Cross-assemblage concordance is lower in disturbed than in reference wetlands: implications for monitoring and assessment

Background/Question/Methods

Wetland biomonitoring is based on the premise that the community composition of one taxonomic group is indicative of the overall biology and the underlying environmental conditions at a wetland.  To be a good bioindicator, there must be adequate concordance between the indicator group and other biotic assemblages.  Otherwise, multi-assemblage monitoring is necessary to glean a complete picture of wetland condition.  In 32 sites ranging from reference wetlands to stormwater retention ponds, we evaluated concordance in community composition among the six most commonly monitored wetland assemblages: waterfowl, wetland dependent songbirds, macroinvertebrates, and plants in the wet meadow, emergent, and open-water vegetation zones.  We also assessed agreement in environmental correlates among these six assemblages and investigated the impact of human disturbance on cross-assemblage concordance.  If cross-assemblage concordance is sensitive to human disturbance, then even if measurement of multiple-assemblages is redundant in relatively pristine wetlands, it may be necessary to ensure accurate evaluation of disturbed wetlands.

Results/Conclusions

We found that cross-assemblage concordance was positive (p < 0.06 in all cases), but relatively low (Mantel R values 0.11 to 0.36), suggesting that the assemblages are mediocre surrogates for one another.  We found a high degree of agreement among environmental correlates, especially along the first axis of assemblage-specific community ordinations (mean Spearman rho = 0.923), indicating that despite low concordance, the six assemblages are likely responding to the same environmental gradients.  Thus, while a single assemblage may not provide an adequate assessment of biological conditions at a wetland, it may yield a reasonable estimate of underlying environmental conditions.  Most important amongst the environmental correlates were sediment and water nutrient levels, shoreline slope, and the area of wet meadow and emergent vegetation zones.  Perhaps most interestingly, we found that the strength of cross-assemblage concordance was greatest in reference wetlands, and was lower in constructed wetlands.  This implies that cross-assemblage concordance present in undisturbed sites may not apply in disturbed wetlands, where congruence in community composition deteriorates.  Furthermore, a general change in cross-assemblage concordance may itself be indicative of human disturbance in wetlands.