Does size matter?

[Euan Harvey]

I would like to start a discussion about whether size or length information matters and how accurate the measurements need to be for fisheries, marine conservation and general marine ecology.   I would be interested in hearing peoples thoughts.

Euan

[Kim Jerome Friedman]

From a long term MER perspective (not longitudinal research), I would suggest we need sufficient accuracy to be able to determine length frequency differences of fish under different forms of management controls – especially discernment of juvenile from adult fish, and larger adults from other adult size classes in the capture size range. This will differ by species, but suggests that in many cases, a length measurement accuracy as low as 3 cm might be sufficient for fish that grow to larger sizes. I realise as scientists, and as researchers, we are interested in carrying out some level of cohort analysis so that this level of accuracy might not seem rigorous, but I would rather have more measures of reasonable accuracy that fewer where accuracy was pin-point.

[Tom Holmes]

Many of the questions we ask with regard to length/size/biomass in fisheries and marine conservation are generally directed towards either larger bodied or target fish groups. My thoughts here is that an accuracy to the nearest ~3-5cm is sufficient for these larger bodied species. Personally I don't see the point in acquiring extremely precise length estimates as I don't think it really makes that much of a difference to the results or outcomes. My focus would be more on getting MORE estimates, rather than extremely precise ones.

[Steve Newman]

If you want to use length based data in any meaningful assessment in regard to sustainability then both accuracy and precision are required when it comes to length based assessments.  For example, the length frequency data used in many fishery assessments is very accurate (i.e. nearest mm), therefore when assessing the sustainability of a fishery, using length based indicators such as Pmat: the fraction of the catch that is above the length at maturity (Lm); Popt: the fraction of the catch that is within ±10% of the optimal length of harvest (Lopt); and Pmega: the fraction of fish that are more than 10% larger than Lopt, then you require both accuracy and precision over time. You also need large sample sizes. Without adequate precision you have little effective power to detect change. When it comes to BRUV and DOV based data, the same principles apply, as it can be very useful in an EBFM context to assess bycatch and discard sustainability. To do this you will have little information except length based metrics. Again in order to detect change, you need both accuracy and precision. I would think that as soon as you diverge to bin lengths of +/- 3-5mm, then it will be extremely difficult to have the power to detect change. Indeed it would make a nice simulation study. 

[Tim Langlois]

Converting length measures to biomass estimates is often done (or requested by reviewers) in studies of the effects of fishing comparing no-take and fished areas.

I have found that inaccuracies in a few length measures can lead to large differences in biomass that can have a large effect on patterns - particularly for the larger fish. This can also be due to length-weight conversions being non-optimal, which could be due to regional variation in growth rates.

This is one of the reasons I am keen on pursing more length based indicators like Steve suggests, as you are not introducing additional error from the length-weight conversions.

Length and/or biomass information is essential for trophic questions (such as in EBFM) and I think it is therefore really important for studies to try and obtain the most accurate length estimates you can across the size distribution of the assemblage - from the little one to the big ones. So that we can represent the patterns accurately - but like Tom said above - precision between studies could be more important than - exact measurement! But I would say that if you are designing a stereo sampling program to give really good precision between studies you will end up with great accuracy as well.