Seafloor substrate (hard and soft bottom) maps derived from satellite imagery for the islands and atolls of the Pacific Remote Island Areas

Email preferred

via the Pacific Islands Benthic Habitat Mapping Center (PIBHMC) website

Email preferred

Email preferred

The geographic extent of Wake Atoll

Johnston Atoll, Baker Island, Howland Island, Jarvis Island, Kingman Reef, and Palmyra Atoll in Pacific Remote Island Areas. The geographic extent of Johnston Atoll (to the North), Howland and Baker (to the West), and Jarvis Island (to the South and East).

Time frame of WorldView 2 or 3 imagery for Howland, Baker, Jarvis, Johnston Atoll, Kingman Reef, Palmyra Atoll, Wake Atoll in Pacific Remote Islands

Newly developed Hard and Soft Bottom Seafloor Substrate Map of Baker Island, Pacific Remote Islands. The zip file contains the product in GeoTIFF format with the PDF map showing the extent of the product. The readme text file describing the detailed processing steps is also included. Source WorldView 2 satellite image: 2012 and WorldView 3 satellite images: 2016. The data are provided by the Pacific Islands Fisheries Science Center and hosted by UH on SOEST servers.

Newly developed Hard and Soft Bottom Seafloor Substrate Map of Howland Island, Pacific Remote Islands. The zip file contains the product in GeoTIFF format with the PDF map showing the extent of the product. The readme text file describing the detailed processing steps is also included. Source WorldView 2 satellite images: 2011 and 2016. The data are provided by the Pacific Islands Fisheries Science Center and hosted by UH on SOEST servers.

Newly developed Hard and Soft Bottom Seafloor Substrate Map of Jarvis Island, Pacific Remote Islands. The zip file contains the product in GeoTIFF format with the PDF map showing the extent of the product. The readme text file describing the detailed processing steps is also included. Source WorldView 2 satellite image: 2016. The data are provided by the Pacific Islands Fisheries Science Center and hosted by UH on SOEST servers.

Newly developed Hard and Soft Bottom Seafloor Substrate Map of Johnston Atoll, Pacific Remote Islands. The zip file contains the product in GeoTIFF format with the PDF map showing the extent of the product. The readme text file describing the detailed processing steps is also included. Source WorldView 2 satellite image: 2014. The data are provided by the Pacific Islands Fisheries Science Center and hosted by UH on SOEST servers.

Newly developed Hard and Soft Bottom Seafloor Substrate Map of Kingman Reef, Pacific Remote Islands. The zip file contains the product in GeoTIFF format with the PDF map showing the extent of the product. The readme text file describing the detailed processing steps is also included. Source WorldView 2 satellite image: 2015. The data are provided by the Pacific Islands Fisheries Science Center and hosted by UH on SOEST servers.

Newly developed Hard and Soft Bottom Seafloor Substrate Map of Palmyra Atoll, Pacific Remote Islands. The zip file contains the product in GeoTIFF format with the PDF map showing the extent of the product. The readme text file describing the detailed processing steps is also included. Source WorldView 2 satellite images: 2014. The data are provided by the Pacific Islands Fisheries Science Center and hosted by UH on SOEST servers.

Newly developed Hard and Soft Bottom Seafloor Substrate Map of Wake Atoll, Pacific Remote Islands. The zip file contains the product in GeoTIFF format with the PDF map showing the extent of the product. The readme text file describing the detailed processing steps is also included. Source WorldView satellite images: 2014 (WorldView 3) and 2017 (WorldView 2). The data are provided by the Pacific Islands Fisheries Science Center and hosted by UH on SOEST servers.

An example hard- and soft-bottom substrate map of Tutuila (prepared using a different methodology compared with the substrate maps prepared for the project described here).

NOAA ESD source data are accessible through this website.

Describes methods development and decision tree.

the UNESCO Digital Library website

The Coral Reef Conservation Program's (CRCP) information portal that provides access to NOAA coral reef information and data products with emphasis on the U.S. states, territories and remote island areas.

Worldview 2 or Worldview 3 satellite images are acquired through the NEXTVIEW Imagery License Agreement between DigitalGlobe, Inc. and National Geospatial Intelligence Agency (NGA). Under the agreement, the images are available for the use by the U.S. Government. Standard product level (2A) of multispectral and pancrhomatic images were acquired to produce hard soft maps. The images meeting the following criteria are carefully selected : cloud and shadow free, relatively small surf breaks, little sunglint on the water surface and little turbidity in the water.

The satellite images were preprocessed before generating the depth invariant indices using the tools available in ENVI 5.3. The images were first orthorectified using the RPC Orthorectification Workflow tool. The images are then radiometrically corrected and converted into the radiance using the Radiometric Calibration tool and Apply Gain and Offset tool. Radiometric calibration factors for each image are provided by DigitalGlobe in the associated metadata. The absolute calibration factors are available in the online documentation made available by DigitalGlobe.

The atmospheric correction steps described in Mumby and Edwards (2000) were applied to the radiance images remove some of the atmospheric influences on the images. The calculation is sequenced in ENVI 5.3/IDL 8.5.

Depth Invariant Index in Mumby and Edwards (2000) were derived for 10 band pairs of each image. The band pairs are: Coastal - Blue, Coastal - Green, Coastal - Yellow, Coastal - Red, Blue - Green, Blue - Yellow, Blue - Red, Green - Yellow, Green - Red, Yellow - Red. Following equation was applied to each band pair using using the Band Math tool in ENVI 5.3 to derive the Depth Invariant Index.

Depth Invariant Index of band pair i and j = Li -[(ki/kj)*Lj]

where

Li = Radiance or Reflectance values of band i

Lj = Radiance or Reflectance values of band j

ki/kj = Ratio of attenuation coefficients of band i and j

Ratio of attenuation coefficients were derived in Excel spreadsheet using the variance and covariance of the radiance values extracted over shallow and deep sandy bottoms of each image.

The results were exported as GeoTiff for the further analysis in ArcGIS.

The ISO Cluster Unsupervised Classification tool in ArcGIS 10.5.1 was employed to segment the depth invariant indices into multiple classes. Number of iterations were set to be 5, and minimum class size were set to be 3. Up to 15 classes were generated for each band pair of the depth invariant index.

Each band pair classification was visually examined by inspecting the cluster results against the substrate boundaries that can be identified in the original satellite images. The best band pair classification result detecting the substrate boundaries was chosen, and interpreters made an informed judgement in attributing the predicted classes into hard and soft. Where required, area-specific class attribution and minor manual editing were undertaken to remove inaccurate predictions in ArcGIS 10.5.1. Output file type is GeoTIFF.

Where multiple images are available, the same steps are applied to the second image, and the gaps of the classification result in the first image will be filled with the second image.