Rapid poleward range expansion of tropical reef corals in response to rising sea surface temperatures

1. Introduction

[2] Sea-surface temperature (SST) is an environmental predictor closely linked with marine biodiversity [Tittensor et al., 2010], and climatic warming is likely affecting species distribution ranges [Hoegh-Guldberg and Bruno, 2010]. Large-scale, long-term observations are needed to detect such changes, but compared to terrestrial studies, such research on marine species remains scarce, partly due to the size and complexity of the ocean, but also to the relative difficulty of taking measurements in marine environments [Hoegh-Guldberg and Bruno, 2010]. In this study, we show the first large-scale evidence of the poleward range expansion of corals, based on 80 years of national records from the temperate areas of Japan, where century-long measurements of in situ sea-surface temperatures have shown statistically significant rises.

[3] Corals play a fundamental role in primary production and habitat formation for numerous other species in tropical and subtropical areas. Thus, their poleward range expansions due to climatic warming could cause fundamental modifications of temperate coastal ecosystems. Although most studies of climate change effects on corals have focused on temperature-induced coral bleaching in tropical areas, poleward range shifts and/or expansions may also occur in temperate areas, as suggested by geological records and present-day eyewitnesses in several localities [Greenstein and Pandolfi, 2008; Precht and Aronson, 2004]. In addition to their importance in ecosystem function, corals are also sensitive detectors of long-term climatic warming effects. Adult coral colonies are sessile, and several years after larval settlement are required for a colony to develop sufficiently to be recognised in situ. Corals are basically long-lived, but are extremely sensitive to temperature. Both high and low temperatures can lead to bleaching, which causes coral mortality [Hoegh-Guldberg et al., 2005]. Therefore, detection of range shifts and/or expansions of corals would provide solid baselines to discuss changes of coastal marine biodiversity and ecosystems in temperate areas.

2. Setting

[4] Japan covers a wide latitudinal range, stretching from subtropical to temperate areas, and latitudinal limits of coral reefs and coral distributions are observed around the Japanese islands [Veron, 1992b; Veron and Minchin, 1992; Yamano et al., 2001] (Figure 1). In addition to SST limits, other reasons why Japan provides a unique opportunity for examining baselines of species range shifts and/or expansions due to climatic warming over a large spatial scale include the following. First, the Japanese islands form an almost continuous chain, and corals are distributed in a stepping-stone sequence along a latitudinal gradient, due to the Kuroshio and Tsushima warm currents, which potentially transport coral larvae from south to north [Veron and Minchin, 1992]. Second, because of Japan's relatively small landmass in comparison with other temperate settings (e.g., the United States, Africa and Australia), distribution is likely to be less disturbed by terrestrial influence. Third, SSTs, which constrain coral distributions [Veron and Minchin, 1992], have risen in the last 100 years as indicated by the Japan Meteorological Agency [Takatsuki et al., 2007]. This was revealed by analyses of in situ observations of SSTs from ICOADS [Diaz et al., 2002] and the Kobe Collection [Manabe, 1999], along with in situ data collected by related agencies. Finally, corals have been observed periodically in Japanese waters since the 1930s. Our dataset on corals is thus one of the largest-scale, longest-term records among marine species reported to date [Helmuth et al., 2006; Hoegh-Guldberg and Bruno, 2010].

3. Methods

[5] We collected records of coral species occurrence from eight temperate regions of Japan along a latitudinal gradient (Figure 1), where past coral occurrence records were available in the form of literature and specimens (auxiliary material). When the previous northern range margin was estimated to have been farther south than the study regions, we also examined occurrence in the Tokara–Tanegashima region (Figure 1), based on literature (auxiliary material). Four distinct periods of coral exploration have occurred. The first period was in the 1930s and was undertaken by Drs Yabe and Sugiyama at Tohoku Imperial University (Tohoku University) of Japan. Some of the specimens collected by them were stored in Tohoku University Museum. The second period was from the 1960s and 1970s. Comprehensive surveys of marine environments and organisms were conducted as part of the designation planning of marine protected areas. Coral identification was performed mainly by Drs Utinomi and Eguchi. The third period was from the 1980s to 1990s, when Dr. Veron, who revised and integrated the recent taxonomic framework of corals, visited Japan and examined coral distribution [Veron, 1992b; Veron and Minchin, 1992]. The last period covers the late 1990s to the present. Japanese researchers, including the authors, conducted field surveys based on the recent taxonomic framework. In this work, we conducted additional field surveys in regions lacking detailed recent information. In addition to observing coral occurrence, colony sizes of the selected corals (Table 1) were measured to infer past occurrence. The eight study regions are all located in seas showing significant SST rises in winter (January–March) (1.1°C–1.6°C) [Takatsuki et al., 2007], which is critical for corals to survive at the latitudinal limits of their distribution [Veron and Minchin, 1992]. Since all the sites in these regions are remote from densely populated cities and most sites are located in marine protected areas, anthropogenic impacts are expected to be minor.

