Research papersEstimation of euphotic zone depth in shallow inland water using inherent optical properties and multispectral remote sensing imagery
Introduction
Deteriorating water quality of inland waterbodies is a pressing challenge because of growing population and climate change impacts (Spears et al., 2021, Qing et al., 2021, Chawla et al., 2020, Filatov et al., 2019, Salerno et al., 2018, Shi et al., 2018). Specifically, many inland waterbodies show increasing sediment-bound nutrient load, phytoplankton blooms, and eutrophication (Hamilton, 2021, Jalil et al., 2019, Ficek et al., 2012, Paerl et al., 2011, O’neil et al., 2012). Suspended sediment and phytoplankton are known to modify underwater light environment (Zhou et al., 2018, Song et al., 2017, Zheng et al., 2016, Shi et al., 2014). Optimal light conditions regulate growth and distribution of aquatic plants (Sánchez et al., 2015). The depth of light penetration in a water column (euphotic zone depth, Zeu) and the rate at which light availability diminishes with water depth (diffuse attenuation coefficient, Kd) are two critical parameters used for quantitatively describing underwater light environment (Kirk, 1991). Optically-active constituents (OACs) in water column control these parameters through absorption and scattering processes (Scheffer, 2004, Kirk, 1991, Preisendorfer, 1986). Because both spatial and temporal distribution of OACs significantly influence underwater light environment (Gomes et al., 2020, Brewin et al., 2014, Zheng et al., 2016), frequent mapping of Zeu and Kd are needed for managing aquatic ecosystems.
In oceanic waters, light attenuation is mainly dominated by seawater and chlorophyll contents along with other photosynthetic pigments of living phytoplankton (Lee et al., 2020, Shang et al., 2011). However, light attenuation in coastal and complex inland waterbodies is largely influenced by suspended particulate matter (SPM) and colored dissolved organic material (CDOM) (Simon and Shanmugam, 2016, Matsushita et al., 2015, Shi et al., 2014). Although Secchi disc depth (ZSD) is commonly used as an approximate measure of light attenuation (Jiang et al., 2019, Alikas and Kratzer, 2017, Lee et al., 2015), Kd and Zeu are more quantitative measures of water clarity and light availability (Gomes et al., 2020, Majozi et al., 2014). With the availability of different satellite data such as MODIS, MERIS, SPOT, and Landsat, remote sensing approaches have been developed over the last decade to estimate Zeu and Kd values over photosynthetically active radiation (PAR) for oceans (Wang et al., 2021, Mitchell and Cunningham, 2015, Son and Wang, 2015, Chen et al., 2014, Saulquin et al., 2013, Shang et al., 2011) and several inland waterbodies (Yang et al., 2020, Huang and Yao, 2017, Shen et al., 2017, Song et al., 2017, Liu et al., 2016, Majozi et al., 2014, Zhang et al., 2012). The Copernicus Sentinel-2 from the European Space Agency (ESA) is a relatively new remote sensing mission providing multispectral instrument (MSI) image data from its two platforms (S2A and S2B) with the goal to maintain continuity of SPOT and Landsat-type image data. The S2A/S2B MSI data have not been used for estimating Zeu although higher spatial (10 to 60 m) and temporal (5 day) coverage with high signal to noise ratio make sentinel S2A/S2B MSI data an attractive choice for evaluating inland waterbodies (Pahlevan et al., 2017).
In remote sensing approaches, Zeu values are typically estimated from Kd(PAR):
Kirk (1991). Estimation of Kd(PAR) requires that the downwelling irradiance (Ed) is measured at multiple water depths. Most sensor assemblies required for measuring Ed are designed primarily for ocean water monitoring. It becomes a challenging task to deploy such assemblies in inland waterbodies many of which are quite shallow. Therefore, Kd(PAR) is often estimated using a) empirical relationship between Kd(PAR) and OACs (Doron et al., 2011, Morel et al., 2007), b) semi/quasi analytical inversion models based on radiative transfer modeling that relates Kd(PAR) with remote sensing-derived a(λ) and backscattering coefficient (bb) (Saulquin et al., 2013, Lee et al., 2005), and c) empirical relationship between Kd(PAR) and Rrs (Majozi et al., 2014, Zhang et al., 2012). The OAC-based algorithms are generally unsuitable for estimating Kd(PAR) in optically-complex inland waters (Shi et al., 2014). The quasi-analytical algorithm (QAA) and its updated versions (Lee et al., 2002, Lee et al., 2009, Lee, 2014) are often less accurate in highly turbid/eutrophic waters (Vandermeulen et al., 2015, Shanmugam et al., 2010). Moreover, a fundamental assumption considered in developing QAA for inland water is that the absorption coefficient of water constituents (at-w) value is negligible over the near-infrared (NIR) region. Although the reference wavelength (λ0) selected from the NIR region differs from model to model (e.g., λ = 753 nm in Yang et al., 2013; λ = 709 nm in Li et al., 2013; and λ = 750 nm in Xue et al., 2019), such an assumption often leads to the underestimation of at-w over 400 to 550 nm wavelength region (Röttgers et al., 2013, de Carvalho et al., 2015). Moreover, with the inherent complexity of light propagation in turbid water, many ocean color-algorithms based on inherent optical property (IOP: a and bb) (Gordon et al., 1988, Kirk, 1991, Lee et al., 2005, Lee et al., 2013) often fail to accurately estimate Kd values in turbid and eutrophic waters (Chen et al., 2014, Simon and Shanmugam, 2016). Therefore, many studies adopted empirical models to estimate Kd (PAR) and Zeu directly from in-situ measured or satellite sensor acquired Rrs products for different inland waters (Song et al., 2017, Liu et al., 2016). For oceans and coastal turbid waters Kd data are routinely provided by ocean color satellite sensors. However, in case of inland water, Rrs-based algorithms must be re-evaluated for their applicability to estimate Kd.
