Climate-driven regime shift of a temperate marine ecosystem

See allHide authors and affiliations

Science  08 Jul 2016:
Vol. 353, Issue 6295, pp. 169-172
DOI: 10.1126/science.aad8745

No turning back?

Ecosystems over time have endured much disturbance, yet they tend to remain intact, a characteristic we call resilience. Though many systems have been lost and destroyed, for systems that remain physically intact, there is debate as to whether changing temperatures will result in shifts or collapses. Wernburg et al. show that extreme warming of a temperate kelp forest off Australia resulted not only in its collapse, but also in a shift in community composition that brought about an increase in herbivorous tropical fishes that prevent the reestablishment of kelp. Thus, many systems may not be resilient to the rapid climate change that we face.

Science, this issue p. 169


Ecosystem reconfigurations arising from climate-driven changes in species distributions are expected to have profound ecological, social, and economic implications. Here we reveal a rapid climate-driven regime shift of Australian temperate reef communities, which lost their defining kelp forests and became dominated by persistent seaweed turfs. After decades of ocean warming, extreme marine heat waves forced a 100-kilometer range contraction of extensive kelp forests and saw temperate species replaced by seaweeds, invertebrates, corals, and fishes characteristic of subtropical and tropical waters. This community-wide tropicalization fundamentally altered key ecological processes, suppressing the recovery of kelp forests.

Broad-scale losses of species that provide the foundations for habitats cause dramatic shifts in ecosystem structure, because they support core ecological processes (13). Such habitat loss can lead to a regime shift, in which reinforcing feedback mechanisms intensify to provide resilience to an alternate community configuration, often with profound ecological, social, and economic consequences (46). Benthic marine regime shifts have been associated with the erosion of ecological resilience through overfishing or eutrophication, altering the balance between consumers and resources, rendering ecosystems vulnerable to major disturbances (1, 2, 6, 7). Now, climate change is also contributing to the erosion of resilience (8, 9), because increasing temperatures are modifying key physiological, demographic, and community-scale processes (8, 10), driving species redistribution at a global scale and rapidly breaking down long-standing biogeographic boundaries (11, 12). These processes culminate in novel ecosystems where tropical and temperate species interact, with unknown implications (13). Here we document how a marine heat wave caused the loss of kelp forests across ~2300 km2 of Australia’s Great Southern Reef, forcing a regime shift to seaweed turfs. We describe a rapid 100-km range contraction of kelp forests and a community-wide shift toward warm-water species, with ecological processes suppressing kelp forest recovery.

To document ecosystem changes, we surveyed kelp forests, seaweeds, fish, mobile invertebrates, and corals at 65 reefs across a ~2000-km tropical-to-temperate transition zone in western Australia (14). Surveys were conducted between 2001 and 2015, covering the years before and after an extreme marine heat wave affected the region.

The Indian Ocean adjacent to western Australia is a “hot spot” where the rate of ocean warming is in the top 10% globally (15), and isotherms are shifting poleward at a rate of 20 to 50 km per decade (16). Until recently, kelp forests were dominant along >800 km of the west coast (8), covering 2266 km2 of rocky reefs between 0 and 30 m depth south of 27.7°S (Fig. 1). Kelp forests along the midwest section of this coast (27.7° to 30.3°S) have experienced steadily increasing ocean temperatures since the 1970s, recently punctuated by three of the warmest summers in the past 215 years (Fig. 2) (17, 18, 19). In December 2010, immediately before an extreme marine heat wave, kelp forests covered over ~70% of shallow rocky reefs in the midwest (Fig. 2 and fig. S1), with no differences in kelp cover or biomass among reefs along the west coast (Fig. 1 and figs. S1 and S2) (8). During this time, seaweed and fish communities in the midwest were similar to those of the temperate southwest (~500 km farther south) and clearly distinct from those of tropical reefs in the northwest (~500 km farther north) (Fig. 3, A and B, and fig. S3) (17).

Fig. 1 Extent of kelp forests in western Australia before and after the 2011 marine heat wave.

(Left) The extent of kelp forests from 0 to 30 m depth before 2011 is shown (map), with a color scale indicating the proportion lost by 2013. (Right) The area of kelp forests before 2011 in 0.5° bins (bars). Before 2011, kelp forests covered 2266 km2 along the >800 km of coastline. However, by 2013, 43% of these kelp forests had disappeared (red bars). On the left side of the map, gray squares (southwest region) and black circles (midwest region) mark locations where reefs were surveyed by scuba divers to establish proportional kelp loss (table S1).

Fig. 2 Regime shift from kelp forests to seaweed turfs after the 2011 marine heat wave.

