allows estimation of the effects of limiting factors on species abun-
dance and the description of potential (or maximal) habitat (Vaz
et al., 2008).
GLM and linear QR were used for modelling the habitat use of
European plaice (Pleuronectes platessa) in the eastern English
Channel, an important area for marine resources and fisheries
(Carpentier et al., 2005). European plaice are bentho-demersal
fish that around the British Isles spawn and recruit mainly (75–
80%) in the English Channel (Nash et al., 2000). They are
exploited by several countries (France, Belgium, and the UK)
and represent the most abundant benthic fish species landed
from the eastern Channel (Mahe
´
et al., 2006). However, in the
past 20 years, plaice landings have more than halved (from
8366 t in 1978 to 3146 t in 2006; ICES, 2008).
In this context, it is necessary to improve our knowledge of the
spatial distribution of the species at various life stages to provide
information for effective management. Previous studies have
tried to relate the distribution of flatfish to single environmental
factors, e.g. depth, temperature, and sediment type (Smale et al.,
1993; Albert et al., 1998; Ellis et al., 2000; Amezcua and Nash,
2001), but the general knowledge of how these species use and
change their habitat during different life stages using a complex
approach (testing several environmental predictors together)
remains poor. Younger fish, because of their smaller body size,
may be more limited in their swimming and competitive abilities
and may be confined to smaller or different areas from older,
bigger fish. Moreover, the environment of the eastern English
Channel is subject to seasonal variation (Carpentier et al., 2009),
notably in terms of temperature and salinity spatial patterns.
Models were therefore developed for two seasons (summer and
autumn) and two life stages to study possible seasonal and onto-
genetic changes in habitat utilization.
Material and methods
Survey data
The Centre for Environment, Fisheries and Aquaculture Science
(Cefas) conducted a Beam Trawl Survey (BTS) annually in July/
August between 1989 and 2006. Samples were collected using a
commercial 4-m beam trawl at fixed stations, following a depth-
stratified sampling strategy. The standard annual survey had a
total of 100 stations (Figure 1a). The gear had a minimum
mesh size of 40 mm and was towed at 4 knots (ground speed)
for 30 min at a warp length adjusted to bottom depth. Water
column depth, temperature, and salinity were recorded using
sensors attached to the beam trawl; temperature/salinity data
were not available for all stations and years.
The Channel Ground Fish Survey (CGFS), a survey by the
Institut Franc¸ais de Recherche pour l’Exploration de la Mer
(Ifremer), was carried out each year in October between 1988
and 2006 on board the research vessel “Gwen Drez”. The survey
extended from the eastern English Channel to the southern
North Sea (Figure 1b). The study area was subdivided into rec-
tangles of 15
′
latitude and 15
′
longitude using a systematic
sampling strategy. The sampling gear was a very high (3 m) vertical
opening bottom trawl (also known as a GOV trawl) with a
minimum 10-mm mesh in the codend. Trawls lasted 30 min at
an average speed of 3.5 knots. This sampling gear is well adapted
to the various seabed types encountered in the study area.
Between 1997 and 2006, temperature and salinity (surface and
bottom) were also measured, using a sensor attached to the head-
line of the trawl.
For both surveys, the fishing method was standardized:
sampling stations each year were at similar locations, and identical
sampling gear was used. After each haul, all fish were sorted, ident-
ified, counted, and their length measured.
Abundance indices at each station were standardized to
density (number of fish per km
2
), which was tested for normality
(using histograms, skewness, and kurtosis) and log-transformed
[log
10
(x + 1), where x is density, to reduce the skewness of the
distribution (Legendre and Legendre, 1998)]. Such a transform-
ation is widespread in numerical ecology (Legendre and
Legendre, 1998) and has a fairly similar effect on the data to a
direct log(x) transformation when data values are .0.1, as is
the case here. Length at age 1 was estimated using the von
Bertalanffy growth function (von Bertalanffy, 1938), then used
to separate data for 0-group (, 18.0 cm) and 1+ group
(≥18.1 cm) fish. Both surveys were designed to target different
components of the fish fauna and did not catch all species
with equal efficiency. Nevertheless, for plaice, both life stages
were well represented in the catches (Figure 2).
Environmental predictors
For habitat modelling, sea surface temperature and salinity, water
column depth (all three measured during the surveys described
above), bed-shear stress, and seabed sediment type (Larsonneur
et al., 1979) were used as predictors (see a summary of the obser-
vations in Supplementary Table S1). Five main categories of
seabed deposit were used: pebble, gravel, coarse sand, fine sand,
Figure 1. Stations sampled in 2006 during (a) the BTS, and (b) the
CGFS surveys (fixed station sampling schemes).
Using habitat modelling and GIS to predict habitat utilization
1501
at University of Plymouth on July 26, 2011icesjms.oxfordjournals.orgDownloaded from
剩余11页未读，继续阅读
都行。
- 粉丝: 0
- 资源: 1
我的内容管理
展开
最新资源
- WebLogic集群配置与管理实战指南
- AIX5.3上安装Weblogic 9.2详细步骤
- 面向对象编程模拟试题详解与解析
- Flex+FMS2.0中文教程：开发流媒体应用的实践指南
- PID调节深入解析：从入门到精通
- 数字水印技术：保护版权的新防线
- 8位数码管显示24小时制数字电子钟程序设计
- Mhdd免费版详细使用教程：硬盘检测与坏道屏蔽
- 操作系统期末复习指南：进程、线程与系统调用详解
- Cognos8性能优化指南：软件参数与报表设计调优
- Cognos8开发入门：从Transformer到ReportStudio
- Cisco 6509交换机配置全面指南
- C#入门：XML基础教程与实例解析
- Matlab振动分析详解：从单自由度到6自由度模型
- Eclipse JDT中的ASTParser详解与核心类介绍
- Java程序员必备资源网站大全
资源上传下载、课程学习等过程中有任何疑问或建议，欢迎提出宝贵意见哦~我们会及时处理！
点击此处反馈