In this study, we revisited Rrs based empirical algorithms (method 3) to estimate Kd(PAR) following Zeu using S2A/S2B MSI Rrs data in Chilika lagoon. Being Asia’s largest brackish water lagoon and home to over 800 species of fauna and several aquatic weeds (Rath and Adhikary, 2008), this productive ecosystem provides livelihood to more than 200,000 fishermen in the surrounding area (Mohapatra et al., 2007). With the opportunity of freely-available S2A/S2B MSI data at high spatial and temporal resolutions, specific objectives of this study are to 1) evaluate the performance of existing Rrs based empirical models to estimate Kd(PAR) using S2A/S2B MSI Rrs data, 2) develop a robust empirical algorithm to estimate Kd(PAR) directly from S2A/S2B MSI-Rrs product to map Zeu over Chilika lagoon, and 3) estimate spatial and temporal dynamics of Zeu in Chilika lagoon.
Section snippets
Study site and in situ data collection
Chilika lagoon is located on the eastern coast of Odisha along the Bay of Bengal (19°28′–19°54′ N and 85°06′–85°35′ E) spreading north–south over a stretch of about 65 km. It has an average water depth of 1.8 m (Pradhan et al. 2017) and water-spread area of about 906 km2 to 1165 km2 (Pal and Mohanty 2002). Daya, Bhargavi, Luna and Makara rivers from the Mahanadi basin and 47 streams from the Western Catchment feed freshwater to the lagoon. Seawater enters through narrow mouth along the Bay of
In situ data
Table 2 lists the descriptive statistics for SPM, Chl-a, ZSD, and water depth values measured at different sampling locations in the Chilika lagoon. Sampling locations were distributed over about 60 % of the total area covering the central and northern part of the lagoon (Fig. 1). Salinity values measured with the WETLab’s SBE37 sensor ranged 0.20 to 1.91 psu in the northern sector and 0.14 to 7.95 psu in the central sector during our field campaigns. Sea water entry through the mouth (Fig. 1)
Discussion
Euphotic depth is a key parameter used in characterizing underwater light environment. In general, light attenuation measurements may be used to estimate Zeu. Specifically, Zeu may be readily estimated from Kd(PAR) values, which requires the downwelling irradiance Ed be measured at multiple water depths. Most sensor assemblies required for measuring photometric and radiometric quantities are designed primarily for ocean water monitoring. It becomes a challenging task to deploy such assemblies
Conclusion
Knowledge of Zeu plays a crucial role in studying biogeochemical cycles and ecosystem of our ecologically fragile inland water bodies such as the Chilika lagoon. Capability to estimate Kd(PAR) from S2A/S2B MSI data with efficient models offer an opportunity to monitor such fragile ecosystems in an efficient manner. In this study, we collected water quality parameters covering about 60 % of the total area of the Chilika lagoon coinciding the S2A/S2B MSI data acquisition schedule. Standard Kd
CRediT authorship contribution statement
Sourav Roy: Conceptualization, Data curation, Formal analysis, Methodology, Software, Writing – original draft, Writing – review & editing. Bhabani S. Das: Funding acquisition, Project administration, Resources, Formal analysis, Writing – review & editing.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgement:
This research was supported under the Networked Programme on Imaging Spectroscopy and Applications (NISA) by the Department of Science & Technology, Ministry of Science & Technology, Government of India (Grant #: BDID/01/23/2014-HSRS/13, WAT-I). We thank the Chilika Development Authority, Department of Forest and Environment, Government of Odisha, India for their support and assistance. We thank European Space Agency (ESA) to make Sentinel 2 MSI satellite data freely available. We thank to Ms.
References (83)
- et al.
Improved retrieval of Secchi depth for optically-complex waters using remote sensing data
Ecol. Ind.
(2017) - et al.
On the temporal consistency of chlorophyll products derived from three ocean-colour sensors
ISPRS J. Photogramm. Remote Sens.
(2014) - et al.
A review of remote sensing applications for water security: Quantity, quality, and extremes
J. Hydrol.
(2020) - et al.
A neural network model for remote sensing of diffuse attenuation coefficient in global oceanic and coastal waters: Exemplifying the applicability of the model to the coastal regions in Eastern China Seas
Remote Sens. Environ.
(2014) - et al.
Ocean transparency from space: Validation of algorithms estimating Secchi depth using MERIS, MODIS and SeaWiFS data
Remote Sens. Environ.
(2011) - et al.
Inherent optical properties and remote sensing reflectance of Pomeranian lakes (Poland)
Oceanologia
(2012) - et al.
Satellite estimates of euphotic zone and Secchi disk depths in a colored dissolved organic matter-dominated inland water
Ecol. Ind.
(2020) - et al.
Sunglint correction of the Multi-Spectral Instrument (MSI)-SENTINEL-2 imagery over inland and sea waters from SWIR bands
Remote Sens. Environ.
(2018) - et al.
Wind-induced hydrodynamic changes impact on sediment resuspension for large, shallow Lake Taihu, China
Int. J. Sediment Res.
(2019) - et al.
New spectrophotometric equations for determining chlorophylls a, b, c1 and c2 in higher plants, algae and natural phytoplankton
Biochemie und physiologie der pflanzen
(1975)