Ecological Applications, 19(3), 2009, pp. 761–773
? 2009 by the Ecological Society of America
Trawl disturbance on benthic communities:
chronic effects and experimental predictions
HILMAR HINZ,1VIRGINIA PRIETO, AND MICHEL J. KAISER
School of Ocean Sciences, University of Wales, Bangor, Menai Bridge, Anglesey, LL59 5AB United Kingdom
While the direct impacts of trawl disturbances on benthic communities have been extensively
studied, the consequences from long-term chronic disturbances are less well understood. The
response of benthic macrofauna to chronic otter-trawl disturbance from a Nephrops norvegicus
(Norway lobster) fishery was investigated along a gradient of fishing intensity over a muddy
fishing ground in the northeastern Irish Sea. Chronic otter trawling had a significant, negative
effect on benthic infauna abundance, biomass, and species richness. Benthic epifauna
abundance and species richness also showed a significant, negative response, while no such
effect was evident for epibenthic biomass. Furthermore, chronic trawl disturbance led to clear
changes in community composition of benthic infauna and epifauna. The results presented
indicate that otter-trawl impacts are cumulative and can lead to profound changes in benthic
communities, which may have far-reaching implications for the integrity of marine food webs.
Studies investigating the short-term effects of fishing manipulations previously concluded that
otter trawling on muddy substrates had only modest effects on the benthic biota. Hence, the
results presented by this study highlight that data from experimental studies can not be readily
extrapolated to an ecosystem level and that subtle cumulative effects may only become
apparent when fishing disturbances are examined over larger spatial and temporal scales.
Furthermore, this study shows that data on chronic effects of bottom trawling on the benthos
will be vital in informing the recently advocated move toward an ecosystem approach in
fisheries management. As bottom-trawl fisheries are expanding into ever deeper muddy
habitats, the results presented here are an important step toward understanding the global
ecosystem effects of bottom trawling.
Bottom trawling has widespread impacts on benthic communities and habitats.
impacts; Irish Sea; Nephrops fishery; Nephrops norvegicus; Norway lobster; otter trawling.
benthic regime shifts; chronic fishing disturbance; estimation of fishing effort; fishing
Over the past two decades, intensive research focused
on the effects of towed bottom-fishing gear on marine
benthic communities (Collie et al. 2000, Kaiser et al.
2006) has led to the recognition of bottom trawling as
one of the greatest and most widespread causes of
anthropogenic change in shelf seas (Jennings and Kaiser
1998, Bergman and van Santbrink 2000, Kaiser et al.
2002). The direct physical impact of towed bottom
trawls causes varying levels of disturbance by altering
seabed morphology, removing, damaging, or killing
biota, ultimately leading to substantial changes in
benthic community structure (Auster and Langton
1999, Kaiser et al. 2002).
While the mechanisms by which bottom-trawl distur-
bance directly affects benthic communities are well
understood (Kaiser et al. 2002), quantifying the effects
of fishing impacts under realistic conditions has re-
mained challenging, as the spatial and temporal scale of
disturbance generated by an entire fishing fleet is
unfeasible in an experimental context. Accordingly,
those fishing impact studies conducted to date typically
have been small-scale experimental studies at spatial
scales measured in ha cf. km2(Collie et al. 2000, Kaiser
et al. 2006). In such studies, an area of supposedly
pristine seabed has been experimentally trawled by
single or multiple passes of the fishing gear under
investigation, and the response of the benthic biota to
the disturbance, as well as its recovery through time,
have been quantified. The findings of such studies have
provided useful insights into the instantaneous effects
(e.g., direct mortality) and relative severity of impacts
for different gear types on various benthic habitats
(Collie et al. 2000, Kaiser et al. 2006). Most importantly,
these studies have revealed the importance of the
interaction between fishing-gear type and habitat, and
the timescales involved in recovery following fishing-
gear disturbance (Kaiser et al. 2006).
Nevertheless, experimental studies have significant
limitations such that care is required when extrapolating
their results to ecosystem-level effects, which will be
required if fisheries managers are to adopt the recently
advocated move toward an ecosystem approach in
Manuscript received 19 February 2008; revised 24 June 2008;
accepted 12 August 2008. Corresponding Editor: K.
fisheries management (Link 2002a, b, Pikitch et al.
2004). Physical disturbance from trawling on fishing
grounds occurs chronically over large spatial scales, and
therefore can be expected to lead to more severe effects
and much slower recovery rates than assumed from
experimental studies. The recovery of chronically
disturbed benthic communities on fishing grounds will
be dependant on recruitment and population growth
rather than on immigration from adjacent untrawled
areas, which is likely to be more important in smaller
scale experimental trawling studies (Hiddink et al.
2006a). Therefore, to assess the effects of trawling
disturbance at the relevant scale, it is necessary to test
the predictions based on experimental studies in areas
subjected to active fisheries.
Due to a lack of spatially resolved, quantifiable
fishing-effort data, past investigations into the effects
of chronic fishing disturbance were restricted to a
comparison of areas defined as either high or low
disturbance or unfished control areas (Simboura et al.
1998, Ball et al. 2000, Smith 2002). As such classifica-
tions of fishing intensities are relative measures that are
specific to each study, the findings are of limited
generality and comparability. Furthermore, areas cho-
sen as untrawled control sites are often inaccessible to
the fisheries due to the presence of a fishing obstruction
(e.g., wrecks, rock reefs, or similar structures), which
generate potential confounding effects (Davis et al.
1982, Page 1997, Bomkamp et al. 2004). Similarly,
confounding effects arise in studies which, due to a lack
of untrawled areas within the fishing ground, sampled
remote control sites that experience different environ-
mental conditions (e.g., see Simboura et al. 1998).
Additionally, due to a general lack of comparable
control sites on fishing grounds most of these studies
lack sufficient replication to allow for firm conclusions
about the effects of chronic fishing disturbance.
The advent of vessel monitoring systems in northern
European fisheries has provided a unique opportunity to
examine the response of biota to a gradient of fishing
activity at an appropriate scale (Jennings et al. 2001,
Hiddink et al. 2006b, Queiro ´ s et al. 2006, Hinz et al.
2008). These comparative studies are imperative as they
enable us to quantify the effects of trawl disturbance
under realistic conditions and allow us to test the
predictions generated from smaller scale experimental
manipulations (Kaiser et al. 2006). Furthermore, these
studies are critical if we are to inform fisheries managers
and policy makers as to the potential outcome of
different fishing effort management scenarios.
