAV imagery captured a larger spectral range suggesting a greater radiometric sensitivity to plant responses to water deficit in the UAV camera than the PlanetScope imagery.” Not updated to reflect change in explanation earlier in the manuscript.


Author Response

Reviewer 3, Round 2

The following sentence appears to be repeated twice: “PlanetScope Imagery has 4 bands: blue (455-515 nm), green (500-590 nm), red (590-670 nm), and NIR (780-860 nm).”

Response: you are correct. We have removed the second sentence.

“We resampled each of the radiometrically calibrated bands from the UAV flight on July 26 2018 at 11:10am (the closest temporal match to PlanetScope Imagery), to 3 m to match the native pixel resolution of PlanetScope imagery.” Later on, the resampled and PlanetScope imagery are both referred to as 3.7 m, not 3 m. Which spatial resolution was used?

Response: That was a mistake. It should read 3.0 throughout. We have made the change.

“Although there was a positive correlation between NDVI derived from the UAV and PlanetScope, the UAV imagery captured a larger spectral range suggesting a greater radiometric sensitivity to plant responses to water deficit in the UAV camera than the PlanetScope imagery.” Not updated to reflect change in explanation earlier in the manuscript.

Response: You are correct, this was missed. We have changed the text to read:

However, PlanetScope data was not able to capture subtle changes in water content. Although there was a positive correlation between NDVI derived from the UAV and PlanetScope, the UAV imagery captured a larger spectral range suggesting a greater spectral sensitivity to plant responses to water deficit with the UAV camera than with the PlanetScope imagery. In this case a narrow band sensor such as the RedEdge provided more detailed information by capturing a more precise measurement of specific wavelengths than did the PlanetScope sensor.”


Back to TopTop