A Novel MERIS Algorithm to Derive Cyanobacterial Phycocyanin Pigment Concentrations in a Eutrophic Lake: Theoretical Basis and Practical Considerations

Lin Qi, Chuanmin Hu, Hongtao Duan, Jennifer Cannizzaro, Ronghua Ma

Research output: Contribution to journalArticlepeer-review

Abstract

<p> <p id="x-x-sp0005"> A novel algorithm is developed and validated to estimate phycocyanin (PC) pigment concentrations of the most common cyanobaterium, <em> <a title="Learn more about Microcystis from ScienceDirect's AI-generated Topic Pages"> Microcystis </a> </em> <em> aeruginosa </em> , in Taihu Lake (China's third largest freshwater lake) using data from the <a title="Learn more about MERIS from ScienceDirect's AI-generated Topic Pages"> Medium Resolution Imaging Spectrometer </a> (MERIS). First, a PC index (PCI) is defined as <a title="Learn more about Remote Sensing from ScienceDirect's AI-generated Topic Pages"> remote-sensing </a> reflectance (Rrs, sr <sup> &minus; 1 </sup> ) at 620 nm normalized against a baseline formed linearly between Rrs(560) and Rrs(665). Because PC has a local absorption peak at ~ 620 nm, PCI is shown in both theory and model simulations to exhibit a monotonic functional relationship with PC concentration. Field measurements also support this hypothesis, from which an empirical algorithm is developed to estimate PC concentrations between ~ 1 and 300 &mu;g/L with unbiased RMS uncertainties of 58%. <a title="Learn more about Radiative Transfer from ScienceDirect's AI-generated Topic Pages"> Radiative transfer </a> simulations further show that such a field-based algorithm can be applied directly to <a title="Learn more about MERIS from ScienceDirect's AI-generated Topic Pages"> MERIS </a> Rayleigh-corrected reflectance (Rrc) after adjusting algorithm coefficients. Spectral analyses and image comparisons indicate that such an algorithm is nearly immune to perturbations from thick aerosols, thin clouds, significant sun glint, and extreme water turbidity. Mean usable data coverage increases from &lt; 1% to &gt; 50% when using Rrc as compared to Rrs processed using standard atmospheric correction algorithms in SeaDAS. The robust algorithm performance and significantly improved data coverage lead to the establishment of a long-term (2002&ndash;2012) time-series of PC concentrations in Taihu Lake, which show both seasonal and inter-annual changes. Test of the algorithm over other lakes, including Dianchi Lake (a typical plateau lake of China), suggests that the same band-subtraction approach might be applicable to other <a title="Learn more about Inland Water from ScienceDirect's AI-generated Topic Pages"> inland water </a> bodies, although local bio-optical conditions due to different sediment and <a title="Learn more about Phytoplankton from ScienceDirect's AI-generated Topic Pages"> phytoplankton </a> compositions need to be considered when applying such an empirical approach. </p></p>
Original languageAmerican English
JournalRemote Sensing of Environment
Volume154
DOIs
StatePublished - Jan 1 2014

Keywords

  • Remote sensing
  • Phycocyanin
  • Microcystis aeruginosa
  • MERIS
  • Chlorophyll-a
  • Lake Taihu
  • Sun glint
  • Aerosols
  • Clouds

Disciplines

  • Life Sciences

Cite this