less than 10 percent. This observation reinforces the conclusion by most investigators that the very low abundance throughout the Lower Bay Complex in January 1973 was real.

Are the differences in community structure between the earlier (1957-60 and 1973) and later benthic surveys (1986-87 and 1994-95) because of habitat changes or because of changes in faunal associations over time? To examine this question, I compared the regional studies using Mantel tests, with the comparison centering on the 1986-87 sur-veybecause ofits compatibility in methods to prior studies. Almost all comparisons of faunal associations were significant, providing evidence of stability over decadal time scales (Table 18.7). Correlations ranged from 0.24 to 0.60, values that are comparable in magnitude to those obtainedfor annual faunal associations. It is especially interesting to note that the strongest correlation infaunal associations was found between the two winter surveys, January 1973 and January 1987.

In contrast to faunal associations, spatial or habitat associations have changed over decadal time scales within the Lower Bay Complex (Table 18.8). Mantel tests indicated that spatial associations between the early surveys and the 1986-87 study were often nonsignificant, indicating little or no similarity in habitat structure between surveys. In particular, the correlations between 1959-86 and 1973-87 were close to 0 and well below the range indicated for annual variations (0.19-0.43). Even with sieve size differences, strong spatial associations in community structure were indicated in the comparisons between 1986-87 and 1994-95. All four spatial comparisons were significant, and three of the four correlations exceeded the range for

Table 18.8. Mantel tests of spatial associations (normal analysis)

0 0

Post a comment