All thumbnails are clickable.
to Coral Reef Health: Spectral Analysis and Radiative Transfer Modeling
of Coral Reef Ecosystem Health
Liane Guild1, Barry Ganapol2,
Philip Kramer3, Roy Armstrong4,
Art Gleason3, Juan Torres4,
Lee Johnson5, Toby Garfield6,
and Brian So7
Ames Research Center, 2University of Arizona, 3University
of Miami, 4University of Puerto Rico, 5California
State University Monterey Bay, 6San Francisco State University,
and 7Lowell High School, San Francisco
| Study Site |Radiative Transfer
Modeling | Operating version of coral RT model
|Next Phase of RT Modeling | Remote
Sensing | Publications
and Press Releases | Related Links
An important contribution
to coral reef research is to improve spectral distinction between various
health states of coral species in areas subject to harmful anthropogenic
activity and climate change. New insights into radiative transfer properties
of corals under healthy and stressed conditions can advance understanding
of ecological processes on reefs and allow better assessments of the
impacts of large-scale bleaching and disease events. The goal of the
proposed research is to quantify the relationship between the coral
reef ecosystem optical properties and coral reef ecosystem health using
new remote sensing capabilities. Our objective is to examine the spectral
and spatial properties of hyperspectral sensors that may be used to
remotely sense changes in reef community health. In situ reef environment
spectra (healthy coral, stressed coral, dead coral, algae, and sand)
are input into our coral radiative transfer (RT) model (currently under
development) to distinguish important spectral characteristics of corals.
The RT model is being developed to identify the optical properties indicative
of stress exhibited by coral that could lead towards better ecological
assessment, extent, and forecasting of detrimental health conditions
of the reef environment. Further, the RT modeling will be linked to
remote sensing data to better predict coral reef health in remote areas
where only remote sensing data is available.
- Coral reefs
are important for ocean productivity and C cycling and provide coastline
protection and nursery grounds for marine fisheries.
- World's reefs
have shrunk to 1/2 - 1/10 of original size due to climate and human
- Coral reefs
respond immediately to environmental changes and are considered canaries
of the ocean.
the relationship between the coral reef ecosystem optical properties
and coral reef health.
sensor capabilities to remotely sense coral reef health.
representative spectra of healthy, diseased, and dead coral, and
dominant algal species at the level of airborne sensors.
1. 2000 IKONOS image of Andros Island, Bahamas and Long Rock study
In situ spectra
were collected in July and August 2002 at the Long Rock site in the
Andros Island, Bahamas coastal zone coral reef (Figure 1). Fieldwork
was performed in coordination with the Navy's Atlantic Undersea Test
and Evaluation Center (AUTEC) on Andros Island at the Long Rock study
site where University of Miami investigators have extensive ground-truth
data and hyperspectral data collected with the Portable Hyperspectral
Imager for Low Light Spectroscopy (PHILLS) sensor. In situ spectra
collection (Figure 2) included spectra of healthy, unhealthy, and dead
coral, substrate (sand), seagrass, and algae collected by a hand-held
portable spectroradiometer (GER 1500) incased in underwater housing
(Underwater Video Vault). Additionally, spectral measurements over a
range of water depths (Figure 3), relief, and bottom types are compared
to help quantify bottom-water column influences. Downwelling irradiance
and upwelling radiance were measured simultaneously by a Tethered Spectral
Radiometer Buoy (TSRB) instrument and water column optical properties
were measured using the AC-9 instrument. Further, digital camera and
video camera images were collected.
2. Collection of in situ spectra using the portable spectroradiometer.
3. Collection of spectra at depth. Spectralon panel at bottom is used
as a reference standard.
Our primary emphasis
is on Acropora palmata (or elkhorn coral, Figure 4), a major
reef building coral, which is prevalent in the study area, but is suffering
from white band disease (Figure 5). A. palmata is currently being
proposed as an endangered species because its populations have severely
declined in many areas of the Caribbean. In addition to the A. palmata
colonies, we have collected spectra of at least seven other coral colonies
that exist within the study area, each with different coral community
composition, density of corals, relief, and size of corals. We collected
nearly 1,300 spectra including species spectral library and target mixtures
at various depths. Coral spectral reflectance will be input into the
RT model, to provide the absorption spectrum for distinguishing spectral
characteristics of coral health.
