1. Viable egg production for Baltic cod and sprat
1.5 Estimate the reliability of spawning stock biomass as an unbiased index of
viable egg production estimates
Introduction
The present section focused on evaluations of indices and absolute measures of egg production derived from adult biology and stock assessment data, and the sensitivity of these indices to variability in input parameters. A comparison of the theoretical potential egg production with that estimated from extensive ichthyoplankton sampling enabled an independent test of the various indices of egg production, including simple estimates of spawning stock biomass for both cod and sprat.
For cod not only the potential egg production has been estimated, but also the viable egg production. This is defined as the number of surviving eggs after consideration of mortality due to low oxygen concentration dependent on egg size coupled to female size. Our calculations are based on experimentally investigated mechanisms relating female size - egg size - egg buoyancy. Coupling historic size distributions with salinity conditions enabled to model the vertical distribution of eggs. Egg survival is then calculated from a relationship between relative egg survival and ambient oxygen concentration derived within the CORE and validated in the present project.
Further steps in the analysis were: i) validation of potential egg production as a measure of Baltic cod reproductive potential and sensitivity to constant input parameters ii) improving spawning stock biomass (SSB) as a measure of sprat egg production.
Results
Time series of potential egg production in cod were calculated using project derived information on maturity, sex ratios, and fecundity in combination with numbers- and size-at-age in the adult stock. The egg production estimates were calculated assuming different levels of biological realism: constant vs.
interannually-varying maturity ogives and sex ratios, knife-edge ogives vs. age-dependent maturity ogives and constant vs. interannually varying relative fecundity. These estimates of potential egg production were then compared with estimates of realized egg production as determined from historical and ongoing ichthyoplankton surveys in the Bornholm basin, i.e., the main spawning area for cod in the Baltic at the present time. The comparison was performed by linear regression of the potential and realized egg
This empirical comparison demonstrated that estimates of reproductive potential could be significantly improved by including knowledge about the reproductive status of the stock instead of assuming either constant or randomly varying biological inputs. This improvement could be achieved by applying routinely obtained data on sex ratios and female maturity to stock estimates. Fecundity as the most labour and cost intensive factor in the estimation procedure of the potential egg production could be predicted with substantial precision. These findings indicate that the estimated potential egg production of cod is an improvement of and a reasonable alternative to the spawning stock biomass in stock assessments.
We found a high correlation between the viable egg production of cod in the Bornholm Basin and the realized production of a late egg stage based on ichthyoplankton observations in the same area.
Furthermore, the results showed that the oxygen related survival of eggs from first time spawners is lower than for eggs from repeat spawners, especially under unfavourable hydrographic conditions.
A similar evaluation of the reproductive potential of sprat was conducted. In this case, spawning stock biomass explained a significant but modest level of variation in realized egg production as measured in ichthyoplankton surveys. The amount of variation in realized egg production explained by spawning stock biomass was higher for sprat than for cod. This suggests that spawning stock biomass is a better measure of realized egg production for sprat than for cod, and indicates that inter-annual variability in maturation processes, age-dependent sex ratios and individual fecundity are less important for sprat than for cod.
However, the variation in realized egg production explainable by spawning stock biomass could be significantly increased by allowing proxy variables to represent nutritional status and condition of the sprat population. These proxy variables were used because the time series for batch fecundity and spawning frequency in sprat were too short to allow construction of reliable time series. The proxy variables used were the pre-spawning growth anomaly and water temperature during the gonadal maturation period. Hence for sprat, environmental conditions related to growth during the gonadal maturation period have an important effect on egg production.
The results obtained in the present section indicate that estimates of potential and realized egg production do not always follow trends in spawning stock biomass. The decoupling between potential or realized egg production and spawning biomass is particularly critical when assuming constant or randomly varying input parameters in the calculation of spawning stock biomass, as had been done by the Baltic Fisheries Assessment working group until 1996 for cod and is still done for sprat. In these circumstances deviations between potential egg production and spawner biomass can be quite large. When interannual variations in relative fecundity, or reasonably configured proxy variables for relative fecundity, are accommodated in the egg production estimate, the decoupling between SSB and potential egg production is increased even more.
These findings are fully consistent with recent studies involving especially cod in the Atlantic. However, what is particularly striking about the analyses is the degree of improvement in fit between estimated egg production rates and independent field based estimates of egg production rate as measured in ichthyoplankton surveys. The correspondence between field and theoretical estimates improved significantly as more realistic biology and environmental information was added to the egg production models.
Furthermore, size dependent egg viability arrays were established for cod showing that oxygen related egg survival is coupled to female size and spawning experience, which implies that surviving egg production is related to female stock structure. Notably the additional biology (i.e. size distributions, sex ratio, maturity at size information and fecundity) or environmental information (e.g. temperature for sprat) can be obtained
relatively inexpensively as part of routine fisheries surveys or from calibrated proxy variables. These considerations suggest that substantial improvements in stock-recruitment models can be achieved by reducing measurement errors in the independent variable, and that biological reference points derived from such models can be estimated with greater certainty.