Academic
Publications
Hankin and Reeves' Approach to Estimating Fish Abundance in Small Streams: Limitations and Alternatives

Hankin and Reeves' Approach to Estimating Fish Abundance in Small Streams: Limitations and Alternatives,10.1577/1548-8659(2003)132<0069:HARATE>2.0.CO;

Hankin and Reeves' Approach to Estimating Fish Abundance in Small Streams: Limitations and Alternatives   (Citations: 14)
BibTex | RIS | RefWorks Download
Hankin and Reeves' (1988) approach to estimating fish abundance in small streams has been applied in stream fish studies across North America. However, their population estimator relies on two key assumptions: (1) removal estimates are equal to the true numbers of fish, and (2) removal estimates are highly correlated with snorkel counts within a subset of sampled stream units. Violations of these assumptions may produce suspect results. To determine possible sources of the assumption violations, I used data on the abundance of steelhead Oncorhynchus mykissfrom Hankin and Reeves' (1988) in a simulation composed of 50,000 repeated, stratified systematic random samples from a spatially clustered distribution. The simulation was used to investigate effects of a range of removal estimates, from 75% to 100% of true fish abundance, on overall stream fish population estimates. The effects of various categories of removal-estimates-to-snorkel- count correlation levels (r 5 0.75-1.0) on fish population estimates were also explored. Simulation results indicated that Hankin and Reeves' approach may produce poor results unless removal estimates exceed at least 85% of the true number of fish within sampled units and unless correlations between removal estimates and snorkel counts are at least 0.90. A potential modification to Hankin and Reeves' approach is the inclusion of environmental covariates that affect detection rates of fish into the removal model or other mark-recapture model. A potential alternative approach is to use snorkeling combined with line transect sampling to estimate fish densities within stream units. As with any method of population estimation, a pilot study should be conducted to evaluate its usefulness, which requires a known (or nearly so) population of fish to serve as a benchmark for evaluating bias and precision of estimators.
Cumulative Annual
View Publication
The following links allow you to view full publications. These links are maintained by other sources not affiliated with Microsoft Academic Search.
    • ...Although there are drawbacks inherent to visual snorkel surveying (Ensign et al. 1995; Thompson 2003), similar studies of benthic species have suggested visual detection to be as good as traditional methods (Hankin and Reeves 1988) and preferable for use with threatened and endangered species (Jordan et al. 2008)...

    Stephen R. Midwayet al. Habitat Suitability of the Carolina Madtom, an Imperiled, Endemic Stre...

    • ...Factors such as habitat features can contribute to bias in both low- and high-intensity estimates (Thompson 2003), and the relationship between the two indices may vary unpredictably in different contexts (Williams et al. 2004)...

    Amanda E. Rosenbergeret al. Validation of Abundance Estimates from Mark–Recapture and Removal Tech...

    • ...We did not apply electrofishing as a means to assess true abundance of juvenile steelhead, and diver movements in a sample unit may have caused fish to evade diver detection (Thompson 2003); however, we did at times employ an additional diver (or an observer on the creek bank) near the upstream boundary of the sample unit to note any fish leaving the sample unit...

    Anthony P. Spinaet al. Downstream Migration, Rearing Abundance, and Pool Habitat Associations...

    • ...Observer efficiency estimates made from observations of radio-tagged trout were positively related to levels of horizontal visibility in our study reach (Figure 1), but the strength of the predictive relationship was inconsistent among years (see Table 1 for regression data summary). During 2001, variability in horizontal visibility explained only 3.8% of the variability in observer efficiency for that year, and the regression was not significant (P = 0.675). In contrast, observer efficiency was significantly related to horizontal visibility in both 2002 (P = 0.013) and 2003 (P < 0.001), 74% and 99.7%, respectively, of the variability in observer efficiency being explained by visibility changes.

      @@@@ Figure 1. Observer efficiency estimates (number of tagged fish seen relative to the number known to be present) versus horizontal underwater visibility for 3 years of periodic surveys in the Salmo River, British Columbia. Although untransformed data are depicted here, regression analyses were conducted using arcsine-square-root-transformed observer efficiency data

      @@@@Regression data for individual years were not prescreened for outliers because error analysis for observer efficiency regressions with such few data points should not be considered realistic (Thompson 2003)...

    John Hagenet al. Accuracy of Diver Counts of Fluvial Rainbow Trout Relative to Horizont...

    • ...1992; Riley et al. 1993; Thompson 2003) have concluded that such samples are negatively biased, but high catchability should yield relatively precise, repeatable population estimates (McKelvey and Pearson 2001)...

    Michael K. Younget al. Population Characteristics of Greenback Cutthroat Trout in Streams: Th...

Sort by: