Interpretation of Long-Term Datasets depends on both Landscape and Temporal Sampling Design: Examples Using a 50 Year Record of Recreational, Commercial and Nuisance Invasive Fish in a Large River

A benefit of developing long-term monitoring programs for assessment fisheries management actions or policy changes is the value of clear trends despite inter-annual variability or noise.  However there are important considerations that can influence how the trends can be interpreted and acted on.  We present some insights drawn from multi-decade monitoring programs in the Illinois River, a mid-west floodplain system. One program is a randomly sampled and stratified over 4 habitats in an 80 km reach. The second program uses a series fixed stations across 225 km of the same river. Comparison of the resulting trends doesn’t reveal appreciable conflicts. However, the effect of inclusion of habitat-specific sampling is a finer spatial resolution of where the trends are occurring. A second insight comes from the disparity in record lengths: The fixed program has a 5 decade record while the stratified random program has 2 decades.  Comparison shows that this difference leads to subtle perceived shifts in trends of abundance, despite the fact that both records are looking at the same fish community composition. When developing and using programs like these, agencies should be aware that some aspects of interpretation and conclusions are influenced by spatial/temporal attributes of the sampling.