Research papersAssessment of MODIS-Aqua chlorophyll-a algorithms in coastal and shelf waters of the eastern Arabian Sea
Introduction
The unprecedented coverage of ocean colour data, both spatially and temporally over the global ocean, has provided important insights into phytoplankton dynamics and biogeochemical cycling in the marine environment (Bousquet et al., 2006, Mohr and Forsberg, 2002). As a consequence, in areas such as the Arabian Sea Shelf waters, ocean colour Chlorophyll-a (Chla) data, in combination with sea surface temperature (SST) and wind speed, is being used operationally to indicate potential fishing zones and fish conservation zones (Petit et al., 2003, Royer et al., 2004).
Most operational satellite Chla algorithms are empirical, switching band ratios (O'Reilly et al., 1998, O'Reilly et al., 2000). They perform well where the optically active substance in the water column is phytoplankton (Sathyendranath et al., 1999), but tend to fail in regions where the Inherent Optical Properties (IOP) are determined not only by phytoplankton, but also by absorption due to coloured dissolved organic material (CDOM) (aCDOM(λ)) and scattering by Total Suspended Matter (TSM) (bp(λ))(Prieur and Sathyendranath, 1981). The presence of CDOM and TSM in coastal waters however, can cause errors in the retrieval of ocean colour Chla especially when using empirical algorithms (Darecki and Stramski, 2004). This is predicted to become worse under future climate change scenarios as TSM and CDOM become uncoupled from phytoplankton Chla (Dierssen, 2010). CDOM and TSM modify the normalised water leaving radiance (nLw), which can lead to significant errors in Chla retrieval from satellite (IOCCG, 2006). To overcome this problem, semi-analytical ocean colour algorithms have been developed, based on radiative transfer solutions from a knowledge of the IOP, which are usually derived from a large in situ database for a specific area (Lee et al., 2002, Maritorena et al., 2002, Sathyendranath et al., 2001, Tilstone et al., 2012). Parallel to the development of these algorithms, advancements in ocean colour satellite sensors, with more spectral bands and novel atmospheric correction models, has provided more accurate ocean colour products (Zibordi et al., 2006). The Moderate Resolution Imaging Spectro-radiometer onboard the Aqua satellite (MODIS-Aqua) is the current operational, medium resolution mission from NASA, and has built on previous missions, such as Sea-viewing Wide Field-of-view Sensor (SeaWiFS), by having more spectral bands and better spatial resolution. Though the accuracy of Chla algorithms for some satellite sensors in some coastal zones has been evaluated (Zibordi et al., 2009), sensor characteristics, atmospheric correction models and level 2 product algorithms used for each sensor differ (Zibordi et al., 2006). MODIS has recently experienced radiometric drift, that has been addressed through frequent vicarious calibration and re-processing of the data. The reprocessed data subsequently needs to be quantified through on-going validation exercises. An accuracy assessment of ocean products available from MODIS-Aqua is still therefore necessary and especially in coastal regions where the optical properties change more dynamically. Though ocean colour products from Coastal Zone Colour Scanner (CZCS), SeaWiFS and Indian Remote Sensing Satellite (ISR)-P4 Ocean Colour Monitor (OCM) have been validated in the open ocean waters of the Arabian Sea (Banse and English, 2000, Banzon et al., 2004, Chauhan et al., 2002, Desa et al., 2001), there have been few ocean colour validation studies in the coastal waters (Chauhan et al., 2003, Menon et al., 2006), and few for Chla products available from MODIS-Aqua (Shanmugam, 2011). There is therefore, an obvious need to assess MODIS-Aqua algorithms in coastal areas of the Arabian Sea to identify the most accurate Chla product available for the region for the on-going monitoring of phytoplankton biomass.
The principal objective of this study was to assess the performance of Chla algorithms available for MODIS-Aqua in coastal waters of the eastern Arabian Sea. The effect of potential errors in the atmospheric correction and from CDOM and TSM were assessed. Subsequently, using the most accurate Chla algorithms, time series were generated from 2002–2011, and spatial and temporal differences between them were analysed.
Section snippets
Study area
The eastern Arabian Sea shelf is defined as the Western Indian Coastal Province (INDW) (Longhurst, 2007). It exhibits intense upwelling along the southern coast during the southwest (summer) monsoon, resulting in a deepening of the mixed layer and injection of subsurface nutrient rich waters to surface and enhanced biological production. During the northeast (winter) monsoon the INDW becomes oligotrophic or mesotrophic (Brock et al., 1991, Shetye et al., 1994). The area adjacent to Kochi is
Variability in Chla, TSM and aCDOM(λ)
Along the central and south eastern Arabian Sea Shelf, Chla varied from 0.01 mg m−3 to 38.85 mg m−3 (Fig. 2a, b), and reached a minima during pre-monsoon and maxima during monsoon and post-monsoon. TSM varied from 0.33 g m−3 to 61.27 g m−3 (Fig. 2a, c) with maxima during the winter and pre-monsoon. The variability in aCDOM(443) was between 0.03 m−1 and 1.56 m−1 with maxima during the monsoon and minima during post-monsoon (Fig. 2a, b). The co-variation between Chla, TSM and aCDOM(443) was assessed by
Bio-optical properties of the eastern Arabian Sea coast
There was a significant correlation between Chla and TSM (Fig. 2a), suggesting that either (1) at the time of the monsoon blooms, inorganic TSM is also input to the sea surface through re-suspension from the sea floor or river run-off, (2) or that the monsoon blooms result in high organic TSM that co-varies with phytoplankton Chla, (3) or both of these processes occur. There was no significant correlation between Chla and aCDOM(443), suggesting that in these coastal waters aCDOM(443) is
Conclusion
The eastern Arabian Sea coast in the north and south were principally Case 2 Chla–TSM–CDOM type, in contrast to the central eastern coast which was Case 2 TSM–CDOM type. Three MODIS-Aqua ocean colour algorithms were evaluated; OC3M was the most accurate, GSM showed a similar accuracy for the south eastern coast and GIOP was the least accurate. OC3M was more accurate because it principally used Rrs(488):Rrs(547) bands, which are less affected by errors in the atmospheric correction. The GSM and
Acknowledgements
G. Tilstone was supported by the European Space Agency (ESA) Project COASTCOLOUR (Contract no. 22807/09/I-AM), A. Lotliker by a Partnership for Observation of the Global Oceans (POGO) Scientific Committee on Oceanic Research (SCOR) Visiting Fellowship to Plymouth Marine Laboratory, UK and P. Miller by Natural Environment Research Council (NERC) Oceans 2025 funding. In situ data were collected within the framework of the Satellite Coastal and Oceanographic Research (SATCORE) programme
References (43)
- et al.
