OPTIMIZING TEMPORAL SAMPLING STRATEGIES FOR BENTHIC ENVIRONMENTAL MONITORING PROGRAMS

Trends in benthic macrofaunal monitoring data are often confounded by natural seasonal fluctuations in abundance, We used eight years of sea sonally-collected benthic biological data to investigate the relative advantages of multi-season vs single season sampling strategies for as sessing trends in the context of these seasonal fluctuations. The effe ct of seasonality on sampling strategy was examined in three ways. Fir st, we tested long-term trends in benthic responses for homogeneity am ong seasons with the premise that a single season sampling strategy wo uld be inappropriate if the direction of trends differed among seasons . Second, we compared the power for trend detection of several samplin g regimes, simulating the distribution of a pre-determined number of s amples across seasons in several ways. Third, we determined whether th e magnitude of differences in benthic response between reference and d egraded sites changed among seasons. Each test was applied to four ben thic response measures: abundance; biomass; diversity; and proportiona l abundance of opportunistic taxa. The direction and magnitude of long -term trends were quite homogenous between seasons. No contradictory s easonal interactions were detected for any of the trends, Though advan tageous, we found that four season sampling is not necessary for condu cting trends analysis because the power of some two season alternative s (e.g. spring-summer) was only slightly less than the power for four seasons, For each of the four response variables we examined, power fo r trend detection was consistently higher when sampling in four season s than when sampling in one or two seasons, Of the individual seasons, we found that summer is the best season in which to sample, Summer yi elded the greatest power for trend detection, although power differenc es among seasons were mostly limited to the abundance measure, Summer was also the season when difference in benthic response between refere nce and degraded sites was greatest. (C) 1997 Elsevier Science Ltd.