In the present study, we quantified the response of
macrofaunal abundance, biomass, species richness, and
community composition along a gradient of otter-
trawling disturbance from a Nephrops norvegicus [Nor-
way lobster; see Plate 1] fishery over a muddy fishing
ground in the northeastern Irish Sea. Fishing effort was
estimated using high resolution fishing-effort data from
sea fisheries inspectorate over-flights. Confounding
effects related to sampling across different habitat types
(highlighted by earlier studies as limiting the formula-
tion of firm conclusions [Jennings et al. 2001, Queiro ´ s et
al. 2006]) were avoided by ensuring a high degree of
similarity in environmental conditions between sampling
sites. The response of community metrics were com-
pared to those ascertained from previous small-scale
fishing impact studies (Kaiser et al. 2006).
Muddy bottoms cover over 50% of the earth’s surface
(Gage and Tyler 1991) and the fauna inhabiting this
environment fulfill important ecosystem services with
respect to the cycling of organic carbon and nutrients
(Widdicombe and Austen 1998, Howe et al. 2004,
Widdicombe and Needham 2007). As bottom-trawl
fisheries are expanding onto deeper mud bottoms, this
paper is an important step toward understanding and
assessing the global ecosystem effects of bottom
The long-term effects of chronic trawling on benthic
macrofauna communities were investigated over an
active fishing ground in the northeastern Irish Sea off
the Cumbrian coast (Fig. 1). The main bottom-trawling
activity that occurs within this fishing ground is otter
trawling for Norway lobster (Nephrops norvegicus
Linnaeus) and gadoid fish. The fishery operates
throughout the year with a peak in activity from spring
to early summer (Marine Fisheries Agency, unpublished
data). Due to the proximity to the coast, fishing vessels
tend to operate on day trips, and the majority of vessels
are ,20 m in ‘‘length overall’’ (LOA). The area is
characterized by low-energy hydrodynamic conditions,
and consequently the substratum comprises mostly fine
sand and muddy sediment deposits (Swift 1993). The
area was selected for this investigation as on a regional
scale it displayed reasonably similar habitat character-
istics (depths, sediment type, bottom temperatures, and
tidal currents) while at the same time showing a strong
spatial gradient in trawling intensity (Table 1, Fig. 1).
Estimation of fishing effort
Estimates of trawling intensities and their spatial
distribution over the fishing ground were derived by
combining over-flight data and log book records of
hours spent fishing per ICES rectangle collected by the
United Kingdom’s Marine Fisheries Agency.
Over-flight data was obtained from surveys designed
to monitor compliance with fisheries regulations. Flights
over the Irish Sea record the geographical position, gear
type, and activity of fishing vessels indiscriminate of
vessel size. From these data, the spatial distribution of
active fishing vessels fishing with relevant bottom gears
was established. The spatial resolution of these data
largely depends on the number of flights conducted. For
a detailed spatial coverage it was required to monitor the
fishing fleet over a period of 5-year period from October
HILMAR HINZ ET AL.762
Vol. 19, No. 3
1999 to October 2004. The mean number of over-flights
over the fishing ground was 3.23 times per month for the
respective period. The density of vessels was calculated
for 1-km2grid cells over the fishing ground within a
geographic information system (GIS) ArcGIS v.9
(ESRI, Redlands, California, USA). For the fishing-
density calculation, each 1-km2grid cell had a search
radius of 3 km from its center, and vessel sightings
within this radius were summed giving the 1-km2grid
cell its density value. This selected radius was used to
produce a smooth generalized fishing density raster from
the underlying point source data (vessel sightings). For
the estimation of fishing frequency as times trawled per
year, the spatial data was combined with data from log
book entries. UK fishermen are by law required to
report, by gear type, how many hours they spent fishing
within International Council for Exploration of the Sea
(ICES) rectangles (30 3 30 nautical miles). These data
are centrally collated by the UK Marine Fisheries
Agency and include the hours spent fishing per month
for each gear type for all UK-registered fishing vessels.
The fishing ground in the Irish Sea lies entirely within
ICES rectangle 37 E6. Thus the total time reported for
bottom trawling in ICES rectangle 37 E6 for the relevant
period and gear types was divided proportionally over
the fishing ground according to vessel density informa-
tion in the 1-km2grid cells. This calculation resulted in
cell values of hours fished that were converted to times
trawled per year by assuming a mean trawl speed of 2.5
knots (4.62 km/h) and a gear width of 60 m. Thus, for
example, a 1-km2grid cell with a total fishing effort of 53
hours over the 5-year period will have been trawled on
average 2.9 times per year (vessel speed 3 gear width 3
hours fished per year).
lobster) fishing ground in the Irish Sea off the Cumbrian coast, UK. Intensity of fishing effort (times trawled/year) is indicated by
different intensities of gray shading.
Location of study area and spatial distribution of sampling sites within the investigated Nephrops norvegicus (Norway
TABLE 1. Summary of fishing effort and environmental characteristics of sampling sites.
April 2009763 CHRONIC TRAWL DISTURBANCE
It should be noted that the fishing-effort values
calculated are estimates based on the best available
data. Smaller scale spatial and temporal variations in
fishing effort which will have been present could
therefore not be accounted for. Data from the European
Community satellite vessel monitoring system (VMS)
were not available, as the fleet operating in this area was
below the minimum size of 24-m LOA for which VMS
Sampling design and sample processing
Sampling of the macrofauna communities was carried
out in November 2004. Within the fishing ground we
chose 20 sampling sites (1 3 2 km rectangles) to cover
the maximum possible range of fishing intensities (Fig.
1). Fishing intensities were determined for each of the 20
sampling sites using the method detailed in Methods:
Estimation of fishing effort. Thus within each site, all
faunal data were assigned the same fishing intensity.
Within each of these sampling sites, eight randomly
distributed 0.1-m2Day Grabs were taken to study
infaunal communities. Samples were sieved over 1 mm
and fixed in 4% buffered formalin solution. In the
laboratory, species were identified to the lowest taxo-
nomic level possible and their blotted wet mass was
recorded. Mean abundance and biomass of infaunal
data per sampling area were standardized to square
meters prior to analysis.
Three randomly distributed 5-minute, 2-m beam-trawl
hauls, with a towing speed of 1–2 knots (1.9–3.7 km/h),
were conducted within each study area to sample the
epibenthos (for beam-trawl specifications see Jennings et
al. ). Retrieved catches were sorted, identified,
counted, and wet weighed with a motion-compensated
balance onboard of the research vessel. As individual
tows tended to vary in length, resulting abundance and
biomass data were corrected for tow length and
standardized to 1000 m2. Epifauna within the present
study included species of flatfish and small demersal fish
caught with the 2-m beam trawl. Pelagic fish caught were
excluded from the analysis.
