Predictions under uncertainty : fish assemblages and food webs on the Grand Bank of Newfoundland
Six broad zoogeographic areas on the Grand Bank, each characterized by a relatively homogeneous and persistent species composition, are identified, described, and mapped. The contours of these areas consistently recurred around the same geographic positions year after year (1971-82, 1984-87) and wer...
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
Memorial University of Newfoundland
1991
|
| Subjects: | |
| Online Access: | https://research.library.mun.ca/1112/ https://research.library.mun.ca/1112/1/Gomes_ManuelDoCarmoMSc.pdf https://research.library.mun.ca/1112/3/Gomes_ManuelDoCarmoMSc.pdf |
| Summary: | Six broad zoogeographic areas on the Grand Bank, each characterized by a relatively homogeneous and persistent species composition, are identified, described, and mapped. The contours of these areas consistently recurred around the same geographic positions year after year (1971-82, 1984-87) and were found to be strongly aligned with bottom depth and oceanographic circulation on the Bank. Consideration of their overall biological coherence led to the reformulation of the original six areas into four major regions that are suggested to define an appropriate scale for studies at the community level on the Grand Bank. Within the two major zoogeographic regions, Southern and Northeastern, there is a complex network of feeding interactions whose major year-round features have been summarized in the form of stereotyped community food webs. On average, the Grand Bank food chains are short, as is the case in other marine food webs, but they exhibit an uncommon degree of omnivory. Examination of diet overlap of predators on the Grand Bank suggests that body size is the most important structuring factor of niche space. Press perturbations, a modeling exercise in which individuals of a given species are continuously removed from the community, are used to investigate the consequences of two major sources of uncertainty (uncertainty about model structure and about parameter values) that plague attempts to make predictions about long term changes in species abundance. It is shown that long term predictions are highly sensitive to details of interactions in community models and therefore community dynamics must be, to a large extent, indeterminate over time. |
|---|