Liane Guild, Brad Lobitz, Randy Berthold, Jeremy Kerr

1
Imaging Spectroscopy and Field Data in Support of
Coral Reef Biodiversity and Change Research in
Puerto Rico
Liane Guild, Brad Lobitz, Randy Berthold, Jeremy
Kerr Biospheric Science Branch, NASA Ames
Research Center, CA Roy Armstrong, James
Goodman, Yasmine Detres, Fernando Gilbes, Stacey
Williams University of Puerto Rico at Mayaguez,
PR Collaborators AVIRIS Team at NASA
JPL Airborne Science and Technology Lab at NASA
Ames Fireball Information Technologies Funding
NASA Interdisciplinary Research in Earth
Science NASA Ocean Biology and Biogeochemistry
Program Contact Liane.S.Guild_at_nasa.gov http//
earthscience.arc.nasa.gov/sge/coral-health/
2
Science Objectives

• Use NASAs 2005 AVIRIS imagery over La Parguera,
  PR to map coral reef benthic types and
  variability
• Establish permanent reef monitoring transects to
  track change annually
• Use AVIRIS data and field measurements to assess
  the ecological impact of the 2005 coral bleaching
  event in Southwestern Puerto Rico
• Hurricane Dean in August 2007
• Improve the interpretation of reef habitat
  variability and biodiversity from imaging
  spectroscopy (hyperspectral) data
• Processing algorithms could be used for future
  HyspIRI data

3
Climate Change Threats Increased Sea
Temperatures

• Due to sustained high sea temperatures,
  zooxanthellae (photosynthetic algae of coral) are
  expelled, leaving the coral bleached.
• Coral species of interest Acroporids (elkhorn
  and staghorn coral) were listed as threatened in
  2006 under the US Endangered Species Act (ESA).
• This year 7 additional
• Caribbean coral reef
• species (75 Pacific spp)
• are under review for
• listing all found in PR

Bleached A. palmata
Healthy A. cervicornis
Healthy Acropora palmata
Mostly dead A. cervicornis
4
NOAA Coral Reef Watch SST Degree Heating Weeks
Oct. 22, 2005
5
NASA Airborne Missions of Coral Reefs Using
AVIRIS and DCS

• AVIRIS (Airborne Visible Infrared Imaging
  Spectrometer)
• Operated by NASA Jet Propulsion Lab
• NASA Wallops Twin Otter
• 224 contiguous spectral channels (370-2500 nm)
• Visible range 410-700 nm (31 bands)
• 10 nm nominal channel bandwidth
• High signal to noise ratio
• Altitude 3.5 km (Twin Otter)
• AVIRIS spatial resolution 3 m (Twin Otter),
  16 m (ER-2)
• Cirrus and Kodak DCS (Digital Camera System) on
  low altitude aircraft
• Cirrus Operated by UC Santa Cruz Airborne
  Sensor Facility at NASA Ames
• Kodak Operated by NASA Ames and Fireball
  Information Technologies
• Spatial resolution lt1 m

6
NASA Airborne Mission of the 2005 Caribbean Coral
Reef Bleaching Event Using AVIRIS and DCS
AVIRIS flight lines over La Parguera, PR

• December 12-13, 2005
• Twin Otter Platform, Altitude 3.5 km
• Pixel Size 3 m

7
Kodak DCS Imagery of San Cristobal, La Parguera
30 cm resolution
AVIRIS
3 m resolution
ETM
30 m resolution
8
Field Measurements

• Airborne reflectance spectra are composites of
  reflectance from
• the atmosphere (air molecules, particulates,
  aerosols)
• the sea surface
• the water column (water molecules, phytoplankton,
  colored dissolved organic matter) and
• the bottom (sediments, seagrass, corals).
• Atmospheric constituents are measured with
  sunphotometers.
• Water column optical properties are measured with
  spectrophotometers.
• Spectroradiometers are used to measure light
  reflected from the sea surface, the coral reef
  bottom types, and within the water column.

1
radiance
2
3
water optics spectrophotometer
Spectroradiometer field spectra
4
9
Field Measurements

• Reef benthic type spectra
• Photogrids
• GPS
• Spectra of calibration targets
• Aeronet station AOD atmospheric correction
• Water optical profiler
• Upwelling radiance, Lu
• Downwelling irradiance, Ed
• Surface Ed, Es

10
Coral Reef Ecosystem Benthic Types
Coral Rubble
Acropora cervicornis
Acropora palmata
Montastraea annularis
Siderastrea siderea
Porites spp.
Thalassia testudinum
Dictyota spp.
Gorgonians
11
Field Spectra and Spectral Resolution
12
Field Spectra and Spectral Resolution

• Spectra normalized to unity at 550nm show
  differences in spectral shape between benthic
  cover types
• Live coral has a broad peak between green and
  red wavelength
• Most features show a chlorophyll absorption
  feature at 680nm

