The CCAMLR krill synoptic survey

Background papers

  1. Paragraphs from WG-EMM-98

Paragraphs from WG-EMM-98 in Cochin

Plans for a Synoptic Krill Survey in Area 48

9.49 At previous meetings of the Working Group and the Scientific Committee a number of documents and recommendations specifically directed at planning for a synoptic survey in Area 48 had been produced.  All of these discussions and recommendations were summarised in WG-EMM-98/24.
9.50 In addition, members of the synoptic survey steering committee and task groups who were present at the Workshop on Area 48 (La Jolla, USA, June 1998) met briefly to discuss progress of plans for the survey.  A report of these deliberations and tasks carried out immediately following that meeting is presented in WG-EMM-98/25.
9.51  The primary objective of the synoptic survey is to improve estimates of B0 (pre exploitation biomass) used in the krill yield model to estimate sustainable yield for Area 48 (SC-CAMLR-XII, paragraphs 2.39 and 2.41 to 2.47).
9.52 The Working Group considered the implications of interpreting the results from a single large-scale synoptic survey on estimating a long-term annual yield.  The Working Group agreed that the krill yield model was robust with respect to interannual variability in krill biomass and will be able to use the B0 estimate directly in its calculations.
9.53  The Working Group noted that the smaller-scale regional surveys could be used to monitor for long-term trends in krill biomass.  However, the relationship between these surveys and the biomass throughout Area 48 needs to be determined.  In this respect it would be advantageous to ensure that the regular regional surveys (such as the US AMLR survey (USA) and the BAS Core Program (UK)) could be linked to the large-scale synoptic survey in time and space so that the temporal variations observed in the regional surveys may be interpreted in respect of the larger area.

Survey Design

9.54 A number of documents specifically addressing the subject of survey design have been submitted to WG-Krill and WG-EMM over a number of years (summarised in WG EMM 98/24) and a number of general survey designs had been proposed.
9.55 At the Workshop on Area 48 , survey design was discussed in general terms and it was agreed that a randomised design coupled to a design-based analysis would produce the most statistically defensible result (WG-EMM-98/25, Appendix 1; see also conclusions from WG Krill-94/20 which is presented as Appendix 10 in WG-EMM-98/24).
9.56 The Working Group considered the set of four draft survey plans (WG EMM 98/44 and 98/53) which were drawn up at the request of the Area 48 synoptic survey planning meeting.  All plans considered alternative ways of interleaving the tracklines of three ships that each had 30 days of time available to carry out the entire survey (including associated logistic costs).
 (i) Plan 1 comprised a stratified random design with four strata. Three separate strata were placed around the South Shetland Islands, South Orkney Islands and South Georgia.  The fourth strata covered the remaining oceanic regions of Area 48.
 (ii) Plan 2 was an adaptive survey design utilising the same large-scale grid as plan 1 but directing additional survey effort into a series of survey cells (2  latitude x 2  longitude) that were shown to have above-average levels of biomass from the first pass through the area.
 (iii) Plan 3 comprised a random transect design where all the effort for the three ships was put into conducting a series of large-scale transects across the area.
 (iv) Plan 4 was an adaptive survey design where a series of small scale surveys were carried out each time a ship passed through an area where there was a high biomass of krill.
9.57 The Working Group first considered the relative merits of an adaptive survey design against a pre-planned survey design (plans 2 and 4 versus plans 1 and 3).  It was agreed that an adaptive survey design could offer increased understanding of the structure of the system, through a more detailed description of the distribution of krill within high density areas.  However, the Working Group felt that the advantages of an adaptive approach,
particularly as outlined in plan 4, in terms of improving the CV of the biomass estimate were less obvious.  Indeed such techniques appeared to introduce increased complexity in terms of the design, execution and subsequent analysis of the survey.
9.58 The Working Group agreed that a model simulation of the relative merits of the adaptive and pre-planned surveys would be required to quantify the advantages of the two approaches.  However, considerable concern was expressed both over the time-scale of such a simulation and the conclusiveness of the outcome.
9.59 The Working Group was also in agreement that a decision on the overall survey design had to be taken at this meeting.  The Working Group agreed therefore that, given the concerns expressed above, the more conservative approach of utilising a pre-planned survey should be the preferred approach.  This approach had been widely used in the past (for instance FIBEX) and was statistically robust and defensible.
9.60 The Working Group then considered the relative merits of a stratified versus unstratified design (plan 1 versus plan 3).  It was pointed out that if krill were distributed in similar quantities both in the open ocean and in the shelf areas then a design which gives a uniform density of sampling across the whole area should be used (plan 3).  However, if krill are concentrated in particular predictable areas, then a stratified sample design which takes account of this is likely to produce a lower overall CV.  Note, however, that such a design will not change the expected estimate of mean biomass.
9.61 The Working Group was unable to agree on the relative importance of krill occurring on the shelves around the coast of the Antarctic Peninsula and the islands in Area 48. A variety of datasets and published papers (cf. WG-EMM-98/18 and 98/32) illustrate the complexity of the system.
9.62 The Working Group finally agreed that a modification of plan 3 would be adopted. Such a modification seeks to maximise the coverage provided by a series of large scale transects carried out across the Scotia Sea by the three ships undertaking the survey. However, in order to reduce the CV of the biomass estimate, within three regions (north coast of South Georgia, north coast of the South Orkneys and offshore from the South Shetlands) there will be an additional transect between each of the large-scale transects (see Figure 2), in effect doubling the transect density in the three regions described above relative to the rest of the survey area.  For analytical purposes, this allocation of survey effort would provide data from two distinct strata (one being more densely sampled than the other).
9.63 To supplement the core survey as described above the Working Group also agreed that:
 (i) if a fourth or fifth ship was able to contribute time within the period January 2000 then a series of extra transects would be interleaved within the existing large-scale transects;
 (ii) the length of the transects would be tuned to the latitudinal boundaries of the krill distribution; and  (iii) participants would be encouraged to carry out their standard regional surveys either prior to or after the main synoptic survey.  This was extremely important because it is necessary to link the temporal sequence of regional surveys with the wide spatial coverage of the synoptic survey.

