Turbidity flows – underwater avalanches – are large-scale physical disturbances that are believed to have profound and lasting impacts on benthic communities in the deep sea, with hypothesized effects on both productivity and diversity. In this review we summarize the physical characteristics of turbidity flows and the mechanisms by which they influence deep-sea benthic communities, both as an immediate pulse-type disturbance and through longer-term press-type impacts. Further, we use data from turbidity flows that occurred hundreds to thousands of years ago as well as three more recent events to assess published hypotheses that turbidity flows affect productivity and diversity. We find, unlike previous reviews, that evidence for changes in productivity in the studies was ambiguous at best, whereas the influence on regional and local diversity was more clear-cut: as had previously been hypothesized, turbidity flows decrease local diversity but create mosaics of habitat patches that contribute to increased regional diversity. Studies of more recent turbidity flows provide greater insights into their impacts in the deep sea, but without pre-disturbance data, the factors that drive patterns in benthic community productivity and diversity, be they physical, chemical, or a combination thereof, still cannot be identified. We propose criteria for data that would be necessary for testing these hypotheses and suggest that studies of Kaikōura Canyon, New Zealand, where an earthquake-triggered turbidity flow occurred in 2016, will provide insights into the impacts of turbidity flows on deep-sea benthic communities as well as the impacts of other large-scale disturbances such as deep-sea mining.
Turbidity flows are a type of large-scale physical disturbance that is prevalent in the deep sea (i.e. at water depths
The frequency of turbidity flows is highly variable; some can occur annually to decadally (Dennielou et al., 2017; Heezen et al., 1964; Liao et al., 2017; Vangriesheim et al., 2009) while others such as those initiated by earthquakes
Graph summarizing the frequency (years) and extent patch size (
Turbidity flows transport massive volumes of sediment and associated organic material from the near-shore environment into the deep sea. Ancient turbidity flows (e.g. Grand Banks and others noted below) have transported an estimated average volume of 0.525 to 185
Physical disturbances, such as turbidity flows, are a structuring factor for biotic communities in all environments (Churchill and Hanson, 1958; Dayton, 1971; Dial and Roughgarden, 1998; Levin and Paine, 1974; Paine, 1979; Raup, 1957; Sousa, 1979; Weaver, 1951; Webb, 1958). In the marine environment disturbances can be caused by natural physical processes such as the battering of intertidal regions by tides and storms (Dayton, 1971; Levin and Paine, 1974; Paine, 1979), biological processes such as bioturbation of sediment (Hall et al., 1994; Preen, 1996; Reidenauer and Thistle, 1981; Thrush et al., 1991), and anthropogenic impacts such as those arising from bottom trawling (Collie et al., 2000; Lundquist et al., 2010; Thrush et al., 1998). Compared to other physical disturbances experienced by benthic communities, turbidity flows represent a major disturbance (Fig. 2).
Pickett and White (1985) define a disturbance as “
For seabed or benthic communities, as already noted above, the effect of a disturbance varies within a community depending on the characteristics of the disturbance as well as the biological characteristics or traits of the impacted organisms. For example, the more mobile the organism, the greater its likelihood of escaping the disturbance altogether, by being able to either burrow below the area impacted by the event (in the case of small organisms) or leave the area altogether (in the case of larger megafaunal organisms) (Crandall et al., 2003). Additionally, small mobile organisms can sometimes burrow upwards when disturbances bury them rather than being smothered (Maurer et al., 1986; Nichols et al., 1978; Tiano et al., 2020). It can be harder to determine the impact of disturbances on mobile organisms such as demersal fish compared to sessile organisms, because they may not be killed outright by the event. Newly exposed or dead benthic fauna resulting from turbidity flows can provide an immediate and concentrated source of food for fish (Okey, 1997). However, after such short-term benefits are exploited, because disturbances mostly eliminate or create shortages of vital resources such as food or cover for mobile organisms, populations tend to decline until these resources have regenerated (Sousa, 1984).