13
AVIRIS Dec. 2005 (3 m Resolution) Mosaic of La
Parguera, PR
14
LiDAR Bathymetry and Field Measurement Locations

• Airborne data collection 2006
• Bathymetry 50m
• 4m spatial resolution

15
J. Goodman (UPRM)
2005 AVIRIS Preprocessing and Inversion Model
Image Preprocessing Correct Stray Light
Anomaly Hedley Glint Removal Tafkaa Atmospheric
Correction
Inversion Output Water Properties
Spectral Input Parameters Aquatic Absorption
Properties Generic Bottom Reflectance
Inversion Model
Bathymetry
Bottom Albedo (550 nm)
16
J. Goodman (UPRM)
2005 AVIRIS Unmixing Model
Spectral Endmembers
Unmixing Output
Benthic Composition
Forward Model
Sand
Unmixing Model
Coral
Algae/ Seagrass
17
2005 AVIRIS From Imagery to Information
Preprocessed AVIRIS Image
Inversion and Unmixing
Hydrolight Transformation
Unmixing End-Member-Index Classification
Preprocessed Spectral Classification
Hydrolight Spectral Classification
18
Acropora cervicornis (staghorn coral) Transects
2004-2007
2005
2004
2006
2007
19
Acropora cervicornis (staghorn coral) Transects
2008
2008
2009
20
Reef Bottom Type Biodiversity (Percent Cover)
Coral Point Count with Excel Extensions (CPCe)
21
Reef Bottom Type Biodiversity (Percent Cover)

• Coral and gorgonian cover was not significantly
  different from 2007 to 2009
• Sponge cover has increased
• Macroalgae increased in 2008 then decreased in
  2009
• Sand cover decreased from 2007 to 2008 and 2009
  level
• Recovery possibly beginning in benthic
  communities at some sites

• Results are similar to other studies following
  disturbances in coral reefs
• Need at least two more years of monitoring to
  confirm occurrence of recovery

22
Summary and Next Steps
http//earthscience.arc.nasa.gov/sge/coral-health/

• Generalized spectral categories (coral, sand, and
  SAV) used as inputs to the models still allowed
  for detailed output classes (ecological zones)
• Use detailed benthic spectral library for
  unmixing end members
• Compare spectral transects with
  Hydrolight-transformed AVIRIS data
• Substantial image preprocessing may have affected
  the modeling results
• Biodiversity analysis, based on percent cover
  within photogrids acquired in 2007-2009, showed
  recovery possibly beginning in benthic
  communities at some sites
• Complete photogrid analysis for other patch reefs
  and relate 2005 transects with AVIRIS
  classifications
• Complete assessment of changes in biodiversity,
  perhaps related to recovery from 2005 bleaching
  event and 2007 hurricane
• Critical need for hyperspectral data for studies
  in shallow water
• Procedures relevant to HyspIRI

23
Remote Sensing Challenges Spatial Resolution
Montastraea annularis (pale) following 2005
bleaching
24
Background

• Coral reefs are in peril
• rapidly growing local populations
• regional to global atmospheric, oceanographic and
  climatic changes
• increased sea temperatures, rising CO2 and
  falling pH (ocean acidification)
• increased frequencies, intensities and
  distributions of bleaching, disease and other
  disturbances
• deteriorating biologically and physically
• loss of hard corals, increased abundance of
  algae, replacement of corals by algae, predation
  and microbial outbreaks
• 2 species listed as threatened under the US
  Endangered Species Act (2006) and 83 additional
  species are proposed for listing (2009) with
  final determination in 2010

25
Field Data Collection
Cirrus DCS Imagery (lt 1m resolution) Buck Island,
St. Croix, USVI December 16, 2005 Calibration
tarp and boat are visible!
26
US Coral Reef Task Force NASA Airborne Mission
of the 2005 Caribbean Coral Reef Bleaching Event
Using AVIRIS and DCS

• Airborne Mission 2005
• Culebra, PR, Dec. 12
• La Parguera, PR, Dec. 12-13
• Buck Island, St. Croix, Dec. 15-16
• St. John, Dec. 19
• Vieques Is., PR, Dec. 20
• Research Objective
• Assess the ecological impact of coral
  bleaching in Puerto Rico and the USVI

DCS imagery of Buck Island, USVI, December 2005
This collaborative effort was in support of the
US Coral Reef Task Forces (USCRTF) November 2005
resolution to mobilize and assess coral reef
bleaching, coral mortality, coral recovery, and
the ecological impact of the September-October
2005 coral bleaching event in Puerto Rico (PR)
and the US Virgin Islands (USVI).
27
Water Column Spectral Attenuation
Spectral attenuation coefficients (Kd) can be
determined by using Satlantic Hyperpro
downwelling irradiance (Ed) profiles. These
measurements can be expressed as a function of
depth.
28
(No Transcript)
29
J. Goodman (UPRM)
2005 AVIRIS From Imagery to Information
Raw Image
Bottom Albedo
Benthic Composition
Bathymetry
30
J. Goodman (UPRM)
2005 AVIRIS From Imagery to Information
Sand
Algae/ Seagrass
Coral
AVIRIS Data
Field Data DCS Images
Benthic Composition
31

• 89 categories
• Corals 52 species
• Macroalgae 13 genera
• Gorgonians 11 genera
• Dead coral with algae (4) Zoanthids (2)
  Coralline algae (1) Sponges (1) Other live (1)
  Diseased corals (1)