Saudi Red Sea - Cave Reef (Thuwal)
Part of the regional summaries: Saudi Arabia - Jeddah / KAUST (Thuwal) / Rabigh, Saudi Arabia, and Red Sea
More info
Al Madinah and Makkah (NOAA CRW virtual station)Cave Reef, named for its caverns and overhangs, is a field of eroded, heavily bioeroded limestone pinnacles rising from sand, with coral patchy and mostly small — massive Lobophyllia, faviid domes, and giant clams rather than a continuous framework. A single February 2024 visit found 16% cover with 27% of coral tissue bleached despite zero accumulated heat stress at the time, and fresh white bite scars, which Anderson noted as fish bites, were visible on several massive colonies.
شعاب الكهف، المسماة بكهوفها ونتوءاتها، حقل من القمم الكلسية المتآكلة والمثقوبة حيويًا بشدة ترتفع من الرمل، ومرجانها متقطع وصغير في الغالب - Lobophyllia الكتلية وقباب الفافيدات والمحار العملاق بدلًا من هيكل متصل. ووجدت زيارة واحدة في فبراير 2024 تغطية 16% مع ابيضاض 27% من نسيج المرجان رغم انعدام الإجهاد الحراري المتراكم وقتها، وظهرت ندوب عضّ بيضاء حديثة، دوّنها أندرسون على أنها عضّات أسماك، على عدة مستعمرات كتلية.
ترجمة هذا الملخص بالذكاء الاصطناعي؛ النص الإنجليزي هو المعتمد.
One independent survey station shown with the program’s logo — click one to see how far it is from this site.
Three other CRD dive sites with their own pages within 5 km shown in orange — click a marker to open its page.
7 of the 47 photographs on this page are shown beside the text that discusses them under "Site description" below.
Site description
Setting
The site is named for its caverns, and one frame is shot from inside one — looking up through a ragged gap in the reef roof at bright surface light, the walls in silhouette (e.g., Image 42).
Outside the caves the structure is a field of eroded limestone pinnacles and buttresses rising from white sand, tall, undercut, and heavily bioeroded, standing in hazy blue water (e.g., Image 1, Image 2).
General benthic features
Coral is patchy and mostly small. The living component is dominated by massive and encrusting forms rather than branching ones: a large brown lumpy massive colony on a pinnacle (e.g., Image 3), red-brown meandroid colonies sitting free on sand (Anderson's field call: Lobophyllia), olive and khaki massive domes on pinnacle shoulders, purple and lavender pocilloporids, encrusting sheets, and giant clams with electric-blue and green mantles set into the pavement (e.g., Image 22, Image 39).
Between them, a great deal of the surface is old, algal-filmed, coralline-crusted dead framework (e.g., Image 4).
Coral cover
- Reference: Mayfield (2025)
- Data file: (Mayfield, 2025-appendix A)
- CoralNet submission package (accession #5546)
- 16% ± 20% across 68 photographs.
- The standard deviation of 20 is larger than the mean itself, which is the honest signature of a site where individual frames swing between bare pavement and dense colony clusters.
- The 9 February visit recorded 16% cover with 27% of coral tissue bleached, with no accumulated heat stress at the time.
Stressors
Grazing pressure is visible: Anderson noted fish bites on one frame: a scatter of small, fresh white bite scars across the massive colonies in it.
Comparisons with other data
Apart from our own published survey shown here, we found no other surveys at this site as of 2026. Within 5 km, one independent station, Furby et al. (2013) at Offshore reef 1, 1.51 km away, read a mean of 31% in 2010.
Dive record
A single winter dive to 27.2 m on 9 February 2024 at 24.0°C, the second dive of a day that started at Rose Reef, worked with Coastline (KAUST). Anderson logged a maze of old coral structures with very little live coral remaining, and very silty water with poor visibility.
CV model comparison
More info
The reference is this dive site's own published CoralNet point count by the author (Mayfield 2025, Environments 12:248, Appendix A; most frames annotated by CoralNet's AI, trained and checked by the author). The computer-vision rows are this site's own values: Claude is a blind point count of all of the site's edited photographs, DINOv3 ran on the site's benchmark frames (the photographs the author annotated point by point), and CoralSCOP is the site's finished survey of its edited photographs. They are on different frame sets from the published count, a row under 30 frames is indicative only, and no computer-vision method yet estimates coral cover accurately on new images.| Method | Hard coral cover (%) | Photos (n) ± SE |
|---|---|---|
| Published point count (CoralNet-assisted) | 16.5 | |
| In-house DINOv3 | too few photos (n=2) | |
| CoralSCOP edited | 21.8 | n = 70 |
| Claude (edited photos) | 15.7 | n = 65 · SE 2.1 |
The two computer-vision methods agree within 6.1 pp (15.7–21.8%) and read within 5.3 pp of the published point count (16.5%).
In-house DINOv3: this method was run on 2 of 70 photos, under the 10 needed for a value, so it is not shown and is left out of any mean; every other row rests on its own photo count.
See a more detailed, stand-alone version of this table.
Independent surveys nearby
Within 5 km, 1 independent survey station recorded hard coral cover in 2010. The most recent nearby survey is from 2010, so these values predate recent heat stress. This site's own estimate (16%) sits 15 points below the one nearby survey station (31%); the methods differ, so the comparison is context, not a like-for-like test.
Collected by other programs using different methods, so these are context for the figures above, not a like-for-like comparison. Where a station was surveyed more than once, the cell shows either the range of its values or, as a single value, the mean of surveys made in one year. Station names are not dive-site names.
| Agency | Site / location | Date | Hard coral cover (%) | Distance | Link |
|---|---|---|---|---|---|
| Furby et al. (2013) | Offshore reef 1 | 2010 | 31% (3 records) | 1.51 km | doi:10.1007/s00338-012-0998-5 |
Data & links
All images from Coral Reef Diagnostics. Data are from Coral Reef Diagnostics and the organizations named in the data sources above.
© Coral Reef Diagnostics.