While we know that turbidity flows can have damaging impacts on seabed communities in the deep sea (see Sect. 2.2), we do not understand clearly how benthic communities respond to these catastrophic events or how patterns of benthic productivity and diversity are influenced by them. As noted by Glover et al. (2010), turbidity flows have historically been studied via palaeontological proxies due to their size, the timescales at which they occur, and the inaccessibility of the deep-sea environment (Fig. 2). These proxies are identifiable by their characteristic deposition of graded sediment known as turbidites (Kuenen and Migliorini, 1950).
This review is timely; since the previous review on the topic, which was limited to studies of ancient turbidites (Young et al., 2001), a growing body of studies have been published on more recent turbidity flows. This review was further prompted by the 2016 Kaikōura Earthquake (New Zealand) and subsequent turbidity flow, which present an exceptional opportunity to advance our understanding of turbidity flow impacts on deep-sea benthic communities. Here we evaluate published data on turbidity flows to assess the influence of this type of disturbance on (1) benthic community productivity and (2) local and regional diversity of benthic communities. (3) We then consider further research to address the gaps in our understanding of how turbidity flows impact benthic communities in the deep sea.
The classic case study of a turbidity flow is that described by Heezen and Ewing (1952) following the Grand Banks Earthquake in November 1929. Following the 7.2 (
Three well-studied but ancient turbidity flow sites were reviewed by Young et al. (2001): those that have occurred in the Cascadia Channel, the Venezuela Basin, and the Madeira Abyssal Plain. The Cascadia Channel, adjacent to the states of Oregon and Washington on the west coast of the United States of America (USA), has evidence of multiple turbidites originating from the Columbia River drainage, which collectively extend at least 650
Map showing the approximate location of the turbidity flows reviewed in this paper. See Table 1 for additional details.
Metadata for turbidites and studies discussed and analysed in this review.
Turbidity flows act as both press- and pulse-type disturbances with erosional and depositional forces (Harris, 2014). The immediate pulse-type impact is mass mortality, due to either dislodgment by erosional forces or burial by deposition (Fig. 1a and b). Organisms caught up in the erosional forces become part of the material transported to deeper depths, and it is generally assumed that even if they are not killed outright, they would be unable to establish themselves in depositional environments that may be tens to hundreds of kilometres away and at deeper water depths (Griggs et al., 1969). However, some studies have found evidence of range extension of shallow-water taxa in areas with regular sediment mass movements that might be the result of acclimation of transported fauna (Kawagucci et al., 2012; Rathburn et al., 2009; Tsujimoto et al., 2020).
It has been suggested that the impacts of burial in areas of deposition may be greater for deep-sea fauna than the erosional forces (Miller et al., 2002). Whether deep-sea organisms will survive the deposition of a turbidite depends on depth of sediment deposition, the type of material being deposited compared to the sediment already there, and the impacted organisms (Young and Richardson, 1998). In shallow-water environments where sedimentation rates are higher and physical disturbances more frequent, studies have shown some organisms are able to survive burials between 30 and 50
The unconsolidated nature and sometimes quite fine grain size of the sediment transported by turbidity flows mean that the impacts of the mass wasting are often spread over a large area (Lambshead et al., 2001). The clogging impact of increased turbidity, especially to filter and suspension feeders, may persist for long periods following the triggering event. Turbid waters were observed in Sagami Bay for up to 3 h following the 5.4 (
Another way in which turbidity flows act as press-type disturbances and may delay faunal community response to the newly available sediment deposits is the creation of anoxic and hypoxic conditions. The introduction of large volumes of organic matter, either from organic-rich coastal sediments or organisms caught up in the mass sediment movement, can lead to anoxic conditions developing as bacteria break down the newly settled and buried organic matter. The presence of these near-surface reducing zones has been observed as layers of thin, iron-rich crust in the Venezuela Basin turbidites (Briggs et al., 1985). These anoxic or hypoxic conditions restrict organisms that rely on aerobic respiration and as a result may delay recruitment of benthic fauna as colonization cannot begin until oxygen is again present in the bottom water (Froelich et al., 1979). Low oxygen persistence is a function of the volume and organic content of the turbidite, speed and direction of the benthic boundary layer currents, and diagenetic processes in the sediment (Sholkovitz and Soutar, 1975).