Kelp forests were dense in Kalbarri until 2011 (A), when they disappeared from ~100 km of coastline (Fig. 1) and were replaced by seaweed turfs (B). (C) The habitat transition (lines) coincided with exceptionally warm summers in 2011, 2012, and 2013 (red bars), punctuating gradually increasing mean ocean temperatures over the past decades (17). Shown are mean (± SE, n = 3 reefs) kelp forest cover (dark blue circles and line) and seaweed turf cover (dark red circles and line), chronologically aligned with monthly sea surface temperature (SST) anomalies (blue and red bars) relative to monthly climatological means for 1981–2015 (table S1).

Fig. 3 Changes in seaweed and fish communities on affected reefs after the 2011 marine heat wave.

Ordinations (nonmetric multidimensional scaling) of seaweed (A) and fish (B) communities show how community structure on reefs north of 29°S shifted from a close resemblance to temperate reefs farther south to a greater resemblance to tropical reefs to the north. Dark blue = Perth (2005–2007), light blue = Kalbarri before (2005–2007), pink = Kalbarri after (2013–2015), red = Ningaloo Reef (2010) (table S1). Each symbol represents an individual reef. Ordinations are based on Bray-Curtis dissimilarities calculated from ln(x + 1)–transformed data. (C and D) Change in ln(x + 1)–transformed abundance of seaweeds [(C) grams of fresh weight per 1.5 m2] and fish [(D) individuals per 2500 m2] in Kalbarri (2005–2007 versus 2013–2015) clearly show the decline in cool-water species (blue bars) and concurrent increase in warm-water species (red bars), with several species not previously recorded (+) or now absent from the samples (–). Each bar represents an average across six reefs for an individual species. White bars indicate taxa with ambiguous distributions. Species are listed in tables S2 and S3.

By early 2013, only 2 years later, our extensive surveys found a 43% (963 km2) loss of kelp forests on the west coast (Fig. 1). Previously dense kelp forests north of 29°S had disappeared (Fig. 2 and fig. S1) or been severely reduced (>90% loss, Fig. 1), representing a ~100-km range contraction and functional extinction from 370 km2 of reef (a reduction in abundance severe enough to delete ecological function). Instead, we found a dramatic increase in the cover of turf-forming seaweeds (Fig. 2) and a community-wide shift from species characteristic of temperate waters to species and functional groups characteristic of subtropical and tropical waters [Figs. 3 and 4 and fig. S3, mean square contingency coefficient (ϕφ2,52) = –0.70, P < 0.001]. Compared to the composition of the heavily affected midwest reef communities in Kalbarri before the 2011 marine heat wave, differences in community structure (Bray-Curtis dissimilarity) from reefs in Perth in the temperate southwest increased by 91 and 28% for seaweeds and fishes, respectively, whereas differences from Ningaloo Reef in the tropical northwest decreased by 32 and 16%, respectively. This broad-scale community-wide reef transformation reflected consistent decreases in the abundance of taxa characteristic of temperate reefs, coinciding with increases in the abundance of species characteristic of subtropical and tropical reefs, for both seaweeds (Fig. 3C and table S2, ϕφ2,20 = –0.81, P < 0.001) and fishes (Fig. 3D and table S3, ϕφ2,20 = –0.64, P = 0.008). Similar changes were seen for sessile and mobile invertebrates in the southern part of the midwest region, where small hermatypic coral colonies increased almost sixfold in abundance and doubled in species richness (Fig. 4 and table S5), while the abundances of sea urchins and gastropods also increased and decreased in accordance with their thermal affinities (Fig. 4 and tables S4 and S5, ϕφ2,12 = –0.68, P < 0.045).

Fig. 4 Changes in benthic invertebrate abundances in the midwest after the 2011 marine heat wave.

(A and B) Change in ln(x + 1)–transformed abundance of common mobile invertebrates (inverts) [(A) individuals per 30 m2] and small (<6 cm)hermatypic corals [(B) colonies per 1000 m2]. Colors and symbols are as in Fig. 3. Each bar represents an average across 6 and 23 reefs for an individual species of mobile invertebrates and corals, respectively. Mobile invertebrates were counted in Jurien Bay (2005, 2011 versus 2013 and 2014), and corals were counted between Cervantes and Dongara (30.6° to 29.3°S) (2005–2006 versus 2013) (table S1). Species are listed in tables S4 and S5.

Even though the acute climate stressor has abated (Fig. 2 and fig. S1), as of late 2015, almost 5 years after the heat wave, we have observed no signs of kelp forest recovery on the heavily affected reefs north of 29°S. Instead, concurrent with an 80% reduction in standing seaweed biomass (fig. S2), we have recorded subtropical and tropical fish feeding rates on canopy seaweeds that are three times higher than on comparable coral reef systems. Similarly, we have found a 400% increase in the biomass of scraping and grazing fishes, a functional group characteristic of coral reefs, which now display grazing rates on seaweed turfs that are comparable to those observed on healthy coral reefs worldwide (table S6) (10). High herbivore pressure now suppresses the recovery of kelp forests by cropping turfs and kelp recruits (10).

