Hyperspectral sensing of disease stress in the Caribbean reef-building coral, Orbicella faveolata - perspectives for the field of coral disease monitoring

Underwater spectral reflectance measurements: the reflectance standard submersion factor and its impact on derived target reflectance

For Earth observation remote sensing, high quality reflectance spectra are necessary for model input, algorithm development, and validation of derived products. In the aquatic environment, a common approach for making spectral reflectance measurements involves using a calibrated reflectance standard such as a Spectralon plaque underwater. The manufacturer provides a National Institute of Standards and Technology traceable reflectance curve with each standard, measured in air. Here, we demonstrate how the reflectance factor changes when submerged in water based on the standard albedo and viewing geometry. Target reflectances calculated incorrectly with the air calibrated values are 10%-60% lower than those estimated with submerged plaque calibrations. We provide guidelines for proper use and calibration of standards underwater.

Microbial Contamination Survey of Environmental Fresh and Saltwater Resources of Upolu Island, Samoa

Analysis of microbe diversity in freshwater resources and nearshore seawater samples of Upolu Island was performed to investigate the distribution of harmful bacteria. For this, 124 samples were collected from 23 river systems, two volcanic lakes, and 45 locations inside and outside the barrier reef of Upolu Island, Samoa. Physicochemical parameters for general water quality, detection of coliform bacteria and 16S rRNA amplicon sequencing were performed on all samples. Fecal indicator bacteria (FIB) testing indicated a wide distribution of coliform bacteria in all sampled freshwater sites with evidence of fecal coliform in most locations. Importantly, evidence of coliform bacteria was found in most seawater samples inside and along the reef, apart from those samples taken ~20 km offshore. Illumina 16S rRNA amplicon sequencing of the V4 hypervariable region confirmed the presence of various types of harmful bacterial species, namely from the Enterobacteriaceae, Enterococcae, Streptococcaceae, and Vibrionacea families. By combining the sensitivity of FIB testing and next-generation sequencing, we were able to show the extent of potential contaminations in fresh and seawater samples and simultaneously identify the potential pathogenic bacterial genera present. The wide distribution of potential harmful bacteria from river runoff or direct sewage dumping has an impact on human health, leading to many skin and intestinal diseases, and is potentially detrimental to coral reef community health.

UAVs, Hyperspectral Remote Sensing, and Machine Learning Revolutionizing Reef Monitoring

Recent advances in unmanned aerial system (UAS) sensed imagery, sensor quality/size, and geospatial image processing can enable UASs to rapidly and continually monitor coral reefs, to determine the type of coral and signs of coral bleaching. This paper describes an unmanned aerial vehicle (UAV) remote sensing methodology to increase the efficiency and accuracy of existing surveillance practices. The methodology uses a UAV integrated with advanced digital hyperspectral, ultra HD colour (RGB) sensors, and machine learning algorithms. This paper describes the combination of airborne RGB and hyperspectral imagery with in-water survey data of several types in-water survey of coral under diverse levels of bleaching. The paper also describes the technology used, the sensors, the UAS, the flight operations, the processing workflow of the datasets, the methods for combining multiple airborne and in-water datasets, and finally presents relevant results of material classification. The development of the methodology for the collection and analysis of airborne hyperspectral and RGB imagery would provide coral reef researchers, other scientists, and UAV practitioners with reliable data collection protocols and faster processing techniques to achieve remote sensing objectives.

Identification of a consistent polyene component of purple pigment in diseased sclerites of Caribbean corals across region, species, and insult agent

Gorgonians respond to insult (damage and disease) by producing sclerites containing a purple pigment as opposed to the normal white sclerites. Raman microscopy is used to study the purple areas of three species of diseased coral, Gorgonia ventalina, Pseudoplexaura porosa, and Eunicea laciniata obtained from Puerto Rico. These spectra were compared to Gorgonia ventalina samples previously reported that were obtained from San Salvador, Bahamas. Spectra from two samples of G. ventalina that had been infected by different agents, Aspergillus sydowii and a slime mold, were also obtained. The results indicate that the purple compounds (polyenes) generated by the coral in response to infection are similar regardless of region from which the coral were harvested, of species of coral, and of the infective agent. A discussion of the Raman spectra of G. ventalina and the other coral species is presented.