Turbidity flows can create completely new habitats in the deep sea not only by removing existing faunal communities but also by uncovering or creating new resources. For example, chemosynthetic communities, unique assemblages of organisms that are fuelled by the chemosynthesis of reduced chemical compounds rather than photosynthetic detritus (Sibuet and Olu, 1998), will occur where there is enough organic material to support reducing conditions (Gooday et al., 1990). Turbidity flows initiate the development of chemosynthetic communities both through the burial of large volumes of organic material and through exposure of methane-bearing sediments by erosion (Rathburn et al., 2009). Chemosynthetic communities have been observed at the Laurentian Fan in the path of the Grand Banks turbidity flow, in Monterey Canyon slide scarps, and at the Congo deep-sea fan (Embley et al., 1990; Mayer et al., 1988; Savoye et al., 2000). In the Laurentian Fan the chemosynthetic communities were associated with gravel exposed by the Grand Banks turbidity flow that allowed methane-rich fluids to percolate to the surface (Mayer et al., 1988). The chemosynthetic communities in Monterey Canyon were also associated with the erosional environments of
Prior to the introduction of the concept of turbidity flows, it was widely accepted that all non-chemosynthetic fauna in the abyss (i.e. generally 2000 or 3000 to 6000
Young et al. (2001) reviewed several studies from abyssal turbidites (Cascadia Channel-Abyssal Plain, Venezuela Basin, and Madeira Abyssal Plain) to formally evaluate the Heezen et al. (1955) hypothesis. They concluded that the data from the studies they examined did not support the hypothesis. However, the studies cited by Young et al. (2001) are not as conclusive as they were interpreted to be. The Cascadia Channel–Abyssal Plain and Venezuela Basin studies sampled three main sedimentary regimes, termed turbidite, pelagic, and hemipelagic. Hemipelagic sediments have higher biogenic and terrigenous material than pelagic sediments as they are found on continental slopes beneath highly productive surface waters. At the Madeira Abyssal Plain the study sites also included an abyssal plain below more productive surface waters, in this case seasonally eutrophic waters, an oligotrophic turbidite, and oligotrophic non-turbidite abyssal plain. Young et al. (2001) focused on the difference between the faunal abundances and biomass in the turbidites and those in the hemipelagic or eutrophic sediments, rather than considering the nearby pelagic sediments.
In the Cascadia Channel–Abyssal Plain, the abundance (1011
Thurston et al. (1994) observed that megafaunal abundance and biomass were lower at the Madeira Abyssal Plain (turbidite) compared to the Porcupine Abyssal Plain (non-turbidite) in the northeastern Atlantic. Thurston et al. (1994) attributed these differences between the sites to variation in surface productivity, supplying phytodetritus to the Porcupine Abyssal Plain and not the Madeira Abyssal Plain. In this case Young et al. (2001) noted that sediment trap data showed similar overall fluxes of suspended material to the seafloor at both sites (Honjo and Manganini, 1993; Newton et al., 1994) and therefore argued that the turbidite and not phytodetritus was the driving factor of variation observed between megafauna at the two sites. However, a subsequent study comparing foraminiferal communities, as a proxy for the meiofaunal communities, at the two sites showed that not only was there a higher amount of phytodetritus at the Porcupine Abyssal Plain but there were also foraminiferal communities uniquely suited to exploiting phytodetrital aggregates (Gooday, 1996). Similarly, the dominant megafauna seen at the Porcupine Abyssal Plain were “vacuum cleaner” holothurians, which are well suited to utilizing phytodetritus aggregates (Thurston et al., 1994). Further, each of these studies included third sites which are close to the Madeira Abyssal Plain and similar to it in terms of suspended material flux but have not been impacted by turbidites. For both megafauna and foraminifera (meiofauna), there was no significant difference in abundance or biomass between the Madeira Abyssal Plain turbidite site and the additional non-turbidite sites, but abundance and biomass were higher in the Porcupine Abyssal Plain than all