Back

RGB color indices as proxy for symbiont cell density and chlorophyll content during coral bleaching

Ferrara, E. F.; Bauer, L.; Puntin, G.; Bautz, F. R.; Celayir, S.; Do, M.-S.; Eck, F. L.; Heider, M. C.; Wissel, P. M.-C.; Arnold, A.; Wilke, T.; Reichert, J.; Ziegler, M.

2024-12-22 ecology
10.1101/2024.12.20.629333 bioRxiv
Show abstract

Coral bleaching, the breakdown of the symbiosis between the coral host and endosymbiotic microalgae, is the main cause of widespread coral reef degradation. Current methods for assessing coral health based on visual appearance, such as the use of color reference cards, are limited by subjective human color perception and low resolution. Digital photography with RGB (Red, Green, Blue) color channel analyses offers a fast, non-invasive, and standardized alternative to estimate physiological parameters. However, the link between coral color and physiological parameters during bleaching may vary depending on the type of stressor. While such approaches are extensively used in plant studies, their application in estimating Symbiodiniaceae cell density and chlorophyll content in corals requires further attention. In this study, we analyzed the correlation between Symbiodiniaceae cell density and chlorophyll content across three coral species (Acropora muricata, Pocillopora verrucosa, and Stylophora pistillata) with 19 color indices derived from the RGB channels currently established as predictors of chlorophyll content in plants. Corals were exposed to three bleaching conditions (acute short-term and chronic long-term heat stress and menthol bleaching) to identify the best color indices for assessing coral health through image analysis. We found that the Red index had the strongest linear correlation with symbiont cell density and chlorophyll content across species (R2 up to 0.97), so that relative changes in this color index can be directly interpreted as corresponding changes in tissue parameters. To train a model that predicts symbiont densities of a distinct sample set using the Red index, we found 10 to 12 samples to be sufficient to achieve an accuracy of > 95 % of the models trained on the full datasets. This research contributes to improved image analysis as a reliable and non-invasive tool for monitoring, by providing guidelines for a systematic use of RGB data to interpret coral health.

Matching journals

The top 2 journals account for 50% of the predicted probability mass.

1
Frontiers in Marine Science
62 papers in training set
Top 0.1%
38.4%
2
Coral Reefs
21 papers in training set
Top 0.1%
14.6%
50% of probability mass above
3
Marine Ecology Progress Series
21 papers in training set
Top 0.1%
4.2%
4
PeerJ
308 papers in training set
Top 2%
3.9%
5
Global Change Biology
78 papers in training set
Top 0.5%
3.9%
6
Science of The Total Environment
186 papers in training set
Top 1%
3.4%
7
PLOS ONE
5266 papers in training set
Top 39%
3.1%
8
Journal of Experimental Biology
259 papers in training set
Top 1%
2.7%
9
Ecology and Evolution
267 papers in training set
Top 3%
2.3%
10
Frontiers in Ecology and Evolution
69 papers in training set
Top 1%
2.1%
11
Scientific Reports
3612 papers in training set
Top 56%
1.7%
12
Ecological Informatics
33 papers in training set
Top 0.4%
1.7%
13
Communications Biology
993 papers in training set
Top 27%
1.0%
14
Limnology and Oceanography: Methods
11 papers in training set
Top 0.2%
1.0%
15
Methods in Ecology and Evolution
176 papers in training set
Top 2%
0.8%
16
Limnology and Oceanography
32 papers in training set
Top 0.4%
0.8%
17
Royal Society Open Science
214 papers in training set
Top 6%
0.8%
18
Journal of Phycology
14 papers in training set
Top 0.3%
0.8%
19
Hydrobiologia
12 papers in training set
Top 0.5%
0.8%
20
Peer Community Journal
281 papers in training set
Top 5%
0.8%
21
Environmental Microbiology
133 papers in training set
Top 3%
0.6%
22
G3 Genes|Genomes|Genetics
351 papers in training set
Top 4%
0.6%