At four locations within each sampling site, a further
four Day Grabs were taken to collect sediment samples
from which a 50-g surface sample was collected and
frozen. In the laboratory the fraction of particles ,63
lm, i.e., mud (silt and clay), was determined by mass
loss following wet sieving, while the more coarse,
particle-sized fractions of the sediment were determined
using mechanical dry sieving through a stacked set of
Wentworth grade sieves (Holme and McIntyre 1984).
Organic matter content was estimated by mass loss of
;5 g of dried sediments on ignition at 5508C for 6 hours
(Holme and McIntyre 1984).
A conductivity, temperature, and depth (CTD) data
logger was used to assess near-bottom temperature at
the center of each study site to assess the relative
differences between sampling sites. One vertical CTD
drop was performed at each site.
Tidal-bed shear stress data for each of the sampling
areas were derived from a two-dimensional hydro-
graphical model of the Irish Sea. For details on the
shear stress calculations see (Hiddink et al. 2006b).
Analysis of environmental characteristics of sampling sites
Prior to analyzing the response of benthic communi-
ties to trawling intensity, the similarity/dissimilarity of
abiotic habitat characteristics found at respective
sampling sites was analyzed in a multivariate cluster
analysis (Primer v.6, 2006; PRIMER-E, Lutton, Ivy-
bridge, UK). This analysis was carried out to ensure that
only sites of similar habitat type were included in the
further analysis. This approach limited the possibility of
confounding effects of habitat type while assessing the
response of the benthic communities to fishing distur-
bance. For the cluster analysis, the following variables
were used to calculate a Euclidian distance matrix:
depth, mean partial size, percentage of silt and clay,
organic content, shear stress, and near bottom temper-
ature. Variables were normalized before analysis. The
cluster analysis was carried out integrating the SIMPOF
routine (Primer v.6) that determines statistically signif-
icant station clusters within an a priori ungrouped set of
The relationship of environmental variables and
trawling was analyzed using Pearson’s correlations.
These correlations served as an exploratory tool to
identify significant relationships between environmental
variables and fishing intensity. Prior to analysis, median-
grain-size organic content and near-bottom temperature
were log10(n þ 1)-transformed so that the data approx-
imated to normality.
Univariate analysis of benthic community data
Multiple linear regression models were used to
determine if fishing intensity or environmental param-
eters had a significant effect on the measured univariate
community descriptors for infauna and epifauna. Linear
models were not used to find a single ‘‘best’’ model
describing the relationship between predictors (environ-
mental variables and fishing intensity) and dependent
variables (univariate community descriptors), but to
analyze causality between them. Hierarchical partition-
ing, a form of multiple linear regression, was used, as
this method has been described as particularly suitable
for this task (Chevan and Sutherland 1991). Hierarchical
partitioning compares all possible models in a multiple-
regression setting and determines the independent
capacities of the predictive variables to explain the
patterns of variability in the corresponding response
variable. The hierarchical partition approach alleviates
one of the major limitations of multiple regressions,
namely that the contributions of each predictor variable
is dependent upon which other predictor variables
happen to be included in the model (Chevan and
Sutherland 1991, Mac Nally 2000). The independent
explanatory power for each predictor on its dependent
HILMAR HINZ ET AL. 764
Vol. 19, No. 3
variable is characterized by an index I, which reflects the
independent contribution of the predictor to the
variance explained by the models. A second parameter
J measures the interaction between each predictor and
the others. In our study, J values were relatively small
(jI/Jj ratios . 1), and therefore interactions among
predictors were not considered (Mac Nally 2000).
Variables that independently explained a larger propor-
tion of variance than chance were identified using
randomization tests. For each predictor, the observed
contribution to the explained variance (I) was compared
to the distribution of a population of I’s of 1000
randomizations of the data matrix. Significance was
accepted at the upper 95% confidence limit. Hierarchical
partitioning procedures were calculated using the
hierarchical partitioning software for the public domain
package R (Mac Nally and Walsh 2004; available
Predictors that had a significant independent effect on
the response variable were used in an ordinary least-
square linear regression model (OLS) to assess the
overall fit of the model via the r2value.
The responses of univariate community descriptors
(abundance, biomass, and species richness) to trawling
frequency (per year) were further explored using least-
square regression analyses. Prior to analysis abundance
and biomass data were log10(n þ 1)-transformed. The
response in abundance of individual macrofauna taxa
(number per m2for infauna and number per 1000 m2for
epifauna) to trawling intensities was investigated using
least-square regression analysis. The analysis was
performed on log10(n þ 1)-transformed abundance data
of the 20 most dominant infauna and epifauna species.
Step-up false discovery rate (FDR; Garcia 2004) was
applied to the results of this analysis to prevent inflation
of the Type I error due to the use of multiple tests (n ¼
20). The dominance of each species was established by
first ranking abundances of species within every
sampling site (rank 1 signifying the most dominant
species) and subsequently calculating the mean rank
over all sampling sites. Additionally the percentage
occurrence was calculated.
Several studies on fishing disturbances have postulat-
ed that the loss of larger macrofauna would be replaced
to some extent by a proliferation of smaller species such
as polychaetes (Bergman and van Santbrink 2000,
Kaiser et al. 2000, Jennings et al. 2001). To test if
chronic fishing intensities had a positive effect on small
polychaete abundance and biomass data, taxa with a
mean individual wet mass ,0.002 g were pooled for
separate regression analyses. The following eight poly-
chaete taxa were pooled for this analysis: Aphelochaeta
marioni, Chaetozone sp., Eumida sanguinea, Levinsenia
gracilis, Magelona minuta, Pholoe inornata, Poecilochae-
tus serpens, and Prionospio sp.
Multivariate analysis of benthic community data
To investigate the effect of chronic trawl disturbance
on macrofauna composition, infauna and epifauna
abundance data were analyzed using the PRIMER
(v.6) software package. Prior to analysis the data were
square-root transformed to reduce the influence of
dominant species. The transformed data were subse-
quently analyzed using cluster analysis in conjunction
with the SIMPROF routine based on calculations of
Bray-Curtis similarity matrices (Bray and Curtis 1957).
Clusters identified by the SIMPROF test were visualized
using multi-dimensional scaling (MDS) plots. MDS
plots were overlaid with bubble plots that displayed
relative fishing intensity.
Ordination plots and identified clusters were further
analyzed with the BIOENV and SIMPER routines
respectively. The BIOENV analysis was used to identify
which underlying factors (environmental variables mea-
sured at each site and fishing intensity) best correlated
with the observed community ordination patterns. The
BIOENV analysis compares the agreement between the
biotic similarity matrix and the Euclidian environmental
matrix using a weighted Spearman rank correlation
coefficient (qw). Furthermore, the data were tested for
seriation to detect any directional trends in community
pattern (MDS ordination) related to trawling intensity
(Primer v.6 RELATE routine). The SIMPER routine
was used to quantify the percentage contribution that
each species made to the similarity within the clusters
(site groups) and to the difference (dissimilarity) between
Environmental characteristics of sampling sites
The cluster analysis with integrated SIMPROF
(PRIMER v.6 software package) routine identified two
significant site clusters (P , 0.05) which indicated the
presence of two distinct habitats among sampling sites.
Cluster A consisted of 15 sites, and cluster b1 consisted
of three sites (Fig. 2). Stations of cluster b1 together with
two further sites formed part of a larger cluster, B (Fig.
2). In comparison, sites of cluster A were located in
slightly deeper water (31 6 6 m, mean 6 SD) and had
finer sediments with a higher percentage of silt and clay
(67% 6 14%, mean 6 SD), while stations of cluster B
were characterized by shallower depth (30 6 6 m, mean
6 SD) and less fine sediments (sit and clay content 27%
6 8%, mean 6 SD). For a summary of all environmen-
tal variables of respective clusters see Table 2.
For the further analysis, only sampling sites of a
similar habitat type were chosen. As cluster B consisted
of only five stations, a separate analysis into this habitat
type was not feasible due a lack of statistical power.
Further analyses were only performed on the remaining
15 stations of cluster A that had comparable habitat
April 2009765CHRONIC TRAWL DISTURBANCE
Pearson correlations of environmental variables and
fishing intensity of cluster A sites showed that depth was
significantly correlated with fishing intensity (Table 3).
The relationship of depth and fishing effort is intrinsic to
the Nephrops fishery as it specifically targets areas of
greater depth. This is primarily related to the behavioral
adaptations of the main target species Nephrops
norvegicus to ambient light conditions. Nephrops norve-
gicus, a burrowing crustacean, tends to emerge from its
burrow at dawn and dusk when light conditions are low
(Chapman 1980). Deeper waters have prolonged low
light conditions, and hence catch rates of Nephrops are
higher over these grounds (Ball et al. 2000), which is
ultimately reflected in the overall distribution of fishing
effort. As the difference in depth between sampling sites
was relatively small (mean depth 31 6 6 m, mean 6 SD),
its direct influence in structuring benthic fauna, i.e.,
through differences in isobaric pressure, was assumed
negligible and depth was therefore removed from the
further analysis. Besides depth, near-bottom tempera-
ture correlated with fishing intensity but was retained for
the further analysis (Table 3). The sediment descriptor
‘‘silt and clay content’’ correlated significantly with
median particle size (MPS) and near-bottom tempera-
ture (Table 3). As sediment descriptors, MPS, silt and
clay, and organic content are highly interrelated
variables, only the silt and clay content was retained as
a sediment descriptor in the further analysis.
Multiple linear regression analysis in the form of
hierarchical partitioning was used to determine the
independent capacities of the predictive variables
(fishing intensity, silt and clay content, shear stress,
near-bottom temperature) to explain the patterns of
variability in the corresponding response variables:
infauna/epifauna abundance, biomass, and species
richness (Table 3). Fishing intensity was the only
significant predictor for infauna abundance and ac-
counted for 79% of the total independent effect (I). For
infaunal biomass two significant predictors were identi-
fied, fishing intensity (I ¼ 56%) and the silt and clay
content (I ¼ 30%). The species richness of infauna was
affected by trawling intensity and accounted for 68% of
the total independent effect (I). Fishing intensity proved
to be the only significant predictor for the abundance of
B, identified by the cluster analysis of environmental variables.
Summary of abiotic habitat characteristics (mean 6 SD) for station groups, A, b1, and
Environmental variablesA b1B
Median particle size (mm)
Silt and clay content (%)
Organic content (%)
Sheer stress N/m2
Near-bottom temperature (8C)
31 6 6
0.079 6 0.009
67 6 14
4.4 6 2
0.21 6 0.02
10.6 6 0.3
26 6 1
0.157 6 0.031
32 6 1
1.2 6 0.3
0.18 6 0.001
10.4 6 0.3
30 6 6
0.160 6 0.038
27 6 8
1.6 6 1.1
0.19 6 0.03
10.4 6 0.4
routine (PRIMER v.6 software package).
Cluster analysis of environmental variables showing significant station clusters (A and b1) identified by the SIMPROF
HILMAR HINZ ET AL.766
Vol. 19, No. 3
epifauna (I ¼ 39%). No significant predictor was found
for epifauna biomass. Fishing intensity was the only
significant predictor for epifauna species richness with
an independent effect of 56%.
Ordinary least-square linear regression (OLS) models
of the variables with a significant, independent effect on
the response variables were all significant (Table 4) and
displayed relatively high r2values, in particular for
infauna biomass (0.61) and species richness (0.76).
Infaunal abundance, biomass and species richness, and
epifaunal abundance and species richness had significant,
negative relationships with trawling intensity (Table 4).
According to the slopes of these regressions (Table 5,
Fig. 3), infaunal abundance was reduced by 72% from the
lowesttrawling effortrecorded (1.3 times trawled/year) to
the highest (18.2 times trawled/year). For the same range
of trawl intensities, infaunal biomass was reduced by 77%
and species richness decreased by 40%. Epifaunal
abundance from the lowest to the highest trawling
intensity was reduced by 81%. Species richness of
epifauna was reduced by 14%. As univariate community
descriptors were log transformed, it needs to be empha-
sized that the reductions in abundance, biomass, and
speciesrichness aregreatestatlowlevels oftrawling effort
and less severe at higher levels of fishing disturbance.
From the regression analyses of the 20 highest ranking
infauna taxa with respect to abundance, 16 taxa had
negative slopes with increasing trawling intensity of
which nine were significant after FDR correction (Table
6): the phoronidae Phoronis sp.; the polychaetes
Aphelochaeta marioni, Glycera spp., Notomastus sp.,
and Magelona alleni; the bivalve Abra spp.; Nemerteans;
the crustacean Ampelisca spp.; and the echinoderm
Amphiura filiformis. Positive slopes were detected for
four infauna taxa, the polychaete Prionospio spp., the
burrowing shrimp Jaxea nocturna, and for the molluscs
Spisula subtruncata and Corbula gibba. As most of the
individuals of Spisula subtruncata and Corbula gibba
sampled were recently settled juveniles, these trends were
thought unrelated to trawl disturbances. The polychaeta
Prionospio spp. was the only species that had a
significant, positive trend after false discovery rate
(FDR) correction. For the epifaunal taxa, 13 out of
the 20 taxa had negative slopes while seven had positive
slopes. None of these relationships were significant after
FDR correction (Table 6). The pooled abundance and
biomass data did not demonstrate significant responses
to increases in trawling disturbance; indeed, the slopes of
these relationships were negative.
The cluster analysis and SIMPROF test identified two
significant main clusters for both infauna and epifaunal
abundance data (Fig. 4). For the infauna (I) clusters, I1
and I2 were identified at a similarity level of 60%. A
single site station H clustered separately from these
TABLE 3. Correlation analysis of physical factors assessed at sites of cluster A and trawling intensity (times trawled/year).
Silt and clay
content Shear stress
Median particle size
Silt and clay content
?0.231 (0.408)0.142 (0.613)
?0.974 (0.000)*** ?0.051 (0.858) ?0.001 (0.997)
0.100 (0.724) ?0.067 (0.812)
?0.267 (0.336) ?0.318 (0.248)0.302 (0.274)
?0.552 (0.033)* ?0.410 (0.129)
Notes: Pearson correlation coefficients are shown with P values in parentheses. Significant correlations are indicated by asterisks.
* P , 0.05; ** P , 0.01; *** P , 0.001.
TABLE 4.Results of the hierarchical partitioning analysis and linear model of predictors.
Hierarchical partitioning analysis: independent contribution (I%) by predictorLinear model, OLS?
Trawling intensitySilt and clay content Shear stressNear-bottom temperaturer2
Notes: The table presents hierarchical partitioning analysis describing explanatory capacity of fishing effort and four
environmental parameters to explain patterns of variability in the corresponding response variables (abundance, biomass, and
species richness of infauna and epifauna). Plus and minus signs indicate the direction of the response variable. For the linear model
of predictors with a significant explanatory contribution, significant results are indicated by asterisks; n.a., not applicable.
* P , 0.05; ** P , 0.01; *** P , 0.001.
? Ordinary least-square linear regression.
April 2009767 CHRONIC TRAWL DISTURBANCE
station groupings. The clusters E1 and E2 were
identified for the epifauna at a similarity level 55%.
Station R clustered separately from the two main
clusters (Fig. 4).
The BIOENV analysis for the infaunal abundance
data showed that the ordination was best explained by
two variables trawling frequency and silt and clay
content (qw ¼ 0.602). The best single factor that
explained the ordination was trawling intensity with a
qw of 0.562 (silt and clay content qw ¼ 0.214). The
ordination of the epifauna abundance data for all
stations was best explained by two variables: trawling
and species richness) and fishing intensity for infauna and epifauna.
Results of linear regressions describing the relationship between univariate community descriptors (abundance, biomass,
Note: Significant correlations results are indicated by asterisks.
* P , 0.05; ** P , 0.01; *** P , 0.001.
infauna, and no./1000 m2for epifauna); log10-transformed biomass (wet mass, originally measured as g/m2for infauna, and g/1000
m2for epifauna), and species richness with fishing effort (times trawled/year) for (a) infauna and (b) epifauna over stations
characterized by muddy sediments.
Relationship of univariate community descriptors, log10-transformed abundance (originally measured as no./m2for
HILMAR HINZ ET AL.768
Vol. 19, No. 3
each infauna and epifauna group).
Results of regression analysis performed on individual infauna and epifauna species (total number of tests ¼ 20 within
SpeciesCode RankOcc. (%) SlopeInterceptr2
Processa novelli holthusii
Notes: Significant results are indicated by asterisks after step-up false discovery rate (FDR) correction; after FDR, some of the
species at P , 0.05 were not significant. Species were ranked after dominance. Occ. (%) indicates percentage occurrence over
sampling sites analyzed. Abbreviations are: C, Crustacea; E, Echinodermata; M, Mollusca; N, Nemertea; P, Pisces; Ph,
Phoronidae; Po, Polychaeta.
Infauna clusters (I1 and I2) are signified by broken lines as identified by the SIMPROF analysis (similarity of stations within
clusters 60%). Epifauna clusters (E1 and E2) are signified by broken lines as identified by the SIMPROF analysis (similarity of
stations within clusters 50%).
Multidimensional scaling plots of infauna and epifauna abundance data with fishing intensity overlaid as bubble plots.
April 2009769CHRONIC TRAWL DISTURBANCE
frequency and bed shear stress (qw ¼ 0.618). Trawling
intensity was the best single factor explaining the
ordination with a qw of 0.506 (tidal bed shear stress
qw ¼ 0.336).
Results of the seriation test showed that there was a
significant, directional trend that related to increasing
trawling intensity within the ordination plot generated
for infauna (q¼0.562, P¼0.01) and epifauna (q¼0.506,
P ¼ 0.01) abundance data. This directional trend was
clearly visible in multidimensional spacing (MDS) plots
The SIMPER analysis of the infauna abundance data
showed that both clusters I1 and I2 were characterized
by the same four dominant species: a species of
Phoronidae, Phoronis sp., the polychaetes Nephtys spp.
and Prionospio spp., and the bivalve Spisula subtruncata.
The five top-ranking taxa contributing highest to the
dissimilarity between the clusters I1 and I2 (percentage
contribution in brackets) were Phoronis sp. (11.4%),
Spisula subtruncata (6.5%), Aphelochaeta marioni (3.3%),
the brittle star Amphiura filiformis (3.3%), and Magelona
alleni (2.5%). For all species, abundances were higher in
I1 (sites with relatively low fishing intensity 4.1 6 3.2
times trawled/year, mean 6 SD) compared to I2 (sites
had a fishing intensity of 13.2 6 3.2 times trawled/year,
mean 6 SD), except for Spisula subtruncata which
showed the inverse relationship. For the epifauna,
cluster E1 (fishing intensity 3.0 6 2.9 times trawled/year,
mean 6 SD) was characterized by the echinoderms
Asterias rubens and Ophiura ophiura, the shrimp
Processa nouveli holthuis, the Opistobrancia Philine
aperta, and the small flatfish species Buglossidium
luteum. Cluster E2 with a higher fishing intensity
compared to E1 (12.7 6 3.9 times trawled/year, mean
6 SD) was characterized by Asterias rubens, the shrimps
Crangon allmanii and Processa nouveli holthuis, the sand
goby Pomatoschistus minutes, and the hermit crab
Pagurus bernhardus. The highest contributions to the
dissimilarity between the clusters E1 and E2 were made
by Asterias rubens (20.2%), Ophiura ophiura (17.7%),
Processa nouveli holthuis (5.8%), Philine aperta (5.4%),
and Astropecten irregularis (4.7%). Mean abundances
were all higher in cluster E1 compared to E2.
The present study demonstrated that soft-sediment
habitats with their predominantly infaunal benthic
communities can be strongly affected by chronic
trawling activities. Otter trawling from a Nephrops
fishery had a significant, negative effect on benthic
macrofauna communities. Benthic infauna displayed
consistent negative responses to trawling for all univar-
iate community descriptors. From the lowest to the
highest trawling intensity, 1.3 to 18.2 times trawled/year,
infauna abundance was reduced by 72%, biomass by
77%, and species richness by 40%. While epifaunal
abundance and species richness had similar negative
responses as observed for the infauna, reductions of 81%
and 14%, respectively, no such effect was found for
epifauna biomass. Epifauna biomass at the high
trawling intensity sites was dominated by large speci-
mens of the common starfish Asterias rubens, which
possibly as a response to elevated food availability in
form of biota killed or damaged by trawling, aggregated
at these sites. The contribution of these high biomass
individuals perhaps masked some of the effects of
burrow. (Right) Side scan image of the seafloor at station M taken during a survey conducted in 2003 with a side-scan-sonar system
C-Max 800. The total swath width of the scan as seen in this image was 200 m. Photo credits: H. Hinz.
(Left) The main target species of the fishing ground under investigation, Nephrops norvegicus, emerging from its
HILMAR HINZ ET AL.770
Vol. 19, No. 3
trawling on community biomass. Other empirical studies
have demonstrated that starfish species respond rapidly
to prey availability (Freeman et al. 2001), and starfish
are known to be resilient to the damaging effects of
trawls (Ramsay et al. 2001). A similar equivocal
response in the epibenthic biomass was reported in a
study conducted across a fishing gradient in the North
Sea (Jennings et al. 2001). Overall abundances of
common species were reduced at high trawling intensity
sites, while species known as vulnerable, such as the
brittle stars Amphiura filiformis and Ophiura ophiura,
were either present in very low abundances or absent.
However, species which showed no response or had a
positive trend in abundance with increasing trawling
intensity, such as the small opportunistic polycheate
Prionospio spp. and the burrowing crustacean Jaxea
nocturna, characterized the high trawling intensity sites.
It should be noted that the fishing ground studied
experienced relatively high fishing intensities compared
to the spectrum of fishing intensities reported by earlier
studies (Jennings et al. 2001, Hiddink et al. 2006b,
Queiro ´ s et al. 2006). Furthermore the communities
found at sites of low fishing intensities were most likely
already in a chronically disturbed state, as the whole
area has been fished since the latter half of the 19th
century. Even muddy bottoms like this area probably
supported biogenic epifaunal structures, founded either
on large shells such as from Arctica islandica or growing
on the tubes of polychaetes species such as Chaetopterus
spp. (Thrush et al. 2001). Most of these biogenic
structures would have been eliminated by the first few
trawl passes. Some indicators of a more pristine state
came from the very low trawling intensity sites found at
the margins of the fishing ground, which due to slight
differences in habitat conditions, were not included in
the formal analysis. At these sites, species typically
found on silty grounds, such as the erect sea pen
Virgularia mirabilis and the brittle star Amphiura
filiformis, occurred in considerable abundances.
Experimental fishing manipulations using otter trawls
on muddy sediments to examine the short-term effects of
known levels of disturbance on the benthic biota have
thus far reported relatively modest changes in benthic
communities (Tuck et al. 1998, Hansson et al. 2000,
Lindegarth et al. 2000, Sanchez et al. 2000, Sparks-
McConkey and Watling 2001, DeBiasi 2004). All of
these studies identified changes in community structure
following trawling treatments; however, overall effects
on abundance, biomass, and diversity at a community
level were largely inconsistent between studies. Tuck et
al. (1998) reported a general increase in benthic infaunal
abundance following an 18-month trawling experiment
while, in an experiment of similar design, Hansson et al.
(2000) reported on a significant, negative effect of
trawling for only a specific group of echinoderms
(ophiuroids). Although all experimental studies identi-
fied specific species that were negatively or positively
affected by otter trawling, they concluded that otter
trawling on muddy sediments had only a limited overall
effect on benthic communities. This finding was
corroborated by the results of a meta-analysis conducted
on experimental fishing impact studies (Kaiser et al.
2006) which concluded that otter trawling on muddy
sediments, compared to other combinations of gear and
substrata, had one of the least negative impacts on the
benthic biota. The initial reduction of benthic taxa
following otter trawl disturbance on mud was estimated
to be ?29%. The only other combination of gear and
substrata types with even lower initial impacts were otter
trawling on sand (?15%) and otter trawling on gravel
(þ3%; Kaiser et al. 2006).
Other studies that have investigated the effect of
chronic otter-trawl disturbance on muddy substrata,
although compromised in their validity through con-
founding experimental designs (see Introduction), did
report more severe negative responses of benthic
communities from otter trawling similar to the present
study. Ball et al. (2000) showed that the seabed protected
by wrecks on a Nephrops fishing ground in the western
Irish Sea had much higher abundances and biomasses of
benthic biota compared to the chronically disturbed
areas. Similarly, Smith et al. (2000) showed that a high-
intensity trawling lane had significantly lower benthic
abundances, biomass, and species richness compared to
areas of relatively low fishing intensity. However, in
contrast, a more recent study by Simpson and Watling
(2006) investigating the chronic effect of a shrimp fishery
in the Gulf of Maine did not find significant differences
between chronically disturbed and untrawled areas.
However, the shrimp fishery in the Gulf of Maine is
highly seasonal, restricted to only 3–4 months per year
during late winter to early spring, which may not be
comparable to fishing-effort levels experienced in the
other areas where fishing occurs throughout most of the
year (Ball et al. 2000, Smith et al. 2000, and the present
study). In this context, our findings are novel because
they demonstrate strong negative responses of commu-
nity matrices and the majority of species over a
quantified gradient in fishing intensities from otter
trawling. The strength of the relationship is undoubtedly
linked to the power of our sampling design and the wide
range of trawling intensities studied (see Plate 1). In
contrast, many of the studies just cited had low
statistical power to detect fishing effects, were unavoid-
ably pseudoreplicated, and lacked quantifiable fishing-
effort data limiting the comparability and generality of
When comparing the results from short-term manip-
ulative field experiments and chronic-fishing-disturbance
studies, it can be concluded that while the initial impacts
of the otter-trawl gear on muddy habitats are relatively
modest, the cumulative long-term disturbance can lead
to profound changes in benthic communities.