other sites (Gooday, 1996; Thurston et al., 1994). A later study by Thurston et al. (1998) specifically accounted for the turbidite's potential influence on megafauna invertebrates in the NE Atlantic. They selected two sites from the Madeira Abyssal Plain that had not been impacted by turbidites and found that while there were significant differences in abundance and biomass between these two non-turbidite sites, the general trends between the sites (low abundance, low biomass, dominance of non-detritivore taxa) compared to the Porcupine Abyssal Plain non-turbidite site indicated a regional uniformity that supported the previous finding that the Madeira Abyssal Plain turbidite site was similar to other non-turbidite sites (Thurston et al., 1994, 1998). This conclusion was further supported by a study examining polychaete abundance in the northeast Atlantic (Glover et al., 2001). Thus, evidence from the NE Atlantic does not support the hypothesis of Heezen et al. (1955) that benthic productivity in the abyss is affected by carbon delivered by turbidity flows, albeit 930 years after the deposition of turbidites on the Madeira Abyssal Plain.
Samples of the macrofauna and meiofauna communities from throughout the flow path of turbidity flows in the Congo Channel and a nearby control site have been examined by multiple studies that occurred after the review by Young et al. (2001). The study sites include the channel floor which is regularly disturbed by turbidity flows, a levee site that is only impacted by turbidity flows that are large enough to spill over the canyon walls (Savoye et al., 2009) (which occurred during a March 2001 turbidity flow (Khripounoff et al., 2003) but not during a January 2003 event (Vangriesheim et al., 2009)), and multiple sites on the terminal fan that is formed by the periodic deposition of turbidity flows (Savoye et al., 2009). The multiple sites at the fan are from five lobes which are differently impacted by the turbidity flows, including one “abandoned” lobe no longer receiving deposited sediment (Dennielou et al., 2017; Sen et al., 2017). The control site is roughly the same water depth (4000
Three studies evaluating the impact of turbidity flows on the meiofauna and prokaryote communities following the turbidity flow triggered by the 9.0 (
Subsurface peaks in the vertical distribution of meiofauna and macrofauna were also observed in the Congo Channel (Van Gaever et al., 2009; Galéron et al., 2009) and meiofauna in Cap Breton Canyon (Hess et al., 2005). In the Congo Channel the peak for macrofauna was attributed to either the distribution of organic material or the periodic burial by turbidity flows, which favours living deeper as a strategy to avoid disturbance (Galéron et al., 2009). At Cap Breton, the peak was, again, attributed to the increased carbon but other explanatory factors such as the oxygen content, chemical factors, or grain size could not be ruled out (Anschutz et al., 2002; Hess et al., 2005; Hess and Jorissen, 2009).
Nomaki et al. (2016) also found a similar anomalous vertical distribution compared to general deep-sea trends in prokaryote and meiofauna communities 1 year after the 2011 turbidity flow off the coast of Tōhoku (310–880
Graphs of the impacts of turbidity flows on the biomass and abundance (proxies of productivity) of benthic communities in three size classes (mega-, macro-, and meiofauna) and a summary group of all size classes (“All Fauna”). Increase, decrease, or no change in abundance or biomass in turbidite-affected areas was assessed by comparing individual measurements of these productivity metrics at different sites in the studies listed in Table 1.
Although Young et al. (2001) dismissed the productivity-based hypothesis of Heezen et al. (1955), after their review of a small selection of the studies discussed above, they did conclude that turbidity flows likely have an impact on deep-sea faunal diversity as proposed by Angel and Rice (1996). Young et al. (2001) supported the hypothesis that while the initial impacts of turbidity flows can cause local mortalities either by erosional or depositional forces and therefore negatively impact local diversity (Fig. 1c), overall they contribute to the high regional species richness of deep-sea benthic communities by creating mosaics of habitats in time and space (Angel and Rice, 1996).
