Optical water type discrimination and tuning remote sensing band-ratio algorithms: Application to retrieval of chlorophyll and Kd(490) in the Irish and Celtic Seas

David J.C. McKee, Alex Cunningham, Agnes V. Dudek

Research output: Contribution to journalArticle

33 Citations (Scopus)

Abstract

This paper assesses the feasibility of applying remote sensing algorithms based on blue/green reflectance ratios to Case 2 waters. Two algorithms from the SeaDAS (NASA) image processing package, OC4v4 for surface chlorophyll concentration, Chl, and K(490) for the attenuation coefficient for downward irradiance at 490 nm, Kd(490), were investigated using an extensive set of observations from the Irish and Celtic Seas. In situ data from a profiling radiometer were used as inputs for the algorithms to avoid uncertainties in atmospheric correction procedures, and direct measurements of Chl and Kd490 and were used for validation purposes. The standard versions of the algorithms performed poorly: OC4v4 generally overestimated Chl (with a very low coefficient of determination), and K(490) progressively underestimated Kd490 for values greater than 0.3 m−1. A two-step procedure for level 2 product generation was therefore devised in which the numerical coefficients of OC4v4 and K(490) were tuned for the two optical water types known to occur most frequently in this region (McKee, D., Cunningham, A., 2006. Identification and characterisation of two optical water types in the Irish Sea from in situ inherent optical properties and seawater constituents. Estuarine, Coastal and Shelf Science 68, 305-316) by statistical regression on a data set of 102 stations from the Irish and Celtic Seas. The water types were distinguished by the magnitude of their normalised water leaving radiance signals at 665 nm, nLw(665), and appropriate versions of the tuned algorithms applied to each water type. When this procedure was tested on an independent data set of 19 stations from the Bristol Channel, Chl values were recovered with an RMS error of 0.36 mg m−3 and Kd(490) values with an RMS error of 0.095 m−1. The identification of water types from water-leaving radiance signals, and the application of band-ratio algorithms tuned for specific water types, may therefore provide a simple means of improving the quality of remote sensing products in optically complex shelf seas.
LanguageEnglish
Pages827-834
Number of pages7
JournalEstuarine, Coastal and Shelf Science
Volume73
Issue number3-4
DOIs
Publication statusPublished - Jul 2007

Fingerprint

remote sensing
chlorophyll
water
radiance
Irish Sea
radiometers
sea
optical properties
atmospheric correction
shelf sea
reflectance
image processing
optical property
radiometer
regression analysis
irradiance
uncertainty
seawater
image analysis

Keywords

  • ocean colour
  • shelf seas
  • chlorophyll
  • diffuse attenuation
  • Irish Sea
  • water types
  • optics
  • nanoscience

Cite this

@article{6af4f6ae9a8748e8b607f4c4168af594,
title = "Optical water type discrimination and tuning remote sensing band-ratio algorithms: Application to retrieval of chlorophyll and Kd(490) in the Irish and Celtic Seas",
abstract = "This paper assesses the feasibility of applying remote sensing algorithms based on blue/green reflectance ratios to Case 2 waters. Two algorithms from the SeaDAS (NASA) image processing package, OC4v4 for surface chlorophyll concentration, Chl, and K(490) for the attenuation coefficient for downward irradiance at 490 nm, Kd(490), were investigated using an extensive set of observations from the Irish and Celtic Seas. In situ data from a profiling radiometer were used as inputs for the algorithms to avoid uncertainties in atmospheric correction procedures, and direct measurements of Chl and Kd490 and were used for validation purposes. The standard versions of the algorithms performed poorly: OC4v4 generally overestimated Chl (with a very low coefficient of determination), and K(490) progressively underestimated Kd490 for values greater than 0.3 m−1. A two-step procedure for level 2 product generation was therefore devised in which the numerical coefficients of OC4v4 and K(490) were tuned for the two optical water types known to occur most frequently in this region (McKee, D., Cunningham, A., 2006. Identification and characterisation of two optical water types in the Irish Sea from in situ inherent optical properties and seawater constituents. Estuarine, Coastal and Shelf Science 68, 305-316) by statistical regression on a data set of 102 stations from the Irish and Celtic Seas. The water types were distinguished by the magnitude of their normalised water leaving radiance signals at 665 nm, nLw(665), and appropriate versions of the tuned algorithms applied to each water type. When this procedure was tested on an independent data set of 19 stations from the Bristol Channel, Chl values were recovered with an RMS error of 0.36 mg m−3 and Kd(490) values with an RMS error of 0.095 m−1. The identification of water types from water-leaving radiance signals, and the application of band-ratio algorithms tuned for specific water types, may therefore provide a simple means of improving the quality of remote sensing products in optically complex shelf seas.",
keywords = "ocean colour, shelf seas, chlorophyll, diffuse attenuation, Irish Sea, water types, optics, nanoscience",
author = "McKee, {David J.C.} and Alex Cunningham and Dudek, {Agnes V.}",
year = "2007",
month = "7",
doi = "10.1016/j.ecss.2007.03.028",
language = "English",
volume = "73",
pages = "827--834",
journal = "Estuarine, Coastal and Shelf Science",
issn = "0272-7714",
number = "3-4",

}

TY - JOUR

T1 - Optical water type discrimination and tuning remote sensing band-ratio algorithms: Application to retrieval of chlorophyll and Kd(490) in the Irish and Celtic Seas

AU - McKee, David J.C.

