Cloud Adjacency Effects on Top-of-atmosphere Radiance and Ocean Color Data Products: A Statistical Assessment

Lian Feng, Chuanmin Hu

Research output: Contribution to journalArticlepeer-review

Abstract

<p> <p id="x-x-sp0095"> Ocean color measurements taken near cloud boundaries suffer from cloud adjacency effects (AEs). As a result, ~ 50% of the cloud-free ocean data are flagged as low quality. Quantitative assessment of such effects, as well as the methodology required to minimize, or correct for them, is rarely available. The goal of this study is to quantify such effects on top-of-atmosphere (TOA) radiance and ocean color data products for MODIS/Terra, MODIS/Aqua, and <a title="Learn more about Sea-Viewing Wide Field-of-View Sensor from ScienceDirect's AI-generated Topic Pages"> SeaWiFS </a> measurements. The AEs estimation was based on statistics and an objective method applied to carefully selected clear-water scenes (the number of cloud patches was &gt; 100,000 for each instrument) where ocean properties are relatively homogeneous, over both the North Atlantic and South Pacific. The AEs were quantified as the relative difference between the near-cloud pixels and pixels at least 20 km away from any cloud. Results show that the AEs on TOA radiance share similar patterns among the three missions. Specifically, the AEs decrease sharply as distance increases from cloud edges, and the AEs increase monotonically with increasing wavelengths because they were evaluated in relative rather than absolute terms. However, while discernable memory effects (MEs) are observed on cloud-adjacency pixels of both <a title="Learn more about MODIS from ScienceDirect's AI-generated Topic Pages"> MODIS </a> missions, they are insignificant on the SeaWiFS mission, and are found in measurements along the scan direction downstream of the clouds, representing &gt; 15% of the total AEs in TOA radiance. The AEs on the retrieved <a title="Learn more about Remote Sensing from ScienceDirect's AI-generated Topic Pages"> remote sensing </a> reflectance (R <sub> rs </sub> ) data products are different among the three missions possibly due to their differences in vicarious calibration and uncertainties in atmospheric correction, leading to different patterns in the chlorophyll-a (Chl-a) and normalized Florescence Line Height (nFLH) data products. Large AEs (&gt; 50%) are observed in nFLH of both MODIS/Terra and MODIS/Aqua, likely due to the opposite AEs on R <sub> rs </sub> between 667 and 678 nm. Finally, when the OCI Chl-a algorithm is used, the current MODIS stray-light masking window (7 &times; 5) used to mask the AE-contaminated pixels may be relaxed to 3 &times; 3 without sacrificing data quality, leading to &gt; 40% of the previously masked low-quality data being recovered for clear waters. </p></p>
Original languageAmerican English
JournalRemote Sensing of Environment
Volume174
DOIs
StatePublished - Jan 1 2016

Keywords

  • Cloud adjacency effects
  • Memory effects
  • Stray light
  • Ocean color
  • TOA radiance
  • Remote sensing reflectance (Rrs)
  • Chl-a
  • OCI
  • MODIS
  • SeaWiFS

Disciplines

  • Life Sciences

Cite this