A multi-sensor approach for the on-orbit validation of ocean color satellite data products
Remote Sensing of Environment
(2006) - et al.
Geographical differences in seasonality of CZCS-derived phytoplankton pigment in the Arabian Sea for 1978–1986
Deep Sea Research Part II: Topical Studies in Oceanography
(2000) - et al.
SeaWiFS observations of the Arabian Sea southwest monsoon bloom for the year 2000
Deep Sea Research Part II: Topical Studies in Oceanography
(2004) - et al.
An evaluation of MODIS and SeaWiFS bio-optical algorithms in the Baltic Sea
Remote Sensing of Environment
(2004) - et al.
Seasonal variations in bio-optical properties of phytoplankton in the Arabian Sea
Deep-Sea Research Part II-Topical Studies in Oceanography
(1999) - et al.
An assessment of chlorophyll-a algorithms available for SeaWiFS in coastal and open areas of the Bay of Bengal and Arabian Sea
Remote Sensing of Environment
(2011) - et al.
Variability in specific-absorption properties and their use in a semi-analytical ocean colour algorithm for MERIS in North Sea and Western English Channel Coastal Waters
Remote Sensing of Environment
(2012) - et al.
Modeling the spectral shape of absorption by chromophoric dissolved organic matter
Marine Chemistry
(2004) - et al.
Cross-site consistent in situ measurements for satellite ocean color applications: the BiOMaP radiometric dataset
Remote Sensing of Environment
(2011) - et al.
Validation of satellite ocean color primary products at optically complex coastal sites: Northern Adriatic Sea, Northern Baltic Proper and Gulf of Finland
Remote Sensing of Environment
(2009)
Assessment of uncertainty in the ocean reflectance determined by three satellite ocean color sensors (MERIS, SeaWiFS and MODIS-A) at an offshore site in the Mediterranean Sea (BOUSSOLE project)
Journal of Geophysical Research
Contribution of anthropogenic and natural sources to atmospheric methane variability
Nature
Variability in the chlorophyll-specific absorption-coefficients of natural phytoplankton—analysis and parameterization
Journal of Geophysical Research—Oceans
The phytoplankton bloom in the northwestern Arabian Sea during the southwest monsoon of 1979
Journal of Geophysical Research—Oceans
Comparison of algorithms for estimating ocean primary production from surface chlorophyll, temperature, and irradiance
Global Biogeochemical Cycles
Comparative analysis of ocean color measurements of IRS-P4 OCM and SeaWiFS in the Arabian Sea
IEEE Transactions on Geoscience and Remote Sensing
Surface chlorophyll a estimation in the Arabian Sea using IRS-P4 Ocean Colour Monitor (OCM) satellite data
International Journal of Remote Sensing
Sea truth validation of SeaWiFS ocean colour sensor in the coastal waters of the Eastern Arabian Sea
Current Science
Perspectives on empirical approaches for ocean color remote sensing of chlorophyll in a changing climate
Proceedings of the National Academy of Sciences of the United States of America
Cited by (56)
Prediction of water turbidity in a marine environment using machine learning: A case study of Hong Kong
2022, Regional Studies in Marine ScienceLong-term characteristics of water quality around the Guangdong-Hong Kong-Macau Greater Bay Area: Spatiotemporal variations and meteorological effects
2022, Ocean and Coastal ManagementCitation Excerpt :Nutrients were discharge drivers during rainfall events and transported into the ocean, conducive to substantial phytoplankton growth. We infer that this result was related to the beginning of a large number of phytoplankton growth in spring in some areas (Tilstone et al., 2013). Finally, the spring and autumn blooms of phytoplankton was also an important factor affecting the spatial distribution of Chla.
Evaluation of the operational Chlorophyll-a product from global ocean colour sensors in the coastal waters, south-eastern Arabian Sea
2021, Egyptian Journal of Remote Sensing and Space ScienceSeasonality of primary productivity affects coastal species more than its magnitude
2021, Science of the Total EnvironmentCitation Excerpt :Megapixels were used to simplify the analysis of such a large spatial and temporal dataset, and to reduce the high frequency of missing values in individual pixels due to cloud cover or quality control that are usually present in satellite derived measurements. Coastal waters are optically complex areas where chl-a estimates may not only be an estimate of primary productivity but may also reflect other coloured particles derived from sediments, or bottom reflection in shallow waters (Tilstone et al., 2013). Although these shortcomings are acknowledged, and chl-a measures are not a measure of primary production, for simplicity, chl-a is treated here as an estimate of food available for coastal organisms, in the form of phytoplankton or, to a lesser extent, other coloured dissolved organic matter, and treated as a proxy for primary production.
Cyanobacterial blooms in wastewater treatment facilities: Significance and emerging monitoring strategies
2021, Journal of Microbiological Methods