Methods

   Acoustics
9.64 The general philosophy for the acoustic methodology and equipment had been discussed previously.  All participants were using Simrad EK500 systems and it was therefore possible to ensure a high level of standardisation.  A preliminary proposal suggesting standard sampling protocols had been submitted (WG-EMM-98/47).  The Working Group agreed with the general philosophy of the paper; that is whenever possible exact equipment, software and settings should be dictated.  When exact matches were not possible, pertinent comparative information should be specified.
9.65 These protocols which cover (i) instrument settings, (ii) data logging, (iii) system calibration including multifrequency TS calibrations, intership calibrations and characterisation of system noise, (iv) survey operations and (v) additional recommendations should be agreed by the acousticians of the participating nations.
 9.66 Multifrequency acoustics (38, 120, 200 kHz frequencies) would be available on the three ships carrying out the core transects.  However, if additional ships were able to contribute to the survey effort (see paragraph 9.63) then such data would be a valuable addition to the dataset even if only 38 and 120 kHz data could be provided.
   Net Sampling
9.67 The Working Group agreed that the main priority for the net sampling program was the determination of krill population structure demography.  The requirements for estimates of net density and target identification were of secondary importance given that much information on krill identification would come from multifrequency acoustics.
9.68 There was considerable discussion of the sampling strategy used in the draft plans submitted in papers WG-EMM-98/44 and 98/53.  In this case, a single night-time period was allocated for both net sampling and oceanographic sampling.
9.69 Given that the priority was to obtain a good coverage of samples for estimating the population structure, the Working Group agreed that a net sample should be taken around midnight and midday on each day of the survey. The timing of the midnight sample was constrained by the period of darkness. However, the timing of the midday sample was more flexible and the Working Group agreed that consideration should be given to allowing the time of this sample to shift as necessary to maintain a more regular pattern of sampling stations.
9.70 The Working Group considered that given the theoretical variation in catchability and selectivity of nets (see WG-EMM-98/20) it was highly desirable to standardise on the type of net used for krill sampling.  At present the following nets were available:
Country Net Mesh Size (mm)  Mouth Area (m2)
Japan  KYMT 3.4 9.0
UK RMT8  4.0  8.0
USA IKMT 0.5 2.5
Russia IKMT 6.0 6.0
Korea IKMT 0.5 4.5
9.71 The Working Group felt that the most appropriate net in terms of catchability was the RMT8.  However, they recognised also the financial implications of standardising on a single design and size of net.
9.72 Therefore the Working Group made two recommendations to achieve the best possible compromise with regard to net standardisation:
 (i)  to investigate the possibility of obtaining RMT8 systems from other CCAMLR Members which were not able to contribute ship time to the survey or any other sources; and
 (ii)  if the above was not possible then to allow the use of any of the above nets which had a mouth area of between 8 and 10 m2 and a mesh size of between 3.0 and 4.0 mm.
9.73 The Working Group agreed that representative samples of krill would be measured on board ship and results would be entered onto computer prior to the end of each cruise.  Detailed protocols must be established as soon as possible.
 Environmental Sampling
9.74 The Working Group considered that each ship should undertake a CTD to a depth of 1 000 m at each midnight and midday station.  A detailed protocol including the use of water bottle samples for CTD calibration must be produced as soon as possible and the time implications assessed.
 Other Sampling
9.75 The Working Group agreed that the acoustic transects, net sampling for population structure and CTD conducted at the midnight and midday stations would form the core of the sampling program and must be undertaken to standard protocols by all ships participating in the survey.
9.76 There were likely to be a number of other variables that participating countries would normally collect (for instance phytoplankton fluorescence, macro-zooplankton species composition, underway data such as obtained by pumped seawater supply or undulator).  Such measurements were encouraged by the Working Group.  It was emphasised, however, that such measurements must not compromise the collection of the core datasets.
9.77 The Working Group felt that it was appropriate that protocols and arrangements could be discussed between countries intending to collect similar data through the respective cruise leaders.
9.78 Appendix 8 of WG-EMM-98/25 presented a proposal from the IWC to send whale observers to participate on synoptic survey ships.  Plans for such participation would be discussed at an IWC workshop at St Andrews, Scotland in March 1999.  The Working Group felt that such a proposal was a valuable contribution to the synoptic survey.  However, considerable concern was expressed about the number of observers traditionally used on such sighting surveys (six to eight persons per ship).
9.79 Although it was not possible to detail exactly how many spaces would be available on each participating ship it was thought that Japan and Russia would most likely only have room for one observer on each ship, UK was also tightly constrained but might be able to offer two berths while the USA thought it likely that they would be able to accommodate four to six observers.
9.80 The UK felt that independent estimates of whale numbers and incidental bird numbers would make a valuable comparison with the continuous bird and at-sea mammal observations that they conduct on their standard core program cruises and which are likely to be a component part of the synoptic survey.