In general, locally depressed diversity and increased dominance have been observed in turbidites, a pattern typical of disturbed regimes throughout marine environments (Aller, 1997; Glover et al., 2001; Okey, 1997; Paterson and Lambshead, 1995). These general trends of lower local diversity at turbidites as predicted by Angel and Rice (1996) have been found in a number of turbidity flow studies (Briggs et al., 1996; Frutos and Sorbe, 2017; Van Gaever et al., 2009; Glover et al., 2001; Hess and Jorissen, 2009; Kitahashi et al., 2016; Lambshead et al., 2001; Olu et al., 2017; Tsujimoto et al., 2020). Meanwhile, at a regional scale these same studies and others have noted that the turbidites host unique communities of species not seen at other sites in the region and therefore increase the diversity of the region as a whole (Briggs et al., 1996; Frutos and Sorbe, 2017; Van Gaever et al., 2009; Glover et al., 2001; Hess and Jorissen, 2009; Olu et al., 2017; Tietjen, 1984), supporting the Angel and Rice (1996) hypothesis (Fig. 5).
Graphs of the impact of turbidity flows on diversity in three size classes (mega-, macro-, and meiofauna). For local diversity, the increase, decrease, or no change in species richness or other diversity indices was assessed from individual measurements of these metrics made in the studies listed in Table 1. For regional diversity, the assessment was made based on “unique community” as noted by the authors of these studies, or our interpretation of “uniqueness” from the presented community data (e.g. nMDS plots or dendrograms).
Additional studies from the turbidite sites discussed previously attempt to shed light on what characteristics of the turbidites may be driving these diversity patterns. Glover et al. (2001) compared diversity of polychaetes and other macrofauna in the NE Atlantic and Lambshead et al. (2001) reviewed meiofaunal nematode diversity at disturbed sites in the NE Atlantic and Venezuela Basin. The Madeira Abyssal Plain turbidite site was characterized by low polychaete and nematode species diversity and high dominance (Glover et al., 2001; Lambshead et al., 2001), a pattern also observed for nematodes at the HEBBLE site, an area of the deep sea that is regularly disturbed by benthic storms (Aller, 1997; Lambshead et al., 2001). The authors of these studies interpreted the polychaete and nematode communities as being characteristic of a site that has been recolonized after a disturbance (Glover et al., 2001; Lambshead et al., 2001). Similar high dominance by various meiofauna species described as “opportunistic early colonizers” has been noted at the Congo Channel (Van Gaever et al., 2009), Cap Breton Canyon (Hess et al., 2005; Hess and Jorissen, 2009), and off the coast of Japan (Tsujimoto et al., 2020). However, nematode diversity at the Venezuela Basin showed no impact from the turbidite (Lambshead et al., 2001). This difference in results between the Madeira Abyssal Plain and Venezuela Basin nematodes was attributed to the difference in the age of the respective turbidites (Madeira Abyssal Plain: 930; Venezuela Basin: 2000 years) and the difference in local sedimentation rates (Madeira Abyssal Plain: 0.1–1.0
Evidence for the effects of turbidity flows on the productivity of faunal communities in the deep sea is ambiguous at best. From the reviewed studies, it is difficult to draw a conclusion on the general validity of the Heezen et al. (1955) turbidity flow productivity hypothesis in the deep sea. Some studies show a positive influence on proxies of productivity, particularly biomass (Briggs et al., 1996; Carey, 1981; Griggs et al., 1969; Hess and Jorissen, 2009; Richardson et al., 1985; Thurston et al., 1994), while these same studies and others show negative or no influence on other proxies of productivity, specifically abundance (Briggs et al., 1996; Carey, 1981; Van Gaever et al., 2009; Galéron et al., 2009; Glover et al., 2001; Gooday, 1996; Kitahashi et al., 2014, 2016; Lambshead et al., 1995, 2001; Pearcy et al., 1982; Richardson et al., 1985; Woods and Tietjen, 1985). The difficulty in distinguishing a clear effect of turbidity flows on deep-sea benthic productivity among these studies is most likely related to the particular nature of the sites used to evaluate the hypothesis. The location relative to the turbidity flow path and the time since impact will influence a community's response, and the type of measurement used can influence our perception of that response. For example, the location along the flow path determines whether the impact was mostly erosional or depositional (Fig. 1a). The time since impact by a turbidity flow (months to years to thousands of years) determines what, if any, influence is still observable at the site (Fig. 1b and c). In addition, the community response is variable across proxies of productivity, i.e. abundance or biomass.
