19 Recruitment and compensation

Compensatory mechanisms

Sustaining fisheries yield when fishing reduces stock size depends on the existence of compensatory improvements in per capita recruitment, growth, and/or natural mortality rates. Ecosim allows users to represent a variety of specific hypotheses about compensatory mechanisms. Broadly, these mechanisms fall in two categories:

• direct – changes caused over short time scales (order one year) by changes in behaviour of organisms, whether or not there is an ecosystem-scale change due to fishing; and
• indirect – changes over longer time scales due to ecosystem-scale responses such as increased prey densities and/or reduced predator densities. Usually we find the direct effects to be most important in explaining historical response data. In the next three sections we describe how to generate alternative models or hypotheses about direct compensatory responses; these hypotheses fall in three obvious categories: recruitment, growth and natural mortality.

Using Ecosim to study compensation in recruitment relationships

The multi-stanza representation of juvenile and adult biomasses was originally included in Ecosim to allow representation of trophic ontogeny (big differences in diet between juveniles and adults). To implement this representation, we found that it was necessary to include population numbers and age structure, at least for juveniles, so as to prevent “impossible” dynamics such as elimination of juvenile biomass by competition/predation or fishing without attendant impact on adult abundance (graduation from juvenile to adult pools cannot be well represented just as a biomass “flow”).

When we elected to include age-structured dynamics, we in effect created a requirement to think carefully about the dynamics of compensatory processes that have traditionally been studied in terms of the “stock-recruitment” concept and relationships. To credibly describe the dynamics of multi-stanza populations, Ecosim parameters for multi-stanza juvenile stages usually need to be set so as to produce an ’emergent” stock- recruitment relationship that is at least qualitatively similar to the many, many relationships for which we now have empirical data (see the RAM Legacy Stock Assessment Database). In most cases, these relationships are “flat” over a wide range of spawning stock size, implying there must generally be strong compensatory increase in juvenile survival rate as spawning stock declines (otherwise less eggs would mean less recruits on average, no matter how variable the survival rate might be).

When creating multi-stanza dynamics, be careful in setting model parameters that define/create compensatory effects. This begins with the Ecopath input parameters; in order for the juvenile dynamics to display compensatory mortality changes, at least two conditions are needed or helpful:

• the juvenile group(s) must have relatively high P/B, i.e. high total mortality rate (see Multi-stanza life history chapter);
• the juvenile group(s) must have either relatively high EE (so that most mortality is accounted for as predation effects within the model) or else the user must specify a high (near 1.0) value in the Ecosim > Input > Group info form entry for the juvenile group’s Proportion of other mortality sensitive to changes in feeding time column.

Compensatory effects can be increased (the recruitment relationship is flat over a wider range of adult stock sizes, with a steeper slope of recruitment curve near the origin) by:

• Limiting the availability of prey to juveniles (forcing juveniles to use small foraging arenas for feeding) by setting all elements of the Ecosim vulnerability multiplier form column for the juveniles to a low value (1.1-2.0); or
• Setting a higher value for the juvenile group’s Feeding time adjustment rate parameter  (Ecosim > Input > Group info form), which causes the effective time exposed to predation while feeding to drop directly with decreasing juvenile abundance (i.e., simulates the possibility that when juveniles are less abundant, remaining ones may be able to forage “safely” only in refuge sites without exposing themselves to predation risk). This option should preferably be used if you are fairly sure from field natural history observation that the juveniles do in fact restrict their distribution to safe habitats when at very low abundance.

It is especially important to test alternative values for the vulnerability of prey to juveniles. If the vulnerability multiplier is too high, the Ecosim emergent stock-recruitment relationship is likely to look almost like a straight line out of the origin, i.e. without compensatory effect. If the vulnerability multiplier is too low, the relationship may develop a “spurious” dome-shape.

In Ecosim multi-stanza groups, the group that is displayed on the Ecosim > Output > Stock recruitment plot (S/R) is always the oldest stanza. The stock-recruitment relationship between this stage and each of the younger stages separately is calculated and displayed on the S/R form.  This may cause issues when the oldest group is a “senescent” group as is often done for modelling Pacific salmon, (which die after spawning).

Compensatory growth

Compensatory growth rate responses are modelled by setting the feeding time adjustment rate (Ecosim > Input > Group info form) to zero, so that simulated Q/B is allowed to vary with the group biomass (non-zero feeding time adjustment results in simulated organisms trying to maintain Ecopath base Q/B by varying relative feeding time). Net production is assumed proportional (growth efficiency) to Q/B, whether or not this production is due to recruitment or growth. The Q/B increase with decreasing pool biomass is increased by decreasing vulnerability of prey to the pool (Ecosim > Input > Vulnerability multipliers form). In the extreme as vulnerability multipliers approaches unity (donor or bottom-up control indicative of a group being close to its carrying capacity), total food consumption rate Q approaches a constant (Ecopath base consumption), so Q/B becomes inversely proportional to B.

Compensatory natural mortality

Compensatory changes in natural mortality rate (M) can be simulated by combining two effects: non-zero Feeding time adjustment rate (set on Ecosim > Input > Group info form), and either high EE from Ecopath or high proportion of M0 due to predation (unexplained predation > 0). With these settings, especially when vulnerability multipliers of prey to a group are low, decreases in biomass lead to reduced feeding time, which leads to proportional reduction in natural mortality rate.