South East Atlantic Seamounts
Global datasets of predicted large seamount locations have been created from ocean bathymetry. These data were combined with historical catch data from seamount fisheries and other anthropogenic marine impacts to identify areas of low impact including the waters around the Discovery tablemount group in the South East Atlantic.
Download this reportSouth East Atlantic Seamounts.pdf — PDF document, 397Kb
FFormed by tectonic and volcanic activity, seamounts may act to disrupt normal oceanographic conditions, leading to an increase in vertical mixing and circulation (Roden 1987). Such mixing can influence the productivity of the water column above seamounts (White et al. 2007) and encourage the development of productive ecosystems.
While the idea that all seamounts are highly productive environments has come into question in recent scholarship (White et al. 2007), the existence of highly productive “seamount fisheries” confirms that this concept is valid in many areas. Beginning in the late 1960s, seamount fisheries have seen major expansions both in terms of fishing effort and their geographic range over time (Watson et al. 2007). However, many seamounts are uncatalogued scientifically and untouched by fishing gears.
As fishing is potentially the single largest human disruption affecting most seamounts, a comparison of known seamount locations, reported seamount bottom trawl catch, and their proximity to other anthropogenic impacts can inform the evaluation of the “naturalness” of a given seamount or seamount group.
This illustration shows it is possible to identify seamounts in areas with low anthropogenic impacts using available global datasets. In a given biological province/region, less impacted seamounts can potentially serve as a reference baseline. Such a reference baseline is crucial for quantifying anthropogenic impacts elsewhere and to evaluate seamount fishery management options.
How the area with a comparatively higher degree of naturalness was identified
The naturalness criterion seeks to identify areas in the global ocean that have very low or absent levels of anthropogenic impact. Here, three global datasets are used to estimate the naturalness of areas around seamounts:
A global dataset of predicted locations of large seamounts was obtained (based on Kitchingman & Lai 2004). These data were imported into a Geographic Information System (GIS) and then combined based on location, with a dataset of total historical trawl catch on seamounts obtained from the Census of Marine Life Seamounts project (CenSeam) (Clark et al. 2007). This fishing dataset represents seamount-only catch tonnage in 1-degree cells. Since it is possible that several seamounts could exist in any cell, a count of seamounts in each fishing catch cell was performed and used to equally apportion the catch figures over the number of proximate seamounts. In addition, the cumulative scores for anthropogenic impacts on seamounts (demersal destructive fishing, artisanal fishing, sea temperature, ocean acidification, ocean based pollution, benthic structures) were obtained from a global model of human impacts on seamounts (Halpern et al. 2008) and combined with the previous data.
This combined dataset thus contained several metrics of anthropogenic impacts at each predicted seamount location. An initial filtering of the seamounts dataset to only those with summits < 2000m deep was performed (2000m being a current depth limit of existing trawling technology). Seamounts greater than 2000m deep could also be considered as natural. These locations were not included in this analysis as their naturalness is presently beyond the considered impact drivers of this illustration.
Seamounts < 2000 m deep were then identified in areas where seamount catch was known, and was less than 50 tons, and had a “Very Low” to “Low” anthropogenic impact score (Halpern et al. 2008). Only 0.6% of these seamounts met these criteria, including seamounts in the Discovery tablemount group in the South East Atlantic off the west coast of South Africa. Seamounts in this group range from having no catch data reported to very little catch (<25 tons). There are no commercial fisheries currently established with most exploratory trawling taking place in the late 1970s and early 1980s (Clark et al. 2007). The Discovery seamounts have also been identified by the South East Atlantic Fisheries Organization (SEAFO) as “unexploited” and were closed to bottom fishing in 2006 (Bensch et al. 2008). In addition, some information on fisheries bycatch is available (Stocks 2009). For these reasons, we identified the Discovery tablemount group as a region of interest and as a potential ecologically significant area illustrating the naturalness criterion for those seamounts occurring within the depth range of commercial trawling.
Sources of Data
For this illustration we used seamounts data from a predicted seamounts dataset by Kitchingman and Lai (2004). This dataset is derived from global bathymetry data itself derived from soundings, sonar observations, and satellite altimetry data. Over 14,000 large (>1km in height) predicted seamounts are contained in this database (see http://seaaroundus.org/ecosystemsmaps/default.aspx to download these data). We excluded seamounts in this dataset where the estimated summit depth was <10 m, as many of these are known to be mis-identified as seamounts, in fact being low-lying atolls in the Central-Western Pacific Ocean. The final dataset totaled about 13,000 features. This number is thought to be conservative, with some other estimates of global seamount counts ranging upwards of 100,000 (Wessel 2001).
These data were supplemented by historical global seamount catch data compiled by CenSeam (Clark et al. 2007). Although these data are known to be incomplete (data from some nations was unavailable), they represent the most comprehensive confirmed data for seamount fisheries to date and provided a reasonable indication of seamount trawl catch levels. Because these data are incomplete, only those cells with catch data were initially considered in this analysis. Cells with no data reported may have had no catch or may just be missing catch data. This biases the selection towards lightly impacted seamounts rather than undisturbed seamounts. However, the region of interest contains both seamounts for which catch is reported and seamounts for which no data is reported. This combination relies upon the best available data while hedging against the possibility that no data areas actually represented undisturbed seamounts.
In addition a global model of anthropogenic impact on the global ocean was obtained (Halpern et al. 2008). This model contains a suite of global ecosystem distributions and a variety of marine impact drivers specific to each ecosystem. In areas where ecosystems types overlap, a summed impact approach was taken. This somewhat clouds the impact process connection to any given overlapping ecosystem type.
Combinations of multiple datasets across a geographic range can be accomplished using a Geographic Information System (GIS). Such systems allow for the basic overlay and selection procedures represented within this illustration. Using this methodology, additional datasets describing anthropogenic impacts or fishing effort could be added easily and more elaborate filtering and sub-selection performed.
Seamount fishery exploration began in the late 1960s and 70s, establishing seamounts as a target for many global fisheries (Clark et al. 2007). Prior to more stringent reporting requirements, few data were collected on these early fishing efforts. It is believed that underreporting of fishing effort on seamounts continues to be a problem today (Bensch et al. 2008). Because fishing exploration has exceeded scientific exploration, much of what is known about seamounts is collected concurrent with, or after, the beginning of these extraction impacts.
Despite a spate of recent seamount exploration globally, most are virtually unexplored scientifically. Only 300-400 have been the target of scientific sampling efforts, with less than 100 considered intensively sampled (Stocks 2009, Clark 2009). Thus, there is very little information on which seamounts have been impacted by anthropogenic activities. In such a situation, modeling efforts, which include the use of what data there are about the effects of anthropogenic impacts, can be used to predict the composition of seabed communities on seamounts across the sparsely sampled ocean domain. Classification systems of seamounts using environmental proxies can also help provide useful information about seamounts in those areas where less is known. Such information can be used to develop the application of the naturalness criterion.
The methodology used in this illustration is readily grasped and depends only on the co-location of predicted seamounts and global impact datasets at a coarse resolution. No explanation of impact processes is here included and this co-location is assumed to indicate a direct negative impact (or lack thereof) on the seamounts. Nonetheless, the location of seamounts, historical seamount catch data, and co-located marine impacts together can serve as an initial assessment of known anthropogenic seamount impacts using the best available global data.