Underway Hyperspectral Bio-Optical Assessments of Phytoplankton Size Classes in the River-Influenced Northern Gulf of Mexico
Round 1
Reviewer 1 Report
Summary: Shipborne hyperspectral measurements of above-water remote sensing reflectance are used to distinguish phytoplankton size classes at sub-kilometer scales. Community composition varies, and is associated with the discharge plumes of the Mississippi and Atchafalaya rivers. Freshwater inflow into coastal areas contributes to physical and geochemical gradients that influence phytoplankton community composition.
General comments:
This is an exceptionally well written paper, with clear methods. The results demonstrate the strength of hyperspectral measurements, and are an important contribution to the field.
I have no corrections.
Accept without revision.
Author Response
We appreciate Reviewer 1's positive encouragement for our work, thank you.
Reviewer 2 Report
This manuscript describes a study retrieving representative bio-optical signatures of phytoplankton size classes (PSCs) from hyperspectral measurements of above-water remote sensing reflectance using the quasi-analytical algorithm (QAA_v6) and the validation against in situ pigment data and spectrophotometric analyses of phytoplankton absorption. Results were used to interpret sub-kilometer scale variability in PSCs associated with dynamic and spatially heterogeneous environmental processes in river-influenced oceanic waters. The manuscript is well written and organized, the work is solid and publishable after some minor revisions.
- The title emphasizes the reflectance assessment, which is not quite accurate. It is better to include some phrases on bio-optical signatures/properties and reflect the mechanism governing the variation of these signatures.
- In section 2.3, tidal and water level data are mentioned, but their use in this study is not clearly stated or described. On the other hand, salinity and temperature are used in Figure 10 and 11, but how these data were obtained is not fully described.
- Above water reflectance measurement is described in section 2.5, a diagram describing the setup or configuration of this measurement would be helpful.
- Figure 7shows the correlation between HyperSAS reflectance and MODIS reflectance. It is evident that the correlation is poor when reflectance is below 0.001 for bands 412, 667, 678 nm. This low correlation should be explained either in the text of discussion or in the caption of this plot, a low signal to noise ratio of MODIS is obvious for thee bands at low reflectance values.
- Figure 9 shows very scattered relationship for stations in estuary and inner shelf relative to those from midshelf and open ocean, which reflects the poor performance of QAA_6 in retrieving aph for samples from estuary and inner shelf because of possible effects of cdom and sediment. This should be pointed out or at least discussed somehow.
- The first four paragraphs of section 4.1 are redundant, and could be shortened to one or two short paragraph because some arguments are so obvious and no need to use four paragraphs to explain.
- Figure 12 uses horizontal black bars to indicate location of the bottom surface, but how location of the bottom surface as determined should be explained in the caption.
- Two places need rewording: line 241, conducted should be replaced with recorded; line 577, allow should be replaced with allows.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 3 Report
Review of Underway Hyperspectral Reflectance Assessments of Phyto-plankton Size Classes in the River-Influenced Northern Gulf of Mexico (Verma et al.)
This is a very ambitious paper, involving combing through a great deal of measurements taken from a ship. The results look good, and the paper is well written. However, the results depend on heavy data filtering, which is given short shrift:
Satellite imagery: how much of the data was filtered away?
Hyper OCR: only 5% to 10% of the data left, then further filtered to remove ship movement, then outliers, then median filtered. How much was left?
OCR profiles: filtered by r2, then additional quality filter, then outliers. How much was left?
The above data filtering should be elaborated upon.
I am not an expert on phytoplankton, but it seems to me that these shipboard data are problematic to acquire and process, and that hyperspectral remote sensing solutions would be superior.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