Within the meta-analysis on manipulative fishing
impact experiments conducted by (Kaiser et al. 2006),
a fishing-impact level of once per annum was assumed to
April 2009771CHRONIC TRAWL DISTURBANCE
be representative of most fishing systems, in line with the
majority of experimental studies that monitored single
disturbance events, while fishing hotspots have been
assumed relatively rare (Dinmore and Jennings 2004).
However, it is clear that fishing pressures can be intense
in some coastal areas where, for example due to the
habitat preference of the target species, e.g., Nephrops
norvegicus, the fishing ground is restricted to a relatively
small area. In addition, coastal fleets are increasingly
composed of small fishing vessels (,20 m) with a limited
range which may increase the pressure on fishing
grounds close to the port of origin. Similarly, effort
restrictions limiting the range of vessels or increases in
fuel prices may increase pressures on coastal fishing
grounds in the future.
Several studies on fishing impacts have postulated
that the loss of biomass and production associated with
large macro-benthic species would release smaller body-
size classes of biota from competition for space and
other resources leading to an increase in production of
this component of the benthos (Bergman and van
Santbrink 2000, Kaiser et al. 2000, Jennings et al.
2001). Within the present study only one species (an
opportunitic polychaete) Prionospio spp. increased
significantly in abundance with increasing trawling
intensity, and overall, small polychaete species did not
respond positively to increasing trawl disturbance.
Similar to these results (Jennings et al. 2001), investi-
gating the effects of beam trawling over a gradient of
fishing intensities could not demonstrate that the loss of
larger macrofauna was compensated for by an increase
in smaller opportunistic species. Indeed, for mud habitat
described in this study, it appears that all size classes
from macro to meiofauna are negatively affected by
trawl disturbance (Hinz et al. 2008). These results have
far-reaching consequences for the integrity of the
benthic food web and supports fears that fishing may
cause a shift in food-web structure and a breakdown of
benthic pelagic coupling, which potentially could lead to
irreversible ecosystem regime shifts (Choi et al. 2004).
The present study highlights some of the important
similarities and differences between small-scale manip-
ulations of fishing disturbance and larger scale gradients
of disturbance at the scale of the fishery. This is
particularly relevant when the habitat and dominant
fauna are more resilient to sporadic disturbance, as in
the case of mud habitats. In such cases, the subtle
cumulative effects of fishing on the benthos may become
apparent only when chronic, wide-scale disturbance at a
variety of scales is examined. The latter has been
possible only since the advent of high resolution vessel
monitoring systems that have enabled us to measure
gradients of disturbance with adequate precision.
Within a management context, this study demonstrat-
ed that information on trawling impacts derived from
manipulative fishing experiments has to be viewed with
caution and can not readily be extrapolated to an
ecosystem level. Impacts from real fisheries, as shown by
this study, can be cumulative and thus in the long term
cause far greater ecological change than could have been
anticipated from the results of experimental studies
carried out to date.
This study was funded by the European Union as part of the
project ‘‘Response’’ contract QLRT-2001-00787. Over-flight
data for the Irish Sea was provided by DEFRA. Shear stress for
the Irish Sea data was provided by Alan Elliott (University of
Bangor, Wales). The officers and crew of RV Prince Madog,
Jan Hiddink, Ivor Rees, Marika Galanidi, Ana Ruiz Frau,
Sylvia de Juan, and Ana Queiros, are thanked for their help in
collecting and processing of samples.
Auster, P. J., and R. W. Langton. 1999. The effects of fishing on
fish habitat. Pages 150–187 in Fish habitat: essential fish
habitat and rehabilitation. American Fisheries Society,
Hartford, Connecticut, USA.
Ball, B. J., G. Fox, and B. W. Munday. 2000. Long- and short-
term consequences of a Nephrops trawl fishery on the benthos
and environment of the Irish Sea. ICES Journal of Marine
Bergman, M. J. N., and J. W. van Santbrink. 2000. Mortality in
megafaunal benthic populations caused by trawl fisheries on
the Dutch continental shelf in the North Sea in 1994. ICES
Journal of Marine Science 57:1321–1331.
Bomkamp, R. E., H. M. Page, and J. E. Dugan. 2004. Role of
food subsidies and habitat structure in influencing benthic
communities of shell mounds at sites of existing and former
offshore oil platforms. Marine Biology 146:201–211.
Bray, J. R., and J. T. Curtis. 1957. An ordination of the upland
forest communities of southern Wisconsin. Ecological
Chapman, C. J. 1980. Ecology of juvenile and adult Nephrops.
Pages 143–178 in J. S. Cobb and B. F. Phillips, editors. The
biology and management of lobsters. Academic Press,
Chevan, A., and M. Sutherland. 1991. Hierarchical partition-
ing. American Statistician 45:90–96.
Choi, J. S., T. F. Kenneth, C. L. William, and K. Drinkwater.
2004. Transition to an alternate state in a continental shelf
ecosystem. Canadian Journal of Fisheries and Aquatic
Collie, J. S., S. J. Hall, M. J. Kaiser, and I. R. Poiner. 2000. A
quantitative analysis of fishing impacts on shelf-sea benthos.
Journal of Animal Ecology 69:785–798.
Davis, N., G. R. Vanblaricom, and P. K. Dayton. 1982. Man-
made structures on marine sediments—effects on adjacent
benthic communities. Marine Biology 70:295–303.
DeBiasi, A. M. 2004. Impact of experimental trawling on the
benthic assemblage along the Tuscany coast (north Tyr-
rhenian Sea, Italy). ICES Journal of Marine Science 61:1260–
Dinmore, T. A., and S. Jennings. 2004. Predicting abundance–
body mass relationships in benthic infaunal communities.
Marine Ecology Progress Series 276:289–292.
Freeman, S. M., C. A. Richardson, and R. Seed. 2001. Seasonal
abundance, spatial distribution, spawning and growth of
Astropecten irregularis (Echinodermata: Asteroidea). Estua-
rine Coastal and Shelf Science 53:39–49.
Gage, J. D., and P. A. Tyler. 1991. Deep-sea biology—a natural
history of organisms at the deep-sea floor. Cambridge
University Press, Cambridge, UK.
Garcia, L. V. 2004. Escaping the Bonferroni iron claw in
ecological studies. Oikos 105:657–663.
Hansson, M., M. Lindegarth, D. Valentinsson, and M.
Ulmestrand. 2000. Effects of shrimp-trawling on abundance
HILMAR HINZ ET AL.772
Vol. 19, No. 3
of benthic macrofauna in Gullmarsfjorden, Sweden. Marine
Ecology Progress Series 198:191–201.
Hiddink, J. G., S. Jennings, and M. J. Kaiser. 2006a. Recovery
status as an indicator of the large scale ecological impact of
bottom trawling. Ecosystems 9:1190–1199.
