Abstract
The Operational Land Imager (OLI) onboard Landsat 8 has found successful application in inland and coastal water remote sensing. Its radiometric specification and high spatial resolution allows quantification of water-leaving radiance while resolving small water bodies. However, its limited multispectral band set restricts the range of water quality parameters that can be retrieved. Identification of cyanobacteria biomass has been demonstrated for sensors with a band centered near 620 nm, the absorption peak of the diagnostic pigment phycocyanin. While OLI lacks such a band in the orange region, superposition of the available multispectral and panchromatic bands suggests that it can be calculated by a scaled difference. A set of 428 in situ spectra acquired in diverse lakes in Belgium and The Netherlands was used to develop and test an orange contra-band retrieval algorithm, achieving a mean absolute percentage error of 5.39% and a bias of −0.88% in the presence of sensor noise. Atmospheric compensation error propagated to the orange contra-band was observed to maintain about the same magnitude (13% higher) observed for the red band and thus results in minimal additional effects for possible base line subtraction or band ratio algorithms for phycocyanin estimation. Generality of the algorithm for different reflectance shapes was tested against a set of published average coastal and inland Optical Water Types, showing robust retrieval for all but relatively clear water types (Secchi disk depth > 6 m and chlorophyll a < 1.6 mg m − 3 ). The algorithm was further validated with 79 matchups against the Ocean and Land Colour Imager (OLCI) orange band for 10 globally distributed lakes. The retrieved band is shown to convey information independent from the adjacent bands under variable phycocyanin concentrations. An example application using Landsat 8 imagery is provided for a known cyanobacterial bloom in Lake Erie, US. The method is distributed in the ACOLITE atmospheric correction code. The contra-band approach is generic and can be applied to other sensors with overlapping bands. Recommendations are also provided for development of future sensors with broad spectral bands with the objective to maximize the accuracy of possible spectral enhancements.
Highlights
The typical spatial resolution of 250 m to 1 km offered by dedicated polar orbiting Ocean Colour (OC) sensors is often insufficient to fully resolve processes occurring in coastal, transitional and inland waters [1,2,3]
For the Pan and green bands slope coefficients, coefficient of variation (CV) were < 6.5%, while it reached 36% for the red band. This dependency is expected since the dataset was acquired over waters with variable levels of PC and Chl a, which share signals between the orange and red bands, providing variable relations depending on the concentration range of each random subsample of the dataset
Greater variability was observed for the bias with a CV of 52%, which was of little practical consequence due to its small magnitude
Summary
The typical spatial resolution of 250 m to 1 km offered by dedicated polar orbiting Ocean Colour (OC) sensors is often insufficient to fully resolve processes occurring in coastal, transitional and inland waters [1,2,3]. The requirement of higher spatial resolution has led to the development of water applications for sensors originally designed for terrestrial remote sensing [6,7,8], the most relevant of which are the Landsat and Sentinel-2 series The sensors on those platforms offer global coverage at high spatial resolution (10–60 m) and feature near infrared (NIR) and shortwave infrared (SWIR) bands appropriate for atmospheric correction over turbid waters [7,9]. The Operational Land Imager (OLI) on Landsat 8 resolves four visible bands covering the deep-blue (435–451 nm), blue (452–512 nm), green (533–590 nm) and red (636–673 nm) and has found successful application for remote sensing of water quality (e.g., transparency, dissolved organic carbon and algal biomass [10,11,12]) This waveband configuration, does not allow retrieval of direct information related to phycocyanin (PC) abundance, a diagnostic pigment for the presence of cyanobacteria which are a major concern in eutrophic inland waters. High spatial resolution orange bands are only available on commercial sensors, notably the imagers on WorldView-2 and -3 (DigitalGlobe) and on the SuperDove constellation (Planet Labs Inc.)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
