Performance evaluation of three bio-optical models in aerosol and ocean color joint retrievals

Hannadige, Neranga K.; Zhai, Peng-Wang; Gao, Meng; Hu, Yongxiang; Werdell, P. Jeremy; Knobelspiesse, Kirk; Cairns, Brian

doi:https://doi.org/10.5194/amt-2023-142

Preprints

https://doi.org/10.5194/amt-2023-142

Preprints

25 Jul 2023

| 25 Jul 2023

Status: this preprint is currently under review for the journal AMT.

Performance evaluation of three bio-optical models in aerosol and ocean color joint retrievals

Neranga K. Hannadige, Peng-Wang Zhai, Meng Gao, Yongxiang Hu, P. Jeremy Werdell, Kirk Knobelspiesse, and Brian Cairns

Abstract. Multi-angle polarimeters (MAP) are powerful instruments to perform remote sensing of the environment. Joint retrieval algorithms of aerosols and ocean color have been developed to extract the rich information content of MAPs. These are optimization algorithms that fit the sensor measurements with forward models, which include radiative transfer simulations of the coupled atmosphere and ocean systems (CAOS) based on adjustable atmosphere and ocean properties. The forward mode consists of sub-models to represent the optics of the atmosphere, ocean water surface, and ocean body. The representativeness of these models for observed scenes is important for retrieval success. In this study, we have evaluated the impact on MAP retrieval accuracy of three different ocean bio-optical models with 1, 3, and 7 optimization parameters that represent the spectral variation of inherent optical properties (IOP(λ)s) of the water body. The Multi-Angular Polarimetric Ocean coLor (MAPOL) joint retrieval algorithm was used to process data from the airborne Research Scanning Polarimeter (RSP) instrument acquired in different field campaigns. We performed ensemble retrievals along three RSP legs to evaluate the applicability of bio-optical models along geographically varying waters. The average differences between the MAPOL aerosol optical depth (AOD) and spectral remote sensing reflectance (R_rs(λ)) retrievals and the MODerate resolution Imaging Spectroradiometer (MODIS) products are also reported. We studied the distribution of retrieval cost function values obtained for the ensemble retrievals using the 3 bio-optical models under clear to highly turbid waters. For the 1-parameter model, retrieval cost function values show narrow distributions over any type of water, regardless of the cost function values, whereas for the 3 and 7- parameter models, the retrieval cost function distribution is water type dependent, showing the widest distribution over clear, open waters. We observed that the 3 and 7-parameter models have similar MAP retrieval performances relative to the 1- parameter model. We also demonstrated that the 3 and 7-parameter bio-optical models can be used to accurately represent both clear, open, and turbid, coastal waters, whereas the 1-parameter model is most successful over extremely clear waters. Given the computational efficiency requirements, we recommend the 3-parameter bio-optical model as the coastal water bio-optical model for future MAPOL studies. This study guides MAP algorithm development for current and future satellite missions such as NASA’s Plankton, Aerosol, Cloud, ocean Ecosystem (PACE) mission and ESA’s Meteorological Operational-Second Generation (MetOp-SG) mission.

Received: 06 Jul 2023 – Discussion started: 25 Jul 2023

Download & links

Neranga K. Hannadige et al.

Status: final response (author comments only)

RC1: 'Comment on amt-2023-142', Anonymous Referee #1, 10 Aug 2023

Please see attached.

Citation: https://doi.org/10.5194/amt-2023-142-RC1
RC2:
'Comment on amt-2023-142', Anonymous Referee #2, 15 Aug 2023
Summary

This paper evaluates 1-, 3-, and 7-parameter ocean bio-optical models using the Multi-Angular Polarimetric Ocean coLor (MAPOL) joint retrieval algorithm applied to data from the Research Scanning Polarimeter (RSP) in different types of water (clear and open water and turbid coastal water). The authors compare MAPOL retrievals to MODIS products.

The topic of this paper is interesting and the approach of this paper seems sound. However, it is unclear where the novelty lies in this paper. No new algorithms are developed, and the conclusion that the 3- and 7-parameter models are more versatile than the 1-parameter model is fairly intuitive. Furthermore, there are deviations between the MAPOL retrievals and MODIS retrievals which are not fully explained. Thus, I recommend acceptance with major revisions. I suggest that the authors further analyze the differences between MAPOL retrievals and MODIS retrievals, and that the authors revise the paper to clarify the major novel contributions of this paper to the reader.

In general, I agree with the comments posted by Anonymous Referee #1 as RC1.
Comments

Figure 4: At a wavelength of 470 nm, it seems like the MAPOL retrievals significantly underestimate AOD when compared to MODIS. Generally, MODIS retrievals are outside of the error bars for MAPOL retrievals. The authors note this (in lines 310-311) but do not explain it -- what are the potential causes? This also appears to be an issue at 550 nm and 670 nm, but at these wavelengths, MODIS retrievals are at least generally within 1-sigma error bars for MAPOL retrievals.

Figure 5: It might improve clarity to include a line on the plot showing where the coast actually is, or make the x-axis the distance from coast rather than longitude. At first glance, it can be difficult to understand where the coast is.

Figure 6: I recommend including the full figure caption (and see above comment about deviations from MODIS).

Figure 9: At 410 nm, spectral remote sensing reflectance from MODIS is consistently lower than from MAPOL (and is sometimes negative). Why is this?

Figure 9: Please include the full figure caption.

General note on figures: I recommend including the full figure caption for all figures.

Lines 123-124: I recommend further explanation of POLYAC and this paper’s relevance to POLYAC. Perhaps POLYAC performance improvements can be used to motivate this paper (if it can be shown that novel contributions in this paper improve MAPOL retrievals).

Lines 316-322: Why does MODIS estimate higher spectral remote sensing reflectance than MAPOL for pixels other than the ones closest to the coast?

Line 327: The authors note that the uncertainty for AOD is less than 0.01 at all wavelengths -- but is the uncertainty adequately capturing error/differences from MODIS retrievals (especially noting the differences between MODIS and MAPOL)?
Citation: https://doi.org/10.5194/amt-2023-142-RC2

Neranga K. Hannadige et al.

Viewed

Total article views: 290 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
232	45	13	290	6	5

HTML: 232
PDF: 45
XML: 13
Total: 290
BibTeX: 6
EndNote: 5

Views and downloads (calculated since 25 Jul 2023)

Month	HTML	PDF	XML	Total
Jul 2023	81	16	4	101
Aug 2023	113	22	7	142
Sep 2023	38	7	2	47

Cumulative views and downloads (calculated since 25 Jul 2023)

Month	HTML	PDF	XML	Total
Jul 2023	81	16	4	101
Aug 2023	113	22	7	142
Sep 2023	38	7	2	47

Viewed (geographical distribution)

Total article views: 284 (including HTML, PDF, and XML) Thereof 284 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 23 Sep 2023

Short summary

We evaluated the impact of three ocean optical models with different numbers of free parameters on the performance of an aerosol and ocean color remote sensing algorithm using the multi-angle polarimeter (MAP) measurements. It was demonstrated that the 3 and 7-parameter bio-optical models can be used to accurately represent both open and coastal waters, whereas the 1-parameter model does have smaller retrieval uncertainty over open waters.


Total:	0
HTML:	0
PDF:	0
XML:	0