Predicting the distribution of coldwater corals & protections to date Jeff Ardron Director High Seas Marine Conservation Biology Institute The importance of the Consultative Process in light of previous topics undertaken, and its important contribution to the oceans agenda Wednesday, 17 June 2009 UN Headquarters, New York City Alex Rogers Andrew Davies John Guinotte Derek Tittensor
Outline 1. Background: Seamount coral habitat as an example of a Vulnerable Marine Ecosystem (VME) 2. Results: Predicting the distribution of coldwater corals and protections to date
UN Resolution 61/105 80. Calls upon States to take action immediately [...] to sustainably manage fish stocks and protect vulnerable marine ecosystems, including seamounts, hydrothermal vents and cold water corals, from destructive fishing practices, recognizing the immense importance and value of deep sea ecosystems and the biodiversity they contain;
Seamounts, corals, and their importance to pelagic ecosystems (and also fisheries) Some seamounts are fished Some seamounts have corals
Seamounts are significant habitats for many ocean predators Cetaceans Sharks Tunas and billfish Cephalopods Pinnipeds Turtles Seabirds Photoes: Greenpeace, National Geographic, Monterey Whale Watching, NOAA, AD Rogers
Demersal / bentho-pelagic fish Chimaera Hatton Bank (DTi SEA Prog.) Spectrunculus Davidson (NOAA/MBARI) Alfonsino (NOAA) Oreo (NOAA)
Seamounts harbour distinct communities with a high diversity Tasmanian Seamounts 297 species of animals on 14 seamounts 16-33% are new to science Low overlap in species present on different seamounts Even lower co-occurrence of species on different chains
Trawling Trawling gear is extremely robust doors are several tonnes each.
Tasmanian Seamounts
Tasmanian Seamounts - fished
Evidence of fishing
Octocorals up to >4000 years old Recovery will be slow (~ 100s 1000s years)
Distribution of live coral Fished Graveyard Morgue + survey location; O corals found Unfished Clark & Rowden in Prep. Diabolical Gothic
Sustainability of Fishing..? Time series of bottom fisheries catches by depth (Morato et al. 2005. Fish & Fisheries 7: 24-34)
Deep-sea commercial species
Deep-sea fisheries: NW Atlantic example All 5 species Critically endangered under IUCN criteria. Devine et al. (2006) Nature 439: 29
UNCLOS provides the legal framework for deep-sea fisheries on the high seas But scientific guidance is required (i) Low productivity species and (ii) Fisheries where gear contacts the seabed in the normal course of operations. Dealing specifically with impacts to Vulnerable Marine Ecosystems including chance encounters. www.iucn.org
Part 2 1. Background: Seamount coral habitat as an example of a Vulnerable Marine Ecosystem (VME) 2. Results: Predicting the distribution of coldwater corals and protections to date
How can we find where a species is? Niche The range of environmental conditions (biological and physical) under which an organism can exist. Tolerance The ability of organisms to exist at the fringes of its niche.
Data Presences Environmental data 95% Processing Model 2% Output Idealised niche = Possible distribution 3%
2008, on our first attempt, We encountered limitations Low resolution of environmental data. Incomplete geographical distribution of environmental data. Paucity of presences. Lack of absences reduces the available modelling techniques. How can we address these?
* Not yet verified by independent measures * Bottom temperature generalised 1 km 2 grid. Constructed using 33 depth layers from World Ocean Atlas.
New methodology Uses Maxent model (Phillips et al. 2006). 26 global variables created at 1km cell size. Geophysical: slope, BTM, rugosity etc. Chemical: carbonate chemistry, nutrients, salinity etc. Biotic: food supply. Physical: currents, temperature etc.
New results Results presented here are for stony corals (Scleractinia mostly L. pertusa, M. oculata, D. dianthus, S. variablis, G. dumosa and includes approximately 270 other species). In total, we modelled the niche using 4140 records split into 75% training and 25% test data sets. Have data for Octocorals, Antipatharians etc, but not showing those results here.
Variable Unit Name Cell size (x,y) Depth Range (levels) Reference Geophysical variables - 1 Depth m DEPTH 0.0083 Ocean depth (1 level) Becker et al. (in press) 2 Slope m m -1 SLOPE 0.25, 0.2 Ocean depth (1 level) Becker & Sandwell (2008) 3 Rugosity - RUGOS 0.0083 Ocean depth (1 level) Derived from Becker et al. (in press) 1 4 Aspect eastings - ASPE 0.0083 Ocean depth (1 level) Derived from Becker et al. (in press) 2 5 Aspect northings - ASPN 0.0083 Ocean depth (1 level) Derived from Becker et al. (in press) 2 6 Bathymetric position index - BTM1 0.0083 Ocean depth (1 level) Derived from Becker et al. (in press) 1 1 5 1 7 Bathymetric position index - BTM2 0.0083 Ocean depth (1 level) Derived from Becker et al. (in press) 1 2-10 1 8 Slope 2 SLOPE2 0.0083 Ocean depth (1 level) Derived from Becker et al. (in press) 1 Hydrographic variables 9 Regional current flow cm s -1 REGFL 0.5 5-5374m (40 levels) Carton et al. (2005) 3 10 Vertical flow m/s VERTFL 0.5 5-5374m (40 levels) Carton et al. (2005) 3 Chemical variables 11 Alkalinity μmol cm -3 ALK 12 Apparent oxygen utilisation 13 Aragonite Ω arag ARAG 14 Calcite Ω calc CALC 3.6, 0.8-1.8 6-4775 m (25 levels) Steinacher et al. (2008) 4 ml l -1 AOXU 1 0-5500 m (33 levels) Garcia et al. (2006a) 3.6, 0.8-1.8 3.6, 0.8-1.8 6-4775 m (25 levels) Steinacher et al. (2008) 4 6-4775 m (25 levels) Steinacher et al. (2008) 4
Database of all coral records on seamounts Rogers et al., 2007; Clark et al. (2006)
Comparing two different algorithms: MaxENT and ENFA predicted habitat suitability for Scleractinia Model results similar but MaxEnt performed significantly better than ENFA
Results: Predicted global distribution of stony corals 1 2 3
1. N Atlantic Closures
2. S Atlantic Closures
3. S Indian Ocean Vol. Closures
What are the major environmental drivers? Combination: Carbonate chemistry. Temperature. Salinity. Nutrients. Dissolved O 2. Topography. Food supply. Caveat: Habitat prediction can only provide an indication of niche. It does not resolve temporal variability, competitive exclusion/competition. We still need experiments and observations.
Conclusions Better accuracy & precision: Species - habitat prediction has improved greatly over the past couple years; Ad hoc approach to fisheries closures: Fisheries management has to date relied on anecdotal information; Incomplete protection: Current VME protections appear to be v. incomplete; Collaboration required: Ecologists and fisheries scientists have different skills which need to be combined in order to meet the requirements of UNGA Res. 61/105
Acknowledgements Thanks to the two analysis teams: Dr.s Andrew Davies & John Guinotte Dr.s Alex Rogers & Derek Tittensor Thank you Jeff.Ardron@MCBI.org