Number Of Sampling Events Research Articles

Temporal variability in stream fish assemblage metrics and implications for long-term monitoring

High natural variability in the condition of fish communities in headwater streams complicates detection of long-term responses to changes in water quality. As a result, little is known about the impacts and recovery of fishes from acid deposition in streams of New York. Twenty-one fish metrics from annual electrofishing surveys at 13 streams sites in the Catskill and Adirondack mountains were assessed to quantify temporal variability and identify effective monitoring strategies for detecting change in headwater stream fish assemblages. Metrics included the density and biomass of Brook Trout populations and entire fish communities using length-, area-, and effort-based standardization techniques. Linear mixed models were used to estimate changes in coefficients of variation (CV) for different classes of metrics and standardization techniques, and a simulation-based power analysis was conducted to assess differences in statistical power for each metric with various sampling designs. Metric CV varied significantly as a result of standardization technique and whether metrics were calculated for the entire community or Brook Trout only. The sampling effort necessary to detect a 30% change with power of 0.80 was strongly correlated with CV (R2 = 0.77). The number of sampling events at the 13 sites needed to detect this change ranged from 1 to >60, suggesting metric selection can strongly affect statistical power and the resources necessary to detect change. Thus, metric selection is a complex decision that must consider study objectives and biological relevance, in addition to natural variability and statistical power. However, adequate statistical power was achieved at relatively small sample sizes using certain metrics, indicating that fish communities in headwater streams can be a valuable component of long-term assessments of acidification impacts and recovery.

Distribution of long-horn beetles (Cerambycidae: Coleoptera) within the Fijian archipelago

Long-horn beetles (Family: Cerambycidae) in Fiji consist of 124 described species within 45 genera, of which 110 (88.7%) species are considered endemic. Despite their conservation value, ecological significance and cultural importance, little scientific research has been conducted on the taxonomy or ecology of Cerambycidae in Fiji. This biogeographical study surveyed Cerambycidae by Malaise trapping on ten Fijian Islands. A total of 438 individuals and 44 species of Cerambycidae were recorded. Thirty three of the species collected are endemic to Fiji; three other species are native and eight species are new records for Fiji and/or new species. Twenty seven species were recorded from only one island and 20 species were recorded only as singletons. There was an expected significant relationship between the number of species collected on an island and the number of sampling events. The highest number of species, 23, was recorded on the largest island, Viti Levu, followed by Gau with 13 species and Vanua Levu and Kadavu with 12 species each. There was a positive relationship between species richness and island size but this was lost if the effect of sample number was taken into account. The results indicate that the species-area relationship may hold for Fijian Cerambycidae, but additional collecting events, over more of the annual cycle, and involving multiple collecting methods may be required to fully catalogue the current Fijian fauna.

Evaluating mark - recapture sampling designs for fish in an open riverine system

Sampling designs for effective monitoring programs are often specific to individual systems and management needs. Failure to carefully evaluate sampling designs of monitoring programs can lead to data that are ineffective for informing management objectives. We demonstrated the use of an individual-based model to evaluate closed-population mark–recapture sampling designs for monitoring fish abundance in open systems, using Murray cod (Maccullochella peelii (Mitchell, 1838)) in the Murray–Darling River basin, Australia, as an example. The model used home-range, capture-probability and abundance estimates to evaluate the influence of the size of the sampling area and the number of sampling events on bias and precision of mark–recapture abundance estimates. Simulation results indicated a trade-off between the number of sampling events and the size of the sampling reach such that investigators could employ large sampling areas with relatively few sampling events, or smaller sampling areas with more sampling events to produce acceptably accurate and precise abundance estimates. The current paper presents a framework for evaluating parameter bias resulting from migration when applying closed-population mark–recapture models to open populations and demonstrates the use of simulation approaches for informing efficient and effective monitoring-program design.