Info

a3

-3.00 *

-3.00 *

-3.00 *

3.87 *

a4

-3.00 *

-3.00 *

3.87 *

-3.00 *

incorrect, are subject to some risk of error. For example, dominant species may well not be indicative of environmental condition £4 as suggested in Table 6.7. Tables 6.8 and 6.9 also show that dominant species a\ is significantly under-represented in environmental condition £4, suggesting that this condition is strongly adverse to the species.

Ecological application 6.4

Legendre et al. (1982) explored the relationship between the abundance of phytoplankton and vertical stability of the water column in a coastal embayment of Hudson Bay (Canadian Arctic). Surface waters are influenced by the plume of the nearby Great Whale River. There were intermittent phytoplankton blooms from mid-July through mid-September. In order to investigate the general relationship between phytoplankton concentrations (chlorophyll a) and the physical conditions, chl a and salinity data from 0 and 5 m depths were allocated to a contingency table (Table 6.10). The null hypothesis of independence being rejected, the correspondence between the two descriptors rests in three cells. (1) At high salinities (> 22 mg L-1), there is a significantly small number of high chl a (> 1.5 mg m-3) observations. At intermediate salinities (18-22 mg L-1), (2) high chl a observations are significantly numerous, whereas (3) low chl a observations are significantly infrequent. At low salinities (< 18 mg L-1), the numbers observed are not significantly different from the frequencies expected under the null hypothesis of independence.

Table 6.9 Statistics (eq. 6.30) for testing the significance of individual cells in a contingency table. The observed and expected values are in Table 6.4. Values larger than Z[1 _ 005 / (2 x 4 x 4)] = z0 9984 = 2.95 are in boldface. They identify cells in which the number of observations (Oj) significantly (p < 0.05) differs (higher or lower) from the corresponding expected frequency (Ej).

Was this article helpful?

0 0

Post a comment