Validation of Abundance Estimates from Mark–Recapture and Removal Techniques for Rainbow Trout Captured by Electrofishing in Small Streams

Abstract

Estimation of fish abundance in streams using the removal model or the Lincoln-Peterson mark-recapture model is a common practice in fisheries. These models produce misleading results if their assumptions are violated. We evaluated the assumptions of these two models via electrofishing of rainbow trout Oncorhynchus mykiss in central Idaho streams. For one-, two-, three-, and four-pass sampling effort in closed sites, we evaluated the influences of fish size and habitat characteristics on sampling efficiency and the accuracy of removal abundance estimates. We also examined the use of models to generate unbiased estimates of fish abundance through adjustment of total catch or biased removal estimates. Our results suggested that the assumptions of the markrecapture model were satisfied and that abundance estimates based on this approach were unbiased. In contrast, the removal model assumptions were not met. Decreasing sampling efficiencies over removal passes resulted in underestimated population sizes and overestimates of sampling efficiency. This bias decreased, but was not eliminated, with increased sampling effort. Biased removal estimates based on different levels of effort were highly correlated with each other but were less correlated with unbiased mark-recapture estimates. Stream size decreased sampling efficiency, and stream size and instream wood increased the negative bias of removal estimates. We found that reliable estimates of population abundance could be obtained from models of sampling efficiency for different levels of effort. Validation of abundance estimates requires extra attention to routine sampling considerations but can help fisheries biologists avoid pitfalls associated with biased data and facilitate standardized comparisons among studies that employ different sampling methods.

Evaluation of an Unsuccessful Brook Trout Electrofishing Removal Project in a Small Rocky Mountain Stream

“…Previous studies (Riley and Fausch 1992;Peterson et al 2004b;Rosenberger and Dunham 2005) have demonstrated that electrofishing removals can produce biased estimates of abundance because CP is lower than what is actually measured, producing abundance estimates that are too low, and in our case, removal efficiency estimates that were probably too high. We have no way of knowing true CP in this study.…”

“…Although CP from multipass removals averaged 0.78, we realized that this was probably an overestimate (Riley and Fausch 1992;Peterson et al 2004b;Rosenberger and Dunham 2005), which would in turn cause underestimation of population abundance and overestimation of removal efficiency (especially for age-0 fish). Estimates of abundance should therefore be considered minimum estimates, and estimates of CP and removal efficiency should be considered maximum estimates.…”

Climate change, forests, fire, water, and fish: Building resilient landscapes, streams, and managers

“…We examined patterns of population density among these three classes of streams (for methods see Rosenberger and Dunham 2005;Dunham and others 2007), and found that density of age 1+ fish differed among the three classes of streams (Kruskal-Wallis Analysis of Variance, X 2 =5.96, df=2, p = 0.02). In streams without a recent history of wildfire densities of age 1+ fish were highest (0.22 fish/m 2 ), those in burned streams were intermediate (0.17 fish/m 2 ), and densities in reorganized streams were lowest (0.05 fish/m 2 ).…”

Fish-Based Indices in Catalan Rivers: Intercalibration and Comparison of Approaches

“…One disadvantage of using fish as ecological indicators is that their population densities are more difficult to estimate accurately and their catchability depends on a number of factors including electrofishing equipment, the characteristics of the river reach [11][12][13], and species-specific features such as morphology or behavior [14,15]. The estimation of catchability and intercalibration of data are important to combine data from different fishing teams and to develop protocols for future work or monitoring [12].…”

“…The estimation of catchability and intercalibration of data are important to combine data from different fishing teams and to develop protocols for future work or monitoring [12]. Habitat quality is often assessed during fish sampling [16,17] and inconsistency of habitat assessment among researchers has been also reported by several researchers (e.g., [18][19][20][21][22]).…”