Frontiers in Marine Science (Oct 2017)

Autonomous Coral Reef Survey in Support of Remote Sensing

  • Steven G. Ackleson,
  • Joseph P. Smith,
  • Luis M. Rodriguez,
  • Wesley J. Moses,
  • Brandon J. Russell

DOI
https://doi.org/10.3389/fmars.2017.00325
Journal volume & issue
Vol. 4

Abstract

Read online

An autonomous surface vehicle instrumented with optical and acoustical sensors was deployed in Kane'ohe Bay, HI, U.S.A., to provide high-resolution, in situ observations of coral reef reflectance with minimal human presence. The data represented a wide range in bottom type, water depth, and illumination and supported more thorough investigations of remote sensing methods for identifying and mapping shallow reef features. The in situ data were used to compute spectral bottom reflectance and remote sensing reflectance, Rrs,λ, as a function of water depth and benthic features. The signals were used to distinguish between live coral and uncolonized sediment within the depth range of the measurements (2.5–5 m). In situRrs, λ were found to compare well with remotely sensed measurements from an imaging spectrometer, the Airborne Visible and Infrared Imaging Spectrometer (AVIRIS), deployed on an aircraft at high altitude. Cloud cover and in situ sensor orientation were found to have minimal impact on in situRrs, λ, suggesting that valid reflectance data may be collected using autonomous surveys even when atmospheric conditions are not favorable for remote sensing operations. The use of reflectance in the red and near infrared portions of the spectrum, expressed as the red edge height, REHλ, was investigated for detecting live aquatic vegetative biomass, including coral symbionts and turf algae. The REHλ signal from live coral was detected in Kane'ohe Bay to a depth of approximately 4 m with in situ measurements. A remote sensing algorithm based on the REHλ signal was defined and applied to AVIRIS imagery of the entire bay and was found to reveal areas of shallow, dense coral and algal cover. The peak wavelength of REHλ decreased with increasing water depth, indicating that a more complete examination of the red edge signal may potentially yield a remote sensing approach to simultaneously estimate vegetative biomass and bathymetry in shallow water.

Keywords