ICES/IOC Workshop onBiological Effects of Contaminants inPelagic Ecosystems |
|
As described in the 2001 Report of the Steering Committee (ICES CM 2001/E:01), workshop participants are requested to report their data to the BECPELAG Scientific Committee for the purpose of a joint analysis of several methods simultaneously. The scope of these analyses is outlined below (see section 5). Workshop participants will receive the basic oceanographic and chemical data which will be provided by the workshop organization (only) after having reported their data to the Scientific Committee.
The data submitted by workshop participants will be not be distributed to third parties, neither inside nor outside the workshop.
Results from the joint analysis of the workshop data will be published in the Workshop Proceedings, with proper reference to the parties involved. Also, results from the joint analysis will be presented during a Theme Session at the ICES Scientific Conference 2002, and at the Result Presentation of the workshop. Beyond that, each workshop participant is expected to describe his/her method and findings in an individual contribution to the Workshop Proceedings.
As the various methods used in the workshop
generate very different types of results, it is hardly possible to provide a
unique template file structure suitable for all methods. The basic rule is that
all information relevant for statistical analysis must be reported. However, a
typical basic record in a report file refers to
In cases of doubt concerning what or how to
report, please contact the Scientific Committee (see section 6).
Aggregated or derived quantities like means,
standard deviations, percentages, ECxx etc. can only be accepted as additional
information for plausibility checks. They cannot replace the basic records
described above.
Each basic record must contain a unique
identification and a description of what the record refers to. This requires
among others the following information
Information |
Example |
project identification |
BE 015 |
date (containing year, month, day) |
270301 for 27Mar01 |
sampling location (in geographic coordinates,
ie latitude and longitude down to seconds) |
60° 35.75’ N for 60° 35’ 45” N (three separate fields) |
station identification (to be given by the
cruise leader) |
XY35 |
sample number (one sample comprises all measurements
referring to one location, one time point) |
same number for all cod taken at 27Mar01 at
60° 35.75’ N, 02° 15.07’ E |
replicate number (identifies sampled object) (could also be pool number) |
5 (for cod number 5) |
subreplicate number (if replicate was split up to determine
analytical variance |
2 (for liver part no.2 from cod no. 5) |
number in pool |
1 (for non-pooled measurement) |
matrix description (may contain several components. If so, use
one field in the record for each component) |
cod (field 1) liver (field 2) |
description of measured quantity |
Cd |
measurement technique |
AAS |
detection/determination limits |
|
unit |
µg/kg |
reference |
liver fat |
additional information |
measurement technique, fish size, weight, age, condition index… |
measurement (the result of the method) |
3.47 |
Quality control procedures |
present? which? |
Data describing concentration-response
information will at least need an additional field for the concentrations used.
Accepted technical formats are
For the use of other formats please contact
beforehand the Scientific Committee (see section 6).
Files must be organized with horizontal rows
referring to records/measurements and vertical columns referring to fields
containing all entries for one parameter/quantity. All files must be
accompanied by a description of their contents. A descriptive head as top of
each column may serve for this purpose.
Files should be sent as soon as possible.
Send your files as e-mail attachments or on
disk to
You will get a statement of receipt
(immediately). Your data will be transformed into a standard form for
statistical analysis, and a safety copy will be made (see the statement of
confidentiality in section 2). You will get a copy of the transformed data,
with the request to check the transformed structure, just to make sure your
data description has been correctly understood. Together with the copy of your
transformed data you will receive the basic oceanographic and chemical data
that is supplied by BECPELAG to the participants.
The aim of the intended comprehensive analysis
is
For the contaminant concentrations, the
information as measured during the workshop will be used as well as results
from the DREAM model (for Statfjord) and backward trajectory calculations (for
the German Bight) provided by the BSH, Hamburg.
A variety of statistical methods will be needed
to pursue these aims. Emphasis will be on multivariate and nonparametric
approaches which are necessary to express the results of the methods on a
common scale required for a simultaneous consideration. It should be noted that
the outcomes of the various methods are of different type (yes/no information,
rank orders, continuous measurements, response curves) so that the raw results
are not a priori suitable for a joint analysis. However, statistical methods
exist which allow to overcome this problem to a certain extent. If the structure
of the final database allows so, also methods from what is sometimes denoted as
"artificial intelligence" will be used to uncover relationships
(regression/classification trees, neural networks, self organizing maps).
Contact person for questions concerning data
reporting and statistical analysis is
Werner Wosniok
Universität Bremen
Institute of Statistics, FB 3
PO Box 330440
D 28334 Bremen
Germany
phone +49 421 218
3471
fax +49 421 218 4020
email wwosniok@math.uni-bremen.de
Last update: 07Feb02 by WW