A Comparison of Full-Station Visual and Transect-Based Methods of Conducting Habitat Surveys in Support of Habitat Suitability Index Models for Southern Ontario

Abstract

Over the years, ecologists have developed several models that reliably relate stream habitat and fish biomass. Under today's development pressures, managers are demanding that inputs to these models be objectively and repeatably applied by minimally trained crews and at low cost. In this study we compared how repeatable three techniques were at measuring the surficial area of morphological habitat features of wadable streams in southern Ontario. Two crews surveyed 20 sites twice with two survey designs (20 sites × 2 crews × 2 visits × 2 survey designs = 160 visits). For the first two visits, the crews surveyed the entire site (minimum length = 40 m, beginning and ending at a crossover) using a technique that entailed visually determining and then mapping the spatial extent occupied by morphological habitat units. For the next two visits, the crews applied the same visual classification system to a point–transect design (evenly spaced observation points located along evenly spaced transects placed at right angles to the flow. At each point, the crews made measurements of six stream variables: depth, velocity, velocity range, water surface slope, vertical roughness, and substrate. These observations were used to characterize each point as a morphological unit by means of a hierarchical key. We compared the percentage of each habitat type and an overall habitat suitability score observed at each site between visits and crews for each of the techniques. We found that the measured point–transect technique was more repeatable than either of the visual techniques for within-crew and especially between-crew comparisons. We speculate that the lack of repeatability for the visual techniques are largely due to difficulties encountered by observers in consistently identifying transition habitats, such as flats and runs, and to differences among observers in the scale at which they recognize individual habitat units. We estimate that as few as 50–60 point observations within a typical third-order stream site (100 m²) should be sufficient to characterize the morphological features of the stream for the purposes of determining habitat suitability with acceptable precision (i.e., a coefficient of variation [100·SD/mean] of <10%).