The cellular stress response of the scleractinian coral Goniopora columna during the progression of the black band disease

Black band disease (BBD) is a widespread coral pathology caused by a microbial consortium dominated by cyanobacteria, which is significantly contributing to the loss of coral cover and diversity worldwide. Since the effects of the BBD pathogens on the physiology and cellular stress response of coral polyps appear almost unknown, the expression of some molecular biomarkers, such as Hsp70, Hsp60, HO-1, and MnSOD, was analyzed in the apparently healthy tissues of Goniopora columna located at different distances from the infection and during two disease development stages. All the biomarkers displayed different levels of expression between healthy and diseased colonies. In the healthy corals, low basal levels were found stable over time in different parts of the same colony. On the contrary, in the diseased colonies, a strong up-regulation of all the biomarkers was observed in all the tissues surrounding the infection, which suffered an oxidative stress probably generated by the alternation, at the progression front of the disease, of conditions of oxygen supersaturation and hypoxia/anoxia, and by the production of the cyanotoxin microcystin by the BBD cyanobacteria. Furthermore, in the infected colonies, the expression of all the biomarkers appeared significantly affected by the development stage of the disease. In conclusion, our approach may constitute a useful diagnostic tool, since the cellular stress response of corals is activated before the pathogens colonize the tissues, and expands the current knowledge of the mechanisms controlling the host responses to infection in corals.

Development and application of molecular biomarkers for characterizing Caribbean Yellow Band Disease in Orbicella faveolata

Molecular stress responses associated with coral diseases represent an under-studied area of cnidarian transcriptome investigations. Caribbean Yellow Band Disease (CYBD) is considered a disease of Symbiodinium within the tissues of the coral host Orbicella faveolata. There is a paucity of diagnostic tools to assist in the early detection and characterization of coral diseases. The validity of a diagnostic test is determined by its ability to distinguish host organisms that have the disease from those that do not. The ability to detect and identify disease-affected tissue before visible signs of the disease are evident would then be a useful diagnostic tool for monitoring and managing disease outbreaks. Representational Difference Analysis (RDA) was utilized to isolate differentially expressed genes in O. faveolata exhibiting CYBD. Preliminary screening of RDA products identified a small number of genes of interest (GOI) which included an early growth response factor and ubiquitin ligase from the coral host as well as cytochrome oxidase from the algal symbiont. To further characterize the specificity of response, quantitative real-time PCR (qPCR) was utilized to compare the expression profiles of these GOIs within diseased tissues (visible lesions), tissues that precede visible lesions by 2–4 cm (transition area), and tissues from healthy-looking colonies with no signs of disease. Results show there are distinctive differences in the expression profiles of these three GOIs within each tissue examined. Collectively, this small suite of GOIs can provide a molecular “finger print” which is capable of differentiating between infected and uninfected colonies on reefs where CYBD is known to occur.

Classification of Tree Species in Overstorey Canopy of Subtropical Forest Using QuickBird Images

This paper proposes a supervised classification scheme to identify 40 tree species (2 coniferous, 38 broadleaf) belonging to 22 families and 36 genera in high spatial resolution QuickBird multispectral images (HMS). Overall kappa coefficient (OKC) and species conditional kappa coefficients (SCKC) were used to evaluate classification performance in training samples and estimate accuracy and uncertainty in test samples. Baseline classification performance using HMS images and vegetation index (VI) images were evaluated with an OKC value of 0.58 and 0.48 respectively, but performance improved significantly (up to 0.99) when used in combination with an HMS spectral-spatial texture image (SpecTex). One of the 40 species had very high conditional kappa coefficient performance (SCKC ≥ 0.95) using 4-band HMS and 5-band VIs images, but, only five species had lower performance (0.68 ≤ SCKC ≤ 0.94) using the SpecTex images. When SpecTex images were combined with a Visible Atmospherically Resistant Index (VARI), there was a significant improvement in performance in the training samples. The same level of improvement could not be replicated in the test samples indicating that a high degree of uncertainty exists in species classification accuracy which may be due to individual tree crown density, leaf greenness (inter-canopy gaps), and noise in the background environment (intra-canopy gaps). These factors increase uncertainty in the spectral texture features and therefore represent potential problems when using pixel-based classification techniques for multi-species classification.