STATISTICAL COMPARISON OF SEASONAL TRENDS FOR AUTOCORRELATED DATA - A TEST OF CONSUMER AND RESOURCE MEDIATED TROPHIC INTERACTIONS

A 2 x 2 factorial enclosure study of food web biomanipulation was conducted in 1986 and again in 1987 to assess the relative importance of fish and nutrients on zooplankton biomass and size, chlorophyll-a concentration, and water clarity. Seasonally autocorrelated data were analyzed using Kendall's median trend slopes and compared with a randomization test. These results were then compared to the within-subjects effects from repeated measures analysis of variance (ANOVAR). Both methods produce complimentary conclusions when contrasting the different results between years. Trend Slope Analysis (TSA) is a valuable alternative to ANOVAR for temporally dependent data and should be used where there is belief that the time series are complicated by non-normality, missing values, non-polynomial fit due to extreme values, or lack of treatment/subject replication.