Skip to main content

    Recruitment variability along the Antarctic Peninsula: What's the best way forward

    Request Meeting Document
    Document Number:
    C.S. Reiss and G.M. Watters
    Submitted By:
    Dr Christian Reiss (United States of America)
    Approved By:
    Dr George Watters (United States of America)

    This document is intended to address several issues related to the development of recruitment indices for Antarctic krill from net-tow data. We address three primary issues: 1) the view that length-frequency data collected by long-term monitoring programs that have used nets other than the RMT-8 cannot be used to parameterize Grym simulations, 2) that proportional recruitment of age 2 krill is somehow the only recruitment index that should be used to parameterize the Grym, and 3) that a reliable parameterization of recruitment can be generated from only one or a few surveys. We conclude from multiple lines of evidence, most developed since the GYM was first parameterized, that long-term monitoring of krill populations along the northern and western Antarctic Peninsula cannot summarily be ignored on the basis that length-frequency data were not collected with an RMT-8. We also conclude that 1) temporal variation in recruitment as inferred from data collected by these long-term monitoring programs is realistic and not too high; 2) length-frequency data from these programs do represent the population with sufficient fidelity; and 3) the Scientific Committee and its working groups should parameterize the Grym with information based on time-series data from the U.S. AMLR Program, Palmer LTER, German surveys, etc. In our view, scenarios 1-6 and 19-24 from Maschette et al. (2021), which do not parameterize recruitment using data from the aforementioned long-term monitoring programs and even include data from East Antarctica, are not appropriate for developing a management strategy for Antarctic krill in Subarea 48.1.

    Here we provide an overview of work previously conducted by the SC’s working groups and demonstrate that some possible recruitment parameterizations suggested for use in some Grym scenarios violate basic assumptions originally used to index recruitment. These violations include an absence of volume calculations to estimate krill length density, net meshes that are too large to effectively represent the range of lengths present in the krill population, and a number of recent surveys that have collected too few samples to confidently represent the populations’ length-frequency distribution. Moreover, given the temporal pattern in recruitment success, that is now understood to relate to recruitment failure rather than availability to or catchability by nets, recruitment parameterizations based on just one or two surveys conducted in immediate succession should be of less value.