Thomas M. Frost

Habitat duration and community structure in temporary ponds

Models of the factors affecting community structure following a disturbance differ in the emphasis placed on adaptations to the physical environment and to biotic interactions. We investigated the hypothesis that the duration of a habitat following disturbance mediates the relative importance of physical and biotic control. We combined detailed information on natural populations in a series of temporary ponds with small-scale experiments on specific processes. We compared data on presence/absence and abundance of taxa from temporary ponds showing a gradient of habitat duration with 3 simple models of community structure that incorporate random forces, life history adaptations, and biotic interactions, respectively. Different scales of resolution in describing community structure provide different emphases on which processes are important in these communities. The Life History Model explained patterns of presence/absence in the ponds, but not patterns of abundance of many taxa which best fit a model in which the importance of predation and competition in structuring communities increases with increasing duration. Experimental studies confirmed that the effect of predation and competition in temporary pond communities depends on habitat duration. Predation on all prey species examined increased with increasing pond duration. Predators were more diverse and more abundant in long-duration ponds where effects of their consumption on the abundance of prey taxa were greater than in shorter-duration ponds. Predators from long-duration ponds fed preferentially on prey taxa that predominate in shorter-duration habitats such as Aedes mosquitoes and the fairy shrimp Eubranchipus. Competition between Daphnia and rotifers occurred primarily in intermediate-duration ponds, where Daphnia at typical denities was capable of drastically reducing abundances of Keratella cochlearis in cultures. However, at densities found in both short- and long-duration habitats, Daphnia had no effect.

Randomized Intervention Analysis and the Interpretation of Whole‐Ecosystem Experiments

Randomized intervention analysis (RIA) is used to detect changes in a ma- nipulated ecosystem relative to an undisturbed reference system. It requires paired time series of data from both ecosystems before and after manipulation. RIA is not affected by non-normal errors in data. Monte Carlo simulation indicated that, even when serial au- tocorrelation was substantial, the true P value (i.e., from nonautocorrelated data) was <.05 when the P value from autocorrelated data was <.01. We applied RIA to data from 12 lakes (3 manipulated and 9 reference ecosystems) over 3 yr. RIA consistently indicated changes after major manipulations and only rarely indicated changes in ecosystems that were not manipulated. Less than 3% of the data sets we analyzed had equivocal results because of serial autocorrelation. RIA appears to be a reliable method for determining whether a nonrandom change has occurred in a manipulated ecosystem. Ecological argu- ments must be combined with statistical evidence to determine whether the changes dem- onstrated by RIA can be attributed to a specific ecosystem manipulation.

Species Compensation and Complementarity in Ecosystem Function

Functional complementarity occurs when ecosystem processes are maintained at constant levels despite stresses that induce shifts in the populations driving those processes. Understanding when such complementarity occurs depends on an integration of perspectives from population and ecosystem ecology. Here, we examine the extent to which functional complementarity is linked with compensatory dynamics among species that carry out a particular ecosystem function. Our approach combines analyses of long-term zooplankton data from a whole-lake acidification experiment with a theoretical treatment of species compensation. Results from the acidification of Little Rock Lake, WI indicated that the biomass of cladocerans, copepods, rotifers, and total zoo-plankton remained at high levels despite the loss of component species from each group. Compensatory increases by other taxa were responsible for this complementarity of function. Theoretical considerations indicated that the degree of compensation occurring among species in response to environmental change increased in response to two different factors: the functional similarity of interacting species and the degree to which an environmental change acts nonuniformly on their interactions. Finally, we tested the extent to which compensation in unperturbed systems could predict functional complementarity. We analyzed a 7-year record from the reference basin of Little Rock Lake and found that substantial compensation occurred only in the natural dynamics of rotifers and cladocerans during some seasons. Complementarity in response to acidification was evident, however, among copepods and total zooplankton in addition to rotifers and cladocerans, and could not have been predicted solely on the basis of compensation prior to stress. Taken together, our analyses reveal how functional complementarity is linked to a variety of compensatory interactions among species responding to stress.

Inferences from Spatial and Temporal Variability in Ecosystems: Long-Term Zooplankton Data from Lakes

Ecosystem patterns and processes exhibit spatial and temporal variability. Analysis of this variability can reveal which aspects of a system are specific to sites or to years. Patterns in this variability yield insights into the forces structuring ecosystems. As an example of this approach, we examined an unpublished zooplankton data set collected by Birge, Juday, and coworkers in the summers of 1929 through 1941 from five neighboring northern Wisconsin lakes. Eight taxa were common to the five lakes: two rotifers (Keratella cochlearis and Kellicottia), four cladocerans (Daphnia pulex, Daphnia galeata mendotae, Diaphanosoma, and Bosmina), and two copepods (cyclopoid and calanoid). Taxa differed in their consistency in depth distribution and seasonal timing. For example, Diaphanosoma peaked at the same relative depth and time in each of our five study lakes. In contrast, Bosmina varied among lakes in the timing of its maximum abundance, and calanoid copepods varied in the depth of their maxima. In general, among-year variability was greater in zooplankton abundance than in seasonal timing, which, in turn, was more variable than depth distribution. Among-year variability in maximum abundance was related to a taxons potential for rapid population increase, r