AU - Cunningham, Alex

AU - Dudek, Agnes V.

PY - 2007/7

Y1 - 2007/7

N2 - This paper assesses the feasibility of applying remote sensing algorithms based on blue/green reflectance ratios to Case 2 waters. Two algorithms from the SeaDAS (NASA) image processing package, OC4v4 for surface chlorophyll concentration, Chl, and K(490) for the attenuation coefficient for downward irradiance at 490 nm, Kd(490), were investigated using an extensive set of observations from the Irish and Celtic Seas. In situ data from a profiling radiometer were used as inputs for the algorithms to avoid uncertainties in atmospheric correction procedures, and direct measurements of Chl and Kd490 and were used for validation purposes. The standard versions of the algorithms performed poorly: OC4v4 generally overestimated Chl (with a very low coefficient of determination), and K(490) progressively underestimated Kd490 for values greater than 0.3 m−1. A two-step procedure for level 2 product generation was therefore devised in which the numerical coefficients of OC4v4 and K(490) were tuned for the two optical water types known to occur most frequently in this region (McKee, D., Cunningham, A., 2006. Identification and characterisation of two optical water types in the Irish Sea from in situ inherent optical properties and seawater constituents. Estuarine, Coastal and Shelf Science 68, 305-316) by statistical regression on a data set of 102 stations from the Irish and Celtic Seas. The water types were distinguished by the magnitude of their normalised water leaving radiance signals at 665 nm, nLw(665), and appropriate versions of the tuned algorithms applied to each water type. When this procedure was tested on an independent data set of 19 stations from the Bristol Channel, Chl values were recovered with an RMS error of 0.36 mg m−3 and Kd(490) values with an RMS error of 0.095 m−1. The identification of water types from water-leaving radiance signals, and the application of band-ratio algorithms tuned for specific water types, may therefore provide a simple means of improving the quality of remote sensing products in optically complex shelf seas.

AB - This paper assesses the feasibility of applying remote sensing algorithms based on blue/green reflectance ratios to Case 2 waters. Two algorithms from the SeaDAS (NASA) image processing package, OC4v4 for surface chlorophyll concentration, Chl, and K(490) for the attenuation coefficient for downward irradiance at 490 nm, Kd(490), were investigated using an extensive set of observations from the Irish and Celtic Seas. In situ data from a profiling radiometer were used as inputs for the algorithms to avoid uncertainties in atmospheric correction procedures, and direct measurements of Chl and Kd490 and were used for validation purposes. The standard versions of the algorithms performed poorly: OC4v4 generally overestimated Chl (with a very low coefficient of determination), and K(490) progressively underestimated Kd490 for values greater than 0.3 m−1. A two-step procedure for level 2 product generation was therefore devised in which the numerical coefficients of OC4v4 and K(490) were tuned for the two optical water types known to occur most frequently in this region (McKee, D., Cunningham, A., 2006. Identification and characterisation of two optical water types in the Irish Sea from in situ inherent optical properties and seawater constituents. Estuarine, Coastal and Shelf Science 68, 305-316) by statistical regression on a data set of 102 stations from the Irish and Celtic Seas. The water types were distinguished by the magnitude of their normalised water leaving radiance signals at 665 nm, nLw(665), and appropriate versions of the tuned algorithms applied to each water type. When this procedure was tested on an independent data set of 19 stations from the Bristol Channel, Chl values were recovered with an RMS error of 0.36 mg m−3 and Kd(490) values with an RMS error of 0.095 m−1. The identification of water types from water-leaving radiance signals, and the application of band-ratio algorithms tuned for specific water types, may therefore provide a simple means of improving the quality of remote sensing products in optically complex shelf seas.

KW - ocean colour

KW - shelf seas

KW - chlorophyll

KW - diffuse attenuation

KW - Irish Sea

KW - water types

KW - optics

KW - nanoscience

UR - http://dx.doi.org/10.1016/j.ecss.2007.03.028

U2 - 10.1016/j.ecss.2007.03.028

DO - 10.1016/j.ecss.2007.03.028

M3 - Article

VL - 73

SP - 827

EP - 834

JO - Estuarine, Coastal and Shelf Science

T2 - Estuarine, Coastal and Shelf Science

JF - Estuarine, Coastal and Shelf Science

SN - 0272-7714

IS - 3-4

ER -