Multiplatform Remote Sensing for Coral Reef Community Assessment Quinta Reunión Nacional de Percepción Remota y Sistemas de Información Geográfica en Puerto Rico September 27, 2007 Roy A. Armstrong, Ph. D. Bio-Optical Oceanography Laboratory University of Puerto Rico, Mayaguez Campus
Bathymetry of Puerto Rico -Virgin Islands Geological Platform Potential Reef Habitat
Remote Sensing Platforms and Sensor Requirements for Benthic Habitat Assessments Insular Shelf (0-30 m) Satellite Sensors Landsat TM, SPOT, IKONOS Airborne Sensors AVIRIS, AISA, Digital Camera Systems Upper Insular Slope (30 100 m) Ship-based Acoustic Sensors Autonomous Underwater Vehicles (AUV) Optical and Acoustic Imaging Remote Operated Vehicles (ROV) Optical Imaging
Direct vs. Indirect Monitoring of Coral Reefs Using Remote Sensing Direct includes benthic mapping and characterization of coral reef and other biotopes using a sensor down approach. Requires local knowledge of reef communities and utilizes image-specific statistics to drive a supervised classification. Requires atmospheric and water column corrections for multi-temporal analysis. Landsat TM data for Los Roques, Venezuela: Benthic Habitat Map Seagrass Biomass Map
Direct vs. Indirect Monitoring of Coral Reefs Using Remote Sensing, cont. Indirect addresses the oceanic environment around the reef. Sea surface temperatures (SSTs) and bleaching events. Low salinities and high turbidity from episodic rainfall events. Water optical properties (AOP and IOP) and coral reef parameters. MODIS K490 November 20, 2003
MODIS Time Series of K490 and Chlorophyll Associated with the November 13-15, 2003 Episodic Rainfall Event 0.40 K 490 (m -1 ) 0.35 0.30 0.25 0.20 0.15 298 305 323 324 327 330 333 0.10 0.05 0.00 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 12 10 Chlorophyll (mg m 3 ) 8 6 4 2 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Distance (km) Julian days 317-319 Rainfall Event
Indirect Methods: Kd (PAR) vs. Percent Coral Cover Irradiance Profiles for Sampling Stations E d (μmol s-1 m-2) 0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 3 5 8 K d vs Percent Coral Cover Depth (m) 10 13 15 18 20 Las Coronas Media Luna Puerto Canoa Puerto de Botes Turmalin Oceanica Manchas Atunera South of Margarita Reef Shelf Edge South of Turrumote 60.00 50.00 K d :0.148 y = 585.4e -16.203x R 2 = 0.9203 40.00 K d 0.201 % CC 30.00 20.00 K d 0.267 10.00 0.00 0.13 0.18 0.23 0.28 0.33 K d
Reef-Up Habitat Mapping Physics based, uses Radiative Transfer Theory Uses spectral libraries of the reef components Its application is independent of site and image-specific statistics. Reflectance (%) 60 40 20 Spectral Reflectance of Nine Scleractinian Species Prp Acc Aga Mna Dis Fav Pra Dps Dpc 0 400 450 500 550 600 650 700 Wavelength (nm)
Spectral Library of Coral Reef Benthic Components Spectral Reflectance Data using GER-1500 Spectroradiometer
Coral Reef Benthic Components Coral Rubble Acropora cervicornis Montastraea annularis Gorgonians Siderastrea siderea Porites porites Dictyota spp. Thalassia testudinum
Coral Reef Bleaching
Coral Bleaching Spectral Response
Spatial Resolution Requirements for Coral Bleaching Detection Landsat 30 m IKONOS 1m 1 m SPOT 20 m CASI 5 m
Sub-meter Digital Camera Imagery NASA s Cirrus Digital Camera System (DCS) Hasselblad camera body, Kodak camera back and CCD array 40 70 cm resolution from 12,000 Fireball Information Technologies DCS 16 Mpixel camera 7 cm at 3,000 NASA DCS - St. Croix, USVI FIT DCS La Parguera
AVIRIS 2005 Hyperspectral Mission Airborne Visible-Infrared Imaging Spectrometer (AVIRIS) 224 spectral bands 224 Spectral Bands, 400-2500 nm Spectral Resolution 10 nm Twin Otter Aircraft Puerto Rico Flighlines December 12-13, 2005 Altitude 3.5 km Pixel Size ~3.5 m Flight lines for La Parguera, Southwestern PR 26 Flighlines Solar azimuth lines
Atmospherically-Corrected AVIRIS Image
Water Column Corrections Lyzenga s depth invariant bands where Benthic Reflectance Model Empirical Line Method Relies on field measurements of homogeneous area or flat fields DN(b) = ρ(b)a(b) + B(b) ρ = reflectance of surface material A = multiplicative term, B = additive term
Underwater Flat-Field Calibration Targets for Water Column Correction
Water Column Correction Validation AVIRIS agrees with field values within 10% from 400-600 nm and up to 18% between 600-700 nm. Spectral features are preserved, mostly corresponding to pigment absorption by microbial layers.
AVIRIS Benthic Spectra After Water Column Correction Sand Seagrass Coral
Acknowledgements Fernando Gilbes, Liane Guild,Yasmin Detrés, Juan Torres, and Maria Cardona Sponsored by: NOAA (DNER) Puerto Rico Coral Reef Monitoring Grant NASA Ocean Biogeochemistry Program Center for Subsurface Sensing and Imaging (CenSSIS) NSF