Experimental acidification of Little Rock Lake, Wisconsin

_\max

COMPENSATORY DYNAMICS IN ZOOPLANKTON COMMUNITY RESPONSES TO ACIDIFICATION: MEASUREMENT AND MECHANISMS

. Variability among years was greater in rotifer taxa than in cladoceran taxa, which were more variable than copepod taxa. The abundance and depth distribution of copepods and cladocerans exhibited relatively greater variability among lakes than among years, suggesting that conditions specific to lakes are important in controlling these parameters. This conclusion was supported by data from Lake Washington; except in one case, the same taxa that showed significant site specificity in Wisconsin changed significantly in abundance in Lake Washington after sewage diversion. In contrast, we inferred that variation in weather was an important determinant of seasonal timing for all eight taxa and of the maximum abundance and depth of maximum abundance of the two rotifers. An ability to predict general site-specific and year-specific ecosystem parameters has significant practical and theoretical importance. For example, taxa or parameters exhibiting large variability among sites are sensitive indicators of change in an ecosystem, whereas those parameters or taxa exhibiting large temporal variation should be studied to understand factors influencing seasonal or yearly dynamics. We suggest that similar analyses of spatial and temporal variability could provide useful comparisons of ecosystems ranging from aquatic to terrestrial habitats and comprising markedly different taxa.

Patterns of Primary Production and Herbivory in 25 North American Lake Ecosystems

The controlled acidification of a two-basin lake is described. The lake was divided by a vinyl curtain in 1984; acidification of one basin began in 1985. Target pH values of 5.5, 5.0 and 4.5 are planned for 2-yr increments. Biotic and chemical responses and internal alkalinity generation are being studied. Baseline studies, initial results at pH 5.5, and predictions of lake responses to acidification are described.

Pigment ratios and phytoplankton assessment in northern Wisconsin lakes

Previous studies indicate substantial variation in ecological responses to perturbation. In some cases, ecosystems are resilient to perturbation due to compensatory dynamics in which losses of sensitive species are offset by population increases of species that perform similar ecological functions. Here, we report a detailed evaluation of compensatory dynamics in zooplankton community responses to the experimental acidification of Little Rock Lake, Wisconsin, USA. We used a variance ratio to quantify compensatory dynamics in functional groups of zooplankton containing species that use similar resources and are vulnerable to the same predators. We also used first-order autoregressive models to explore mechanisms driving the dynamics of each functional group. Our results indicate that responses of functional groups to acidification can be highly variable. Herbivorous copepods and medium-sized herbivorous cladocerans exhibited significant compensatory dynamics in response to acidification, whereas other functional groups exhibited independent or synchronous dynamics. First-order autoregressive models indicated that groups exhibiting compensatory dynamics contained both acid-tolerant and acid-sensitive species that competed. In contrast, groups that contained only acid-sensitive or acid-tolerant species exhibited more independent or synchronous dynamics. Overall, our study highlights the combined roles of sensitivity to environmental perturbation and species interactions in determining the extent of compensatory dynamics in zooplankton functional group responses to acidification.

Little Rock Lake (Wisconsin): Perspectives on an experimental ecosystem approach to seepage lake acidification

The effects of nutrients and herbivory on phytoplankton biomass and production were examined, using data from 25 lakes studied for 2 to 6 years each. Variance among lakes was substantially greater than variance among years, for all physical, chemical, phytoplankton, and zooplankton variates studied. Experimentally manipulated lakes had coefficients of variation within the range exhibited by nonmanipulated lakes. Graphical, correlative, and regression analyses illustrated the significant joint effects of both nutrients and herbivory on phytoplankton biomass and production. A Bayesian analysis of sensitivity to new information showed that the statistical models for chlorophyll are quite robust. Statistical models for primary production were deemed less conclusive, because primary production was measured in fewer lakes. We provide a list of common challenges in comparative statistical analysis of ecosystems and explain their implications for our study. The major pattern apparent in our data—that summer chlorophyll responds positively to nutrients and negatively to herbivore size — is congruent with results of whole-lake experiments in which nutrients or predators were manipulated.

Complex biological responses to the experimental acidification of Little Rock Lake, Wisconsin, USA.

Nine lakes in northern Wisconsin were sampled from February through September 1996, and HPLC analysis of water column pigments was carried out on epilimnetic seston. Pigment distributions were evaluated throughout the water column during summer in Crystal Lake and Little Rock Lake. The purpose of our study was to investigate the use of phytopigments as markers of the main taxonomic groups of algae. As a first approach, multiple regression of marker pigments against chlorophyll a (chl a) was used to derive the best linear combination of the main xanthophylls (peridinin, fucoxanthin, alloxanthin, lutein, and zeaxanthin). A significant regression equation (r2= 0.98) was obtained for epilimnion data. The good fit indicates that the chl a:xanthophyll ratios were fairly constant in the epilimnion of the nine lakes over time. Chlorophyll a recalculated from the main xanthophylls in each sample showed good agreement with measured chl a in epilimnetic waters. A second approach used the CHEMTAX program to analyze the same data set. CHEMTAX provided estimates of chl a biomass for all algal classes and allowed distinction between diatoms and chrysophytes, and between chlorophytes and euglenophytes. These results showed a reasonably good agreement with biomass estimates from microscope counts, despite uncertainties associated with differences in sampling procedure. Changes of pigment ratios over time in the epilimnetic waters were also investigated, as well as differences between surface and deep samples of Little Rock Lake and Crystal Lake. We found evidence that changes in the ratio of photoprotective pigments to chl a occurred as a response to changes in light climate. Changes were also observed for certain light‐harvesting pigments. The comparison between multiple regression and CHEMTAX analyses for inferring chl a biomass from concentrations of marker pigments highlighted the need to take account of variations in pigment ratio, as well as the need to acquire additional data on the pigment composition of planktonic algae.