For nearly three decades, researchers monitored coral growing on concrete breakwaters at Naha Port in Okinawa, Japan, alongside the natural reef sitting a short distance away. The two habitats started out looking very different from each other and, after a severe bleaching event in 1998, they diverged again in a way that surprised the team doing the counting: the artificial structures recovered in about six years, while the adjacent natural reef showed almost no recovery over the same period.
The findings come from a paper published in March 2025 in Scientific Reports, titled “Artificial structures can facilitate rapid coral recovery under climate change,” by Toko Tanaya, Shunpei Iwamura, Wataru Okada, and Tomohiro Kuwae. The dataset runs from 1989 to 2018, which the authors and outside reviewers have both pointed to as unusually long for this kind of study. Bert Hoeksema, a coral taxonomist not involved in the research, described the paper as notable mainly for its duration, telling Mongabay that “this paper is outstanding because it looks at coral settlement on breakwaters over almost 3 decades, which is very long for a study of this kind.” Most comparable studies track recovery over one to ten years.
What the twenty nine years of monitoring actually recorded
Before the 1998 mass bleaching event, coral cover on the natural reef stood at around 45 per cent, well above the breakwater’s cover at the time. The bleaching, driven by a marine heatwave that affected reefs across the western Pacific that year, hit both habitats. Afterwards, natural reef cover fell to 14 per cent. Breakwater coral, dominated by fast growing Acropora species, dropped even further, down to a mean of just 3 per cent by 2000.
What happened next is the part of the paper that has drawn attention. Over the following six years, breakwater coral cover climbed back to roughly 25 per cent, close to its pre bleaching level, at a rate the authors calculate at around 6 per cent a year between 2000 and 2004. The natural reef, over the same stretch, showed minimal recovery. The gap did not close over the following decade of monitoring either. On breakwater sections where the concrete surface had been mechanically roughened during construction, a detail the authors tracked separately, coral cover reached 49 per cent within six years, compared with 18 per cent on unprocessed sections nearby.
This is one paper, drawn from one site, over one specific bleaching event. It is not a survey of Japanese reefs generally, and the authors are not claiming it is.
Why the artificial surfaces recovered faster
The paper’s own analysis points mainly to Acropora’s fast growth and to more favorable conditions on the breakwaters themselves — shallower water, more light, stronger wave-driven flow. Tanaya offered a further possible explanation to Mongabay: that artificial surfaces are more structurally robust than natural reefs, where a dying colony’s calcium carbonate skeleton is left fragile and breaks down under wave action and bioerosion faster than it can be recolonised, leaving a bare, stable settlement surface on concrete almost immediately after a die-off. The authors themselves note this is not yet confirmed.
Tanaya, the paper’s lead author, framed the implication for coastal management plainly rather than as an argument for replacing reefs outright: “Our research suggests that combining natural infrastructure such as coral reefs with artificial infrastructures could be effective to promote coral regeneration, while also demonstrating high disaster prevention,” she told Mongabay. Naha Port’s breakwaters were not built for coral. They were built to blunt wave energy for the harbour. The coral growth on them is, in that sense, incidental. That does not make the recovery pattern any less real, but it is worth being precise about what the structures were designed to do.
What the recovery does not include
The breakwater communities that came back after 1998 were not the same as what came before, in composition if not in raw percentage cover. The paper’s authors note that breakwater coral cover, while quicker to rebound, showed lower species diversity than the natural reef community, and was dominated by fast growing branching species rather than the fuller mix of forms found on natural substrate. A reef that recovers in cover but narrows in composition is occupying the same footprint as a different community, not running a faster version of the original one.
The authors are also careful to say the mechanism is not fully resolved. Structural stability is the leading explanation the paper offers, but the team notes that further work is needed to separate that factor from others, such as differences in herbivory, sedimentation, or local water flow between the two habitats. Single site studies, however long running, cannot settle questions like that on their own.
Where this sits in the wider conversation about artificial reefs
Artificial reef structures, sunken ships, concrete modules, breakwaters retrofitted for coral, have been proposed as restoration tools for years, often on the strength of shorter term or more optimistic framing than the underlying data supports. What the Naha Port dataset adds is not a case for artificial reefs as a substitute for natural ones. It is a specific, long run comparison from one location showing that under one kind of disturbance, an engineered structure’s advantage was resilience of the base itself, not anything about the coral growing on it.
Coastal engineers in Japan have already been retrofitting breakwaters with roughened or textured surfaces to encourage settlement, a practice this dataset appears to support in a narrow sense: the processed sections in this study did better than the unprocessed ones. Whether that finding generalises to breakwaters built from different materials, in different current regimes, exposed to different bleaching severities, is not something a single 29 year record at one port can answer.
The next bleaching event, whenever it comes, will be the more informative test of whether the pattern documented here holds. Okinawa’s reefs, natural and artificial alike, are likely to face one before this decade is out.