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 seasonally-collected benthic biological data to investigate the relative advantages of multi-season vs single season sampling strategies for assessing trends in the context of these seasonal fluctuations. The effect of seasonality on sampling strategy was examined in three ways. First, we tested long-term trends in benthic responses for homogeneity among seasons with the premise that a single season sampling strategy would be inappropriate if the direction of trends differed among seasons. Second, we compared the power for trend detection of several sampling regimes, simulating the distribution of a pre-determined number of samples across seasons in several ways. Third, we determined whether the magnitude of differences in benthic response between reference and degraded sites changed among seasons. Each test was applied to four benthic response measures: abundance; biomass; diversity; and proportional abundance of opportunistic taxa. The direction and magnitude of long-term trends were quite homogenous between seasons. No contradictory seasonal interactions were detected for any of the trends, Though advantageous, we found that four season sampling is not necessary for conducting trends analysis because the power of some two season alternatives (e.g. spring-summer) was only slightly less than the power for four seasons, For each of the four response variables we examined, power for trend detection was consistently higher when sampling in four seasons 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 yielded the greatest power for trend detection, although power differences among seasons were mostly limited to the abundance measure, Summer was also the season when difference in benthic response between reference and degraded sites was greatest. (C) 1997 Elsevier Science Ltd.