4. Acropora palmata with white band disease.
Figure 5. White
band disease on Acropora palmata.
In situ measurements
address basic questions about the interactions of downwelling radiation
with coral reef substrates as well as provide the necessary ground-truth
data to calibrate airborne hyperspectral images. We focused on both
small-scale (cm) species-specific reflectance measurements (spectral
library) as well as intermediate scale (meter) reflectance measurements
at distance from target for mixed spectra of entire micro-habitats.
All field aspects included optical properties (AC-9 instrument) (Figure
6) and light field parameters (TSRB instrument) (Figure 7) so that in
situ measurements of coral reef benthos can be used to test algorithms
that correct for water column and atmospheric properties.
Three coral reef habitats
were examined: shallow reef crest (<3m water depth) dominated by
A. palmata, intermediate depth fore reef zone (7-12 m water depth)
dominated by Montastraea annularis, and deep (20-30 m) marginal
reef dominated by Montasraea franksi.
6. AC-9 instrument.
7. TSRB instrument.
Radiative Transfer (RT)
reflectance spectra are composites of reflectance:
bottom (sediments, seagrass, coral).
light scattering in the coral reef community including scattering
in the water and atmosphere.
Back to Top
version of coral RT model:
reflectance, calculated from in situ coral spectra, is input into a RT
model, CorMOD2 (CM2) (Figure 8), based on a leaf radiative transfer model.
In CM2, input coral reflectance measurements produce modeled reflectance through
an inversion at each visible wavelength to provide the absorption spectrum
Initially, we imposed a scattering baseline that is the same regardless of
the coral condition and that coral is optically thick and no light is transmitted
CorMOD2 coral specific RT model.
9. Absorption profiles output from the coral radiative transfer model.
Healthy Acropora palmata and A. palmata with white band
disease (WBD) profiles are distinct.
2 for coral RT model: Is this
spectral distinction of coral health apparent above the water column and
atmosphere? We will integrate spectra over hyperspectral sensor channels.
phase of RT modeling
is specified from transmission measurements from coral surfaces.
Absorption profile is constructed from individual absorbing coral components
and coral biochemical concentrations.
RT between coral surfaces characterized by non-rotationally invariant
Need orientation of coral surfaces (CAD) like a leaf angle distribution.
Need a coral surface area per unit bottom area (CAI) comparable to
a leaf area index.
Determine the appropriate substrate and diffuse photon source from
the water column.
the appropriate spatial scale for detecting coral features and health
characteristics in hyperspectral data.
Combination 54 (647.5 nm), 37 (565.9 nm), 12 (445.3 nm).
Hyperspectral Imager for Low Light Spectroscopy (PHILLS) data is under
investigation for distinction of coral reef features and variation in
spectral characteristics for healthy and unhealthy coral. The PHILLS sensor
has 128 channels with 65 bands in the visible range
Spatial scale degradation of hyperspectral data will be applied for
assessment of sensor resolution and distinction of reef features.
Back to Top
Guild, Liane, Toby Garfield, and Barry Ganapol, Lee Johnson, Roy Armstrong,
and Philip Kramer, 2002, Clues to Coral Reef Health: Spatial and Spectral
Remote Sensing Detail, International Conference on Remote Sensing of
Marine and Coastal Environments, Miami, May 2002.
Guild, Liane, Barry Ganapol,
Philip Kramer, Roy Armstrong, Art Gleason, Juan Torres, Lee Johnson, and Toby
Garfield, Clues to coral reef health: integrating radiative transfer modeling
and hyperspectral data, Eos. Trans. AGU 83(47), Fall Meet. Suppl., Abstract
Guild, L., 2002. NASA Devising Method to Remotely Monitor Ocean
Environment, NASA News Release: 02-127AR, Dec. 6.
2002-03 NASA Ames Research Center Directors Discretionary Fund
by: Juan Torres, Art Gleason, Liane Guild.
Back to Top
Press Release: NASA devising method to remotely monitor ocean environment.
2003 Deep Coral field trip to USVI and Puerto Rico - San
Juan Star Press Release: includes two picture!
Back to Top
Liane Guild, Ph.D. firstname.lastname@example.org
MS 242-4, NASA Ames Research Center
Moffett Field, CA 94035 USA
Page last updated: July 18, 2003.