In contrast, the evidence in support of the Angel and Rice (1996) diversity hypothesis is relatively clear. However, even in the case of the hypothesis of Angel and Rice (1996), where evidence for lower local diversity (Briggs et al., 1996; Frutos and Sorbe, 2017; Van Gaever et al., 2009; Glover et al., 2001; Gooday, 1996; Hess and Jorissen, 2009; Kitahashi et al., 2016; Lambshead et al., 2001; Olu et al., 2017; Tietjen, 1984) and higher regional diversity (Briggs et al., 1996; Frutos and Sorbe, 2017; Van Gaever et al., 2009; Glover et al., 2001; Hess and Jorissen, 2009; Olu et al., 2017; Tietjen, 1984) seems to exist, the driving factors underlying these patterns are unclear. This review has indicated that turbidity flows may be influencing benthic communities via increased carbon or other nutrient availability, or due to physio-chemical characteristics of the sediments, or some combination thereof (reflective of organic matter degradation and resultant hypoxic or toxic conditions). Attempting to understand how turbidity flows impact the deep sea by looking at ancient turbidites is confounded by other natural processes, such as the flux of carbon from the surface to the benthos, and the hundreds to thousands of years since the turbidity flows occurred. Further, the hypotheses considered here focus on the distal deposition environment, usually abyssal habitats (the environment that Heezen et al. (1955) originally specified), but turbidity flows and studies of them evaluate the impact to the benthic communities all along the flow's path. In Fig. 1b and c we propose potential along-path patterns in proxies of productivity and diversity for a large, generalized turbidity flow in a canyon at three different time points following the event.
A better way to test the hypothesis of productivity of Heezen et al. (1955) and the hypothesis of regional diversity of Angel and Rice (1996), and to understand why these impacts are occurring and in which environments, is to look at more recent turbidity flows such as those triggered by the 1999 storm in Cap Breton Canyon, France; periodic turbidity flows induced by river flooding in the Congo Channel, SE Atlantic; and the 2011 Tōhoku Earthquake off Japan. However, even these studies lack sufficient pre-disturbance data or the spatial spread of data to interpret the impacts of the turbidity flow on the benthic community.
A recent turbidity flow event in Kaikōura Canyon, New Zealand, triggered by the 2016 Kaikōura earthquake (Mountjoy et al., 2018) may provide an ideal opportunity to test turbidity flow hypotheses. About 10 years before the turbidity flow, the canyon head to depths of 1300
Figures 4 and 5 were produced using the ggplot library in R. All data used in this review are available via the published literature referenced in Table 1.
KTB conducted the literature review, wrote the first draft of the paper, and constructed the figures and table. AAR, DL, and DAB contributed to the manuscript drafting plan and figure and table design, wrote portions of the text, and edited drafts of the manuscript. All authors edited and approved the final paper text.
The authors declare that they have no conflict of interest.
We thank the two anonymous reviewers for their constructive comments of the paper. Katharine T. Bigham was supported by the NIWA-VUW PhD scholarship in marine science. Ashley A. Rowden, David A. Bowden, and Daniel Leduc were funded by the Physical Resources Programme through NIWA's Coasts and Ocean Centre.
This paper was edited by Hiroshi Kitazato and reviewed by two anonymous referees.