Reconstructing trophic networks across the Toarcian Ocean Anoxic Event

Karolina Zarzyczny

Karolina Zarzyczny

The Palaeontological Association Undergraduate Bursary – PA-UB01703- June-August 2017

Article published in The Palaeontology Association Newsletter No. 97, March 2018, pp. 86-97.


The early Toarcian extinction event was thought to have been driven by an Ocean Anoxic Event (OAE) associated with the Karoo-Ferrar Large Igneous Province (LIP) eruption (McElwain et al. 2005). The aim of this study is to test, through trophic network modelling, whether the early Toarcian extinction event destabilised trophic interactions of Early Jurassic ecosystems and whether the pelagic and benthic ecosystems become decoupled across the OAE (Danise et al. 2015).


The analysis was based on a field database of 38,670 macrofossil specimens, consisting of 162 different benthic and planktonic species collected from the Yorkshire Coast (Cleveland Ironstone, Whitby Mudstone and Blea Wyke Sandstone formations). Sampling was carried out at metre-scale resolution from a 150m section, spanning the upper Pliensbachian-upper Toarcian succession by Little (1990s); Atkinson (2013); and Zarzyczny, Atkinson, Dunhill and Little (2017). Species were grouped into trophic guilds (Table 1.), based on the Bambach et al. (2007) ecospace model. Trophic guild assignments were inferred from modern relatives and, where possible, from palaeoecological interpretations described in literature.

Trophic interaction matrices were coded for four time bins; pre-extinction, extinction (OAE), early recovery and late recovery. Trophic networks were modelled for each time bin using R v.-3.4.0 and R packages ‘igraph’, ‘network’, ‘MASS’, ‘Matrix’, ‘sna’, ‘indices’ and ‘tnet’ (see Dunhill et al., 2016 for references). Link density (LD) was used to measure the connectance of the network whilst the degree centrality (CD) was calculated to measure the importance of each node within the network (Dunhill et al., 2016).

Guild Code
Attached Facult. Motile Deep Infaunal Chemosymbionts AFDC
Attached Facult. Motile Deep Infaunal Suspension Feeders AFDS
Attached Facult. Motile Epifaunal Suspension Feeders AFES
Attached Facult. Motile Erect Suspension Feeders AFERS
Attached Non-Motile Epifaunal Suspension Feeders ANES
Attached Non-Motile Semi-Infaunal Suspension Feeders ANSES
Detritus D
Fast Epifaunal Predator and Scavengers FEPS
Fast Pelagic Predators FPP
Microbenthos M
Plankton P
Slow Deep Infaunal Suspension Feeder SDS
Slow Epifaunal Deposit Feeders SED
Slow Epifaunal Grazers SEG
Slow Epifaunal Predators SEP
Slow Epifaunal Suspension Feeder SES
Slow Pelagic Predators SPP
Slow Shallow Infaunal Chemosymbionts SSC
Slow Shallow Infaunal Miners SSM
Slow Shallow Infaunal Predators SSP
Slow Shallow Infaunal Suspension Feeders SSS
Unattached Facult Motile Shallow Infaunal Miner-Suspension Feeders UFSMS
Unattached Facult. Motile Deep Infaunal Suspension feeders UFDS
Unattached Facult. Motile Epifaunal Suspension Feeders UFES
Unattached Facult. Motile Shallow Infaunal Miners UFSM
Unattached Facult. Motile Shallow Infaunal Suspension Feeders UFSS
Unattached Non-Motile Epifaunal Suspension Feeders UNES
Table 1. Species grouped into trophic guilds.


Trophic guild richness decreased from 23 in the pre-OAE, to 12 during the OAE as deep infaunal, most motile and benthic predatory guilds disappeared (Figure 1a-b.). Across the OAE, the benthic community was dominated by non-motile suspension feeders. Trophic guild richness increased to 16 in the early recovery with the reappearance of some facultatively motile and slow motile benthos (Figure 1c.). Trophic guild richness increased to 23 by the late recovery with the reappearance of a diverse range of motile guilds, and the appearance of new epifaunal and shallow infaunal predators (Figure 1d.).

The overall connectance of the trophic network decreases markedly during the OAE. There is a slight increase in connectance during early recovery followed by a much greater increase in the late recovery to higher levels than before the OAE. Analysis of lost and gained guilds revealed predatory guilds to have the highest connectance with CD ranging from 17 to 19, in comparison to non-predatory guilds ranging CD 1 to 8 (excluding plankton and detritus).

all webs
Figure 1. Trophic networks depicting trophic guild interactions of the marine macrofauna A– pre-OAE (LD = 5.130), B– OAE (LD = 3.667), C– early recovery (LD = 3.750) and D– late recovery (LD = 7.739). The guilds are arranged by their tiering: pelagic, epifaunal, semi-infaunal, shallow infaunal and deep infaunal (top to bottom). The guilds are defined in Table 1.


The reduction in connectance of the trophic networks across the early Toarcian OAE indicates a reduction in ecosystem complexity. Although there is no direct evidence that pelagic and benthic ecosystems became fully decoupled, there is a major loss of benthic trophic richness.

The decrease in connectance across the OAE can be explained by the loss of 11 guilds, some of which were highly connected. The loss of predatory guilds across the OAE suggests the early Toarcian extinction event was a top-down extinction and although there is a severe reduction in guild richness and network connectance across the OAE, there is no indication of a decrease in ecosystem robustness leading to extinction cascades. This observation can be explained by metabolically demanding motile benthic guilds being more sensitive to anoxia (Vaquer-Sunyer and Duarte, 2008) and thus disappearing during the OAE, unlike the less metabolically demanding non-motile guilds which were able to survive. It is likely that due to appearance of new predatory guilds with high connectivity, the connectance of the network is greater during the late recovery (Figure 1b) than it is pre-OAE.


Thanks to Palaeontological Association for funding this project with the Undergraduate Research Bursary PA-UB01703. Thanks to Dr. Alex Dunhill for this opportunity. Thanks to Dr. Crispin Little and Jed Atkinson for guidance and continuous support throughout the project; and the rest of postgraduate researchers in the palaeontology office who were always willing to help. Finally, thanks to Dr. Andrew Beckerman for advice on network ecology.


BAMBACH, R. K., BUSH, A. M. and  ERWIN, D. H. 2007. Autecology and the filling of ecospace: key metazoan radiations. Paleontology 50, 1-22.

DANISE, S., TWITCHETT, R.J. and LITTLE, C.T.S. 2015. Environmental controls on Jurassic marine ecosystems during global warming. Geology 43, 263-266.

DUNHILL, A.M., BESTWICK, J., NAREY, H. and SCIBERRAS, J. 2016. Dinosaur biogeographical structure and Mesozoic contintental fragmentation: a network-based approach. Journal of Biogeography 43, 1691-1704.

MCELWAIN, J. C., WADE-MURPHY, J. and HESSELBO, S. P. 2005. Changes in carbon dioxide during an oceanic anoxic event linked to intrusion into Gondwana coals. Nature 435, 479-482.

VAQUER-SUNYER, R. and DUARTE, C.M. 2008. Thresholds of hypoxia for marine biodiversity. Proceedings of the National Academy of Sciences of the United States of America, 105, 15452-15457