Hiddink, J. G., S. Jennings, M. J. Kaiser, A. M. Queiro ´ s, D. E.
Duplisea, and G. J. Piet. 2006b. Cumulative impacts of
seabed trawl disturbance on benthic biomass, production and
species richness in different habitats. Canadian Journal of
Fisheries and Aquatic Sciences 63:721–736.
Hinz, H., J. G. Hiddink, J. Forde, and M. J. Kaiser. 2008.
Large scale responses of nematode communities to chronic
bottom-trawl disturbance. Canadian Journal of Fisheries and
Aquatic Science 65:723–732.
Holme, N. A., and A. D. McIntyre. 1984. Methods for the
study of marine benthos. Blackwell Scientific Publications,
Howe, R. L., A. P. Rees, and S. Widdicombe. 2004. The impact
of two species of bioturbating shrimp (Callianassa subterra-
nea and Upogebia deltaura) on sediment denitrification.
Journal of the Marine Biological Association of the United
Jennings, S., T. A. Dinmore, D. E. Duplisea, K. J. Warr, and
J. E. Lancaster. 2001. Trawling disturbance can modify
benthic production processes. Journal of Animal Ecology 70:
Jennings, S., J. Lancaster, A. Woolmer, and J. Cotter. 1999.
Distribution, diversity and abundance of epibenthic fauna in
the North Sea. Journal of the Marine Biological Association
of the United Kingdom 79:385–399.
Jennings, S., and M. J. Kaiser. 1998. The effects of fishing on
marine ecosystems. Advances in Marine Biology 34:201–352.
Kaiser, M. J., K. R. Clarke, H. Hinz, M. C. V. Austen, P. J.
Somerfield, and I. Karakassis. 2006. Global analysis and
prediction of the response of benthic biota to fishing. Marine
Ecology Progress Series 311:1–14.
Kaiser, M. J., J. S. Collie, S. J. Hall, S. Jennings, and I. R.
Poiner. 2002. Modification of marine habitats by trawling
activities: prognosis and solutions. Fish and Fisheries 3:114–
Kaiser, M. J., K. Ramsay, C. A. Richardson, F. E. Spence, and
A. R. Brand. 2000. Chronic fishing disturbance has changed
shelf sea benthic community structure. Journal of Animal
Lindegarth, M., D. Valentinsson, M. Hansson, and M.
Ulmestrand. 2000. Effects of trawling disturbances on
temporal and spatial structure of benthic soft-sediment
assemblages in Gullmarsfjorden, Sweden. ICES Journal of
Marine Science 57:1369–1376.
Link, J. S. 2002a. Ecological considerations in fisheries
management: when does it matter? Fisheries 27:10–17.
Link, J. S. 2002b. What does ecosystem-based fisheries
management mean? Fisheries 27:18–21.
Mac Nally, R. 2000. Regression and model-building in
conservation biology, biogeography and ecology: the distinc-
tion between and reconciliation of ‘‘predictive’’ and ‘‘explan-
atory’’ models. Biodiversity and Conservation 9:655–671.
Mac Nally, R., and C. J. Walsh. 2004. Hierarchical partitioning
public-domain software. Biodiversity and Conservation 13:
Page, H. M. 1997. Importance of vascular plant and algal
production to macro-invertebrate consumers in a southern
California salt marsh. Estuarine Coastal and Shelf Science
Pikitch, E. K., et al. 2004. Ecosystem-based fishery manage-
ment. Science 305:346–347.
Queiro ´ s, A. M., J. G. Hiddink, H. Hinz, and M. J. Kaiser. 2006.
The effects of chronic bottom trawling disturbance on
biomass, production and size spectra of invertebrate infauna
communities from different habitats. Journal of Experimen-
tal Marine Biology and Ecology 335:91–103.
Ramsay, K., M. Bergmann, L. O. Veale, C. A. Richardson,
M. J. Kaiser, S. J. Vize, and S. W. Feist. 2001. Damage,
autotomy and arm regeneration in starfish caught by towed
demersal fishing gears. Marine Biology 138:527–536.
Sanchez, P., M. Demestre, M. Ramon, and M. J. Kaiser. 2000.
The impact of otter trawling on mud communities in the
northwestern Mediterranean. ICES Journal of Marine
Simboura, N., A. Zenetos, M. A. Pancucci-Papadopoulou, M.
Thessalou-Legaki, and S. Papaspyrou. 1998. A baseline study
on benthic species distribution in two neighbouring gulfs,
with and without access to bottom trawling. Marine Ecology
Simpson, A. W., and L. Watling. 2006. An investigation of the
cumulative impacts of shrimp trawling on mud-bottom
fishing grounds in the Gulf of Maine: effects on habitat
and macrofaunal community structure. ICES Journal of
Marine Science 63:1616–1630.
Smith, C. J., K. N. Papadopoulou, and S. Diliberto. 2000.
Impact of otter trawling on an eastern Mediterranean
commercial trawl fishing ground. ICES Journal of Marine
Smith, M. D. 2002. Two econometric approaches for predicting
the spatial behavior of renewable resource harvesters. Land
Sparks-McConkey, P. J., and L. Watling. 2001. Effects on the
ecological integrity of a soft-bottom habitat from a trawling
disturbance. Hydrobiologia 456:73–85.
Swift, D. J. 1993. The macrobenthic infauna off Sellafield
(north-eastern Irish Sea) with special reference to bioturba-
tion. Journal of the Marine Biological Association of the
United Kingdom 73:143–162.
Thrush, S. F., J. E. Hewitt, G. A. Funnell, V. J. Cummings, J.
Ellis, D. Schultz, D. Talley, and A. Norkko. 2001. Fishing
disturbance and marine biodiversity: the role of habitat
structure in simple soft-sediment systems. Marine Ecology
Progress Series 223:277–286.
Tuck, I. D., S. J. Hall, M. R. Robertson, E. Armstrong, and
D. J. Basford. 1998. Effects of physical trawling disturbance
in a previously unfished sheltered Scottish sea loch. Marine
Ecology Progress Series 162:227–242.
Widdicombe, S., and M. C. Austen. 1998. Experimental
evidence for the role of Brissopsis lyrifera (Forbes, 1841) as
a critical species in the maintenance of benthic diversity and
the modification of sediment chemistry. Journal of Experi-
mental Marine Biology and Ecology 228:241–255.
Widdicombe, S., and H. R. Needham. 2007. Impact of CO2-
induced seawater acidification on the burrowing activity of
Nereis virens and sediment nutrient flux. Marine Ecology
Progress Series 341:111–122.
April 2009773 CHRONIC TRAWL DISTURBANCE