Present: Davis (Rapporteur), Beardsley, Runge, Lough, Buckley, Ji
Discussion of 6 themes:
Our working group had a general discussion, with a large amount of overlap between themes.
We first discussed the role of top down versus bottom up control of the ecosystem and target species. Two independent GLOBEC studies, Pershing et al., and Steele et al. have found strong support for the dominance of bottom-up control of the GB/GOM ecosystem. Pershing et al. compared the 1980s and 1990s, the so called MARMAP and GLOBEC decades, respectively, and found that the mean surface salinity of the GOM was lower during the 1990s than during the 1980s. This lower salinity was associated with higher fall phytoplankton, high abundance of small copepods during spring, higher herring stocks, and higher recruitment of haddock. Steele et al. used a combination of historical data and modeling to compare decadal scale changes in the Georges Bank ecosystem over the past 40 years. They concluded that, during the last three decades fish food requirements balanced lower trophic level production, whereas, during the 1960s, intrusions of low-nutrient Labrador Slope Water reduced primary and secondary production and overall fish yields. The group felt that both bottom up and top down processes operate on the system as a whole and different species may be affected in different ways. Bottom up effects do appear to strongly affect variability at the system level.
In considering the effects of remote (climate/basin) forcing on the GB/GOM system, it was felt that, in keeping with the approach of the GLOBEC NWA, we need to understand the interior dynamics of the GB/GOM system in order to understand how the system will respond to external forcing. We felt that the proposed process oriented modeling is the right approach, i.e., conducting targeted numerical experiments to examine different forcing scenarios. Setting up boundary conditions for warm and cold regimes will be examined with respect to their influence on lower food web and copepod species production.
By modeling the GLOBEC years, in which we have a rich data set, we will be able to examine the factors leading to "good" and "bad" years in terms of plankton production. The good years, like 1998 and 1999 provide necessary food conditions for outstanding year classes but are not a guarantee, e.g., haddock recruitment was high in 1998 but not in 1999. (Other factors like post-pelagic juvenile survival and cannibalism, not studied in GLOBEC, may contribute to recruitment success.)
Such production estimates can be used as indices for ecosystem based fisheries management. The models we are developing provide abundance estimates for dominant copepod species (Calanus, Pseudocalanus, Oithona) which can be reduced to simple estimates, such as low, medium, high levels of large and small copepods. These indices can be compared in hindcast with good and bad year classes of cod and haddock.
Other indicators include the biological state variables in the model, including nutrients and phytoplankton. The model variables can be combined into multivariate indices or principal components. Model variables to consider include wind, temperature, N-P-Z levels, copepod species abundances, larval fish growth and survival.
From GLOBEC data, Buckley and Lough have found that 50% of larval growth and mortality can be accounted for by seasonal effects i.e. photoperiod and another 25% is due to a year effect. Thus we can account for a significant proportion of the total variance in larval growth and survival. The cause of the year effect is not yet known, but will be examined using the modeling studies. The remaining percentages are critically important to determine since a small change in mortality rate can have such a large impact on recruitment.
A major finding of the GLOBEC program was that larval cod and haddock are food limited in nature, even in the food rich region of Georges Bank. Using the extensive GLOBEC data set, Buckley and Lough found a strong relationship between abundance of Pseudocalanus and larval fish growth and survival. This finding clearly indicates that bottom up effects are critical for these species.
Thus using the model, we can identify the mechanisms leading to good and bad years of Pseudocalanus which will serve as a proxy index for larval fish growth and survival. Since survival, as well as growth, is affected by copepod abundance, this index may be very powerful.
Historical data analysis indicates that remote forcing has a strong effect on bottom up processes through stratification, nutrient input, and wind driven flows. We will use the models to conduct numerical experiments to gain insight into these mechanisms. The modeling approach is needed to examine year to year effects of winds, warming, and stratification.
One issue raised is that Pershing et al., found a relationship between low salinity and system production, and in another study found a relationship between NAO and Calanus from the CPR GOM transects. Although there is a relationship between NAO and Labrador Slope Water intrusions in the GOM, intrusions of low salinity water from the Scotian Shelf do not appear to be related to NAO (Mountain pers. comm.). Thus, if the enhanced system production in the GOM is in fact caused by low salinity intrusions from the Scotian Shelf, this enhanced production must not be related to NAO, but may be due to a longer term effect, probably polar ice melt. The relation of CPR-derived Calanus abundance to NAO then may be due to intrusions of Labrador Slope Water into the GOM and not to enhanced productivity from Scotian Shelf intrusions.
Numerical experiments need to be done to compare the influence of salinity versus other aspects of remote forcing (global and basin), including wind forcing and surface heating. NAO effects do not appear to impact local meteorological conditions in the GB/GOM region (Beardsley cited Joyce study), but the influence of longer term global trends in warming and wind forcing on the GB/GOM ecosystem need to be examined.
In terms of transitioning the GLOBEC findings and models for use by managers, it was felt that it is not the role of GLOBEC scientists to run operational models, but it certainly is our responsibility to provide models that can be transitioned to operational groups such as NOS. The models we are developing will be tested in hindcast mode and can be run in nowcast and in the end, in forecast mode. The models will provide estimates of lower food-web and copepod species production as well as recruitment of fish through the larval stage. In this way, the models will provide indices of system productivity and good and bad years for larval fish survival as a function of local and remote forcing. The models are not expected to provide high-frequency (e.g., daily) forecasts, but are likely to provide indicators of value for 1-10 years into the future. The models are being calibrated on the GLOBEC years (1995-1999) and extended to present time.
The use of our model-generated ecosystem indices can be used together with current spawning stock biomass estimates to help fisheries managers make better predictions.
These models also will serve as useful tools for determining what we need to measure in the future. In the case of observing systems, the models can be used for siting nodes and for determining what sensors are needed at each node.