POPULATION, COMMUNITY, AND ECOSYSTEM VARIATES AS ECOLOGICAL INDICATORS - PHYTOPLANKTON RESPONSES TO WHOLE-LAKE ENRICHMENT

We quantified the reliability of phytoplankton population, community, and ecosystem variates as indicators of whole-lake enrichment. Variate s with high sensitivity to perturbation and low background variability were assumed to be more reliable indicators than variates with low se nsitivity or high variability. Our data set included weekly data in fo ur lakes (three manipulated and one reference) during two pretreatment summers and two summers of enrichment. We determined background varia bility by evaluating change in each variate from year to year in the r eference lake throughout the experiment and in each manipulated lake d uring the pretreatment period. We evaluated sensitivity to enrichment using the frequency of departure from expected conditions for each var iate during the period of experimental enrichment. Using this informat ion, we then (1) tested the expectation that species populations are m ore reliable indicators of perturbation than ecosystem variates (bioma ss, chlorophyll, and primary productivity), and (2) evaluated whether community variates (genera, taxonomic divisions, allometric variates, and community diversity) were reliable indicators of enrichment. Contr ary to expectations from other perturbations, phytoplankton species po pulations were less reliable indicators of enrichment than community a nd ecosystem variates. Chlorophyll, species diversity, and species eve nness were the most reliable indicators of enrichment: each changed si gnificantly only during the first year of enrichment and only in the t hree enriched lakes. Simultaneous changes in multiple taxonomic divisi ons also signaled enrichment very reliably. In contrast, the frequency of significant changes in species populations differed little between the reference lake and the enriched lakes, even after experimental en richment. Changes in species were difficult to detect reliably due to high background variability in all four lakes: most taxa were not pres ent often enough during a single year to assess reliably whether they had increased or decreased compared to the previous years. Genera and allometric variates were also unreliable indicators due to high variab ility and moderate sensitivity, respectively. Reliable indicators of p hytoplankton responses to enrichment were very different from reliable indicators of animal responses to toxic stressors, suggesting that it may be difficult to make generalizations regarding the use of populat ion, community, and ecosystem variates as indicators of a wide array o f perturbations.