assessments.VPA) (Fig. 9.13a), and of (ii) fisheries catches off Namibia. The fitting
was to year-by-year biomass and catch estimates modified by a water-temperature
function (Fig. 9.13b). The key variables, of course, were the Ci, for which the
fisheries provided the data. Fits for catches are better than those for biomass,
primarily because there are better data. The complexity of such models may in fact be
sufficient to capture a substantial (if uncertain) portion of the significant interactions
among major stocks of nekton (i.e. predator–prey interactions), major habitat factors
(reasonably represented by temperature in an upwelling ecosystem), and fisheries. An
obvious problem, as Ecosim was applied by Heymans et al. (2009) is the absence of
prey-switching. The Ecosim program does allow for that, supposedly provided that
actual data show what the switches are. Surely snoek, for example, are not particularly
fussy about eating sardines when anchovy are unavailable.
Fig. 9.13 Comparisons of data and Ecosim model results for populations of the
northern Benguela upwelling ecosystem: (a) biomass (tonnes) and (b) fisheries
catches (tonnes) from 1956 (baseline Ecopath community model) to 2001. Ecosim
results are lines and annual data are squares, except that the tuna-catch model (and
some other groups not shown) was forced to exactly fit catch data (triangles) to
improve overall model results. The model is an overall best fit of Ecosim to data for
all 32 community categories.
(After Heymans et al. 2009, their Figs. 1 and 2).