Data Set (knb-lter-mcr.5006.2)

MCR LTER: Coral Reef: Computer Vision: Moorea Labeled Corals

  Summary and Data Links People and Organizations Temporal, Geographic and Taxonomic Coverage Methods and Protocols  

These methods, instrumentation and/or protocols apply to all data in this dataset:

Protocols and/or Procedures
Description:

Photoquadrats

Quantitative high resolution digital images, each aproximately 2.5 to 3.2 Mb in JPEG format, with aproximately 2000 x 2000 pixels or 4 megapixels.

Parameters for Photoquadrat Analysis

CPCe software is used to quantify the area of benthic taxa using a matrix of 200 randomly spaced dots each 15 pixels in diameter. Image row and column pixel are counted from (1,1). Prior to the use of CPCe the area of benthic taxa was quantified by superimposing a matrix of 200 randomly located circles on the image and counting the circles landing on the substratum category of interest. Identification has been assigned by one field technician from 2005 through 2007 and another from 2008 through present. This dataset contains photoquadrats from 2008 through 2010 only.

Sampling design for quadrats on the reef

The sampling strategy was designed to facilitate tests of the effect of time and shore on the coral community structure of the fringing, and outer reef habitats of Moorea. For the purpose of this analysis, fringing communities were defined as reefs adjacent to the shore and within ca. 50 m of the land. Outer reef habitats were located on the outer reef, and were censused at 10 m and 17 m depth in order to sample habitats that are representative of the outer reefs of Moorea, and within the depth range for tractable to diving research. A nested, factorial ANOVA experimental design was employed in all sampling strategies. Shore was treated as a fixed factor, site as a random factor nested within shore, and time was a repeated measures factor based on individual quadrats that could be relocated (at the fringe and outer sites), or discrete patches of reef that could be revisited (in the lagoon). In the case of the outer reef, depth also was used as a fixed factor in order to contrast the effects of time and shore between depths. All census methods were designed to quantify coral community structure in terms of the dominant constituents of the benthic community - scleractinian corals, macroalgae, crustose coralline algae, algal turf and, in some cases sand - and for both the fringing and outer reef habitats, allowed the scleractinian community structure to be resolved to genus level. Methodological problems (described below) in the lagoon habitat prevented genus-level resolution of the scleractinians, but such resolution should be attainable in Year II. In addition to establishing an orthogonal contrast of coral community structure, a subset of the sites and habitats were selected for more detailed analyses of the population density of coral recruits and juvenile corals, as well as the demographic analysis of selected coral species. The time-consuming nature of these analyses prevented them from being measured in all site/habitat combinations.

Quality Control issues in benthic community analysis

This data set is based on photographic images recorded using a 0.5 x 0. 5 m photoquadrat and a high resolution SLR digital camera. For the outer reef and fringing reef sampling, photoquadrats are repositioned each year at the same spot along the transect, although the corners of the quadrat are not marked. The photoquadrat is placed with a high degree of precision each year, but currents and waves can result in repositioning errors. To screen for these effects, matched photographs in consecutive years are evaluated using a software routine for changes in coral cover that exceed 20%; such changes are unlikely to occur through biological processes (although they could arise from importation of colonies following storms). All photographs at each site are compared with matched images in the following year, and discrepancies of >20% coral cover are flagged for investigation. Typically <1% of the images require further investigation each year.

Description:

Reference for taxonomy

Veron JEN (2000) Corals of the World - Volumes 1, 2, 3. Australian Institute of Marine Science, Townsville, Australia. 1382pp

Description:

Reference for CPCe software

Kohler, K.E. and S.M. Gill, 2006. Coral Point Count with Excel extensions (CPCe): A Visual Basic program for the determination of coral and substrate coverage using random point count methodology. Computers and Geosciences, Vol. 32, No. 9, pp. 1259-1269, DOI:10.1016/j.cageo.2005.11.009.

Protocol:
Author: Kevin E. Kohler
Title: Coral Point Count with Excel extensions (CPCe)
Available Online:
 View: http://www.nova.edu/ocean/cpce/
Description:

Reference for Computer Vision Methods software

Beijbom O., Edmunds P.J., Kline D.I., Mitchell G.B., Kriegman D., 'Automated Annotation of Coral Reef Survey Images', IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Providence, Rhode Island, 2012. [BibTex] [pdf]

Protocol:
Author: Oscar Beijbom
Title: Moorea Labeled Corals software development kit
Available Online:
 View: http://vision.ucsd.edu/~beijbom/moorea_labeled_corals/devkit.zip