[6] Corals are one of the most difficult taxa to identify at species level. They have few taxonomically significant morphological characteristics because such characters show polymorphism, intraspecific variation and phenotypic plasticity in response to environmental factors [Veron, 2000]. In addition to the difficulties of identification, the coral taxonomic framework before 1990 differed greatly from that in current use. Therefore, careful examination is needed to establish consistent databases of coral records. Species for comparison were selected based on three criteria. First, species that were abundant in the regions of interest [Sugihara et al., 2009] were selected to avoid the possibility that past surveys had missed their occurrence. Second, species distributed typically at depths of less than 10 m were selected, as some past surveys used snorkeling or observation through a water glass from a boat. Finally, and most importantly, species with distinct growth forms that could be identified in situ were selected because in some previous reports, observers identified species in situ without collecting specimens. As a result, nine species categories were selected (Table 1 and auxiliary material). These species, except for Acropora pruinosa/tumida, Caulastrea tumida and Lithophyllon undulatum, are common in the Indo-Pacific region [Veron, 2000], and thus similar assessment is possible for other sites in the Indo-Pacific region by referring to these species. We revised species names from past literature based on the recent coral taxonomic framework formulated by Veron [1992b] (auxiliary material). We also examined descriptions and photographs available in the literature to confirm species names. One of the authors (KS) visited Tohoku University Museum and re-examined the 1930s specimens based on Veron's framework. To minimize local effects, we pooled species lists from several sites in one region to represent occurrence in the region.

[7] We calculated the speed of coral range expansions along latitude based on the occurrence records. Although coral colonies may be recognized several years after larval settlement, we used the year when new colonies were found, in order to keep consistency in calculation based on the occurrence records. Past literature did not describe colony sizes, which made estimation of larval settlement years impossible. Thus, the expansion speed likely was underestimated.

4. Results and Discussion

[8] Four species categories of the nine selected showed poleward range expansion since the 1930s, whereas the other five remained stable (Figure 2), indicating no southward range shrinkage or local extinction. The estimated settlement years of the colonies of expanded species (Table 1) provide further evidence of recent expansions. Of the expanded species, Acropora hyacinthus and Acropora muricata are key species for reef formation in tropical Indo-Pacific regions [Hongo and Kayanne, 2011; Montaggioni, 2005], and clearly indicate the expansion of tropical species ranges to temperate areas. Note that all of the expanded species have been in IUCN (International Union for Conservation of Nature) extinction risk categories of “Near Threatened” or “Vulnerable” since 1998, when temperature-induced mass bleaching occurred [Carpenter et al., 2008] (Table 1). Adult colonies in these regions exhibited spawning [Mezaki et al., 2007; van Woesik, 1995], indicating that corals newly settled as a result of expansion have the potential to reproduce and expand farther northward. Thus, temperate areas may serve as refugia for tropical corals in an era of global warming, while corals in tropical areas suffer declines because of rising SSTs [Hoegh-Guldberg et al., 2007].

[9] These expansions could have major effects on temperate ecosystems. As the expanding A. hyacinthus and A. muricata have higher growth rates than other species (Table 1), several species around mainland Japan, including five endemic species [Veron, 1992a], could suffer declines as a result of competition caused by the invasion of the tropical species. A. hyacinthus and A. muricata generally occur with respective associated specific species that occupy a similar niche [Hongo and Kayanne, 2010]. This means that such associated species might also be expanding their ranges, as shown in Kushimoto, where two of species associated with A. hyacinthus (Acropora gemmifera and Acropora microphthalma) occurred recently [Nomura et al., 2008]. Furthermore, the range expansions of corals may cause range expansions in other tropical marine organisms associated with coral reefs. Recent findings in Japan, although not based on systematic surveys, suggest that such organisms (Acanthaster planci, reef fish and toxic microalgae that cause ciguatera) may also be expanding their ranges [Oshiro, 2008].

[10] The speed of coral range expansions presented in this study reached up to 14 km/year, which is far greater than that of other species, including terrestrial species (average 0.61 km/year) [Parmesan and Yohe, 2003], and generally greater than that of other marine benthic species (as much as 5 km/year) reported so far [Helmuth et al., 2006]. In addition to rising SSTs, the Kuroshio and Tsushima warm currents may be contributing to the quick expansions. Although range expansions of corals have not been detected, range expansions of reef fish were suggested in eastern Australia [Figueira and Booth, 2010], where the Eastern Australian Current, which flows from the equator poleward toward the south, may play a similar role as the Kuroshio and Tsushima warm currents.

[11] Collectively, these findings strongly suggest that fundamental modifications of temperate coastal ecosystems due to climatic warming could be in progress, and in regions with poleward current flows (east coast of the United States [Precht and Aronson, 2004], east coast of South America, east coast of Africa and east coast of Australia [Figueira and Booth, 2010]) the speed would be much greater. Our results not only provide evidence to serve as a baseline for the rapid range expansions of tropical reef corals but they also have strong implications for the potential range expansions of tropical species associated with coral reefs, which would contribute to changes in coastal marine biodiversity and ecosystems in temperate areas.