Data Analysis and Storage

9.81 The Working Group agreed that the core datasets (i.e. acoustic data, krill length frequency data and CTD data) should be analysed in the first instance within a CCAMLR workshop by all survey participants.  Such a workshop should be timed to take place as soon after the cruise as practical and in any event prior to the 2000 meeting of WG-EMM.
9.82 Further the Working Group agreed that the initial dissemination and publication of these core data results should take place as a joint undertaking.
9.83 A copy of all core data and appropriate meta-data must be deposited with the CCAMLR data centre.  The Working Group agreed that the appropriate data storage formats would need to be addressed prior to the cruise.

Coordination of Planning after Meeting of WG-EMM-98

9.84 The Working Group agreed that there was an urgent requirement to appoint a survey coordinator who would undertake the following tasks:
 (i)  serve as a focal point between CCAMLR and cruise participants, and among cruise participants, for all matters relating to the survey planning, conduct and analysis;
 (ii)  organise a planning workshop.  This would include defining all tasks to be accomplished at the workshop, coordinating the preparation of cruise protocols and ensuring such cruise methodologies are uniformly applied;
 (iii)  coordinate cruise plans and preparations amongst participants prior to the beginning of the survey.  This would include coordinating the participation and interchange of key experts;
 (iv)  serve as the at-sea coordinator;
 (v) ensure that data are supplied to CCAMLR and participants;
 (vi) organise data analysis workshop for survey participants; and
 (vii) coordinate report generation.
9.85 The Working Group agreed that the planning workshop should take place in mid- to late March 1999.
9.86 The Working Group also agreed that each participating country should provide a summary of any additional activities relevant to the synoptic survey that they would be undertaking during the 1999/2000 season.
9.87 Such summaries, all protocols and preliminary cruise plans must be completed and circulated to participants one month prior to this planning meeting.
9.88 The Working Group asked Dr Watkins to act as coordinator.  In addition Drs Hewitt, M. Naganobu (Japan) and Watkins were appointed principal contacts for participating nations.
9.89  Any other countries wishing to participate should give a firm commitment prior to the planning meeting and supply the coordinator with a principal point of contact as soon as possible, but in any event no later than 15 March 1999.
9.90 Finally, the Working Group agreed that to facilitate planning it was important to nominate a definite start date as soon as possible.  As a matter of priority all participants should confirm their ability to arrive at South Georgia in the first week of January 2000 to start the first calibration.
 
Introduction Itinerary Station positions Cruise tracks Planning Meeting Sampling Protocols Participants Background papers Contents

Page updated 8 December 1998