We deduce that extreme temperatures beginning in 2011 exceeded a physiological tipping point for kelp forests north of 29°S, and now reinforcing feedback mechanisms have become established that support a new kelp-free state. Similar ecosystem changes have not been observed in the southwest, where heat wave temperatures remained within the thermal tolerance of kelps (17), and the greater distance to tropical bioregions limited the incursion of tropical species. Threshold temperatures for kelp forests appear close to 2.5°C above long-term summer maximum temperatures, consistent with other seaweeds in the region (20). However, the partial loss of kelp forests on reefs between 29° and 32°S suggests that there is variation in threshold temperatures within and between kelp populations.

The consistent responses of both cool- and warm-water species clearly illustrate the important role of temperature. However, the transition in community structure and subsequent persistence of the new regime would have been augmented by the low and high availability of temperate and tropical propagules and immigrants, respectively, as well as changes in competitive interactions after the loss of kelp canopies (21). The oceanography of the region is dominated by the poleward-flowing Leeuwin Current, which delivers warm nutrient-poor water and tropical species into the temperate region, while limiting the supply of propagules, including kelp zoospores from higher-latitude kelp forests (22, 23). Indeed, healthy coral reefs already occur at the Houtman-Abrolhos Islands, 60 km offshore from Kalbarri, directly in the path of the Leeuwin Current. Until now, however, cooler coastal waters have enabled kelp forests to dominate nearshore reefs.

The Leeuwin Current is strongly influenced by the El Niño–Southern Oscillation (24). The strength of the Leeuwin Current and the flow of warm tropical water down the west coast of Australia increase during the La Niña phase of this cycle. These La Niña conditions drive warming anomalies such as the 2011 marine heat wave (24) and are predicted to double in frequency and intensity in the near future (25). Moreover, the southeast Indian Ocean has gradually warmed by at least 0.65°C over the past five decades and will continue to warm until the end of the century and beyond (19). Kelp forest recovery from disturbances can be slow in the nutrient-poor west coast waters, even in the cool southwest (8, 26), providing time for populations of herbivorous fish to become established and seaweed turfs to proliferate. Consequently, the probability of prolonged cool conditions that could reset community structure and ecological processes to facilitate the recovery of kelp forests is becoming increasingly unlikely, while the risk of more heat waves that will exacerbate and expand the new tropicalized ecosystem state is increasing (25).

Short-term climate variability has previously precipitated large-scale destruction of kelp forests (27, 28), which have mostly recovered as environmental conditions returned to normal. The future of kelp forest communities in western Australia is, however, grim. Warming, more frequent heat waves, and the intrusion of tropical species into temperate habitats are unequivocal (13, 19, 25). The current velocity of ocean warming is pushing kelp forests toward the southern edge of the Australian continent (29), where they are at risk of rapid local extinction over thousands of kilometers, due to the east-west orientation of the continent’s poleward coastline and west-to-east flow of surface currents (12). This would devastate lucrative fishing and tourism industries worth more than $10 billion (Australian) per year (30) and have catastrophic consequences for the thousands of endemic species (30) supported by the kelp forests of Australia’s Great Southern Reef.

Supplementary Materials

Materials and Methods

Figs. S1 to S3

Tables S1 to S6

References (3170)

References and Notes

  1. Materials and methods are available as supplementary materials on Science Online.
Acknowledgments: This work was funded by the Australian Research Council (T.W., G.A.K.), the Hermon Slade Foundation (T.W., S.B.), a U.K. Natural Environment Research Council Independent Research Fellowship (D.A.S.), the Australian Institute of Marine Science (T.W., M.D., B.R.), the Australian National University (C.J.F.), the Western Australian Museum (J.F.), the Department of Parks and Wildlife (T.H.H., S.W.), CSIRO Oceans and Atmosphere (R.C.B., F.D., M.A.V.), Fisheries Research and Development Corporation project no. 2008/013 (R.K.H., G.A.K.), The Marsden Fund of The Royal Society of New Zealand (M.S.T.), and the WA Strategic Research Fund for the Marine Environment (R.B., M.A.V., J.F.). T.W. and S.B. conceptualized and wrote the manuscript; T.W., S.B., R.B., T.dB., K.C., M.D., F.D., J.F., C.J.F., J.S.-G., R.K.H., E.S.H., T.H.H., G.K., B.R., B.J.S., D.K.S., M.T., C.T., F.T., M.A.V., and S.W. provided data; and T.W., S.B., R.K.H., J.S.-G., and D.A.S. performed analyses and modeling. All authors discussed the results and commented on the manuscript. The data are provided in the supplementary materials. Additional information can be obtained from T.W. All authors declare no conflicting interests.

Stay Connected to Science


Navigate This Article