We conducted surveys of the fish populations in nine streams across the state of AZ, USA from spring 2016 – spring 2019. Surveys were conducted four times a year at biologically significant seasonal timepoints: after winter storms, before summer monsoons and drought, after summer monsoons, and before winter storms. Not all sites were surveyed in all years, and floods and streamflow intermittence prevented sampling at a few sites in some seasons.
During each survey, we conducted three-pass depletion backpack electrofishing (Model LR-24 Electrofisher, Smith- Root, Vancouver, Washington, USA) of a 100-m stream reach at each of the study sites. The same reaches were revisited during each survey and were blocked at the upstream and downstream ends with 6-mm mesh. All captured fish were anaesthetized with tricaine methanosulphate (MS-222), weighed to the nearest g, measured (fork length; mm), and identified to species.
We estimated the mass of specimens of four species (AGOCHR, CYPLUT, GAMAFF, LEPCYA, and PIMPRO; see taxonomic data in this dataset for species names) that were too small to accurately weigh in the field using length-weight regressions developed from individuals stored in the lab. Regressions to predict weight (Wfinal in g) were created for individuals of each species in the following length ranges:
AGOCHR 18-60 mm, CYPLUT 25-60 mm, GAMAFF 16-45 mm, LEPCYA 24-70 mm, PIMPRO 30-77 mm
Length (L; mm) weight (W; g) regressions for each species:
AGOCHR
W = 0.9796924 - L * 0.0828393 + L^2 * 0. 0019155
CYPLUT
W = 1.9168015 - L * 0.1165794 + L^2 * 0.0020978
GAMAFF
W = 0.4567813 - L * 0.0472909 + L^2 * 0.0013693
LEPCYA
W = 3.4840746 - L * 0. 2149298 + L^2 * 0. 0036709
PIMPRO
W = 2.5359964 - L * 0.1575278 + L^2 * 0.0027142
To calculate fish population sizes, we used the FSA package in R (Ogle et al. 2020) using k-pass depletion and maximum weighted likelihood estimation. We used the “CarleStrub” method for catch data where the number of individuals of a species captured was greatest in pass 1 and declined through pass 3. We used the "Schnute" method when the greatest number of individuals capture was not in pass 1.
REFERENCES
Ogle, D. H., P. Wheeler, and A. Dinno. 2020. FSA: fisheries stock analysis. R package version 0.8.31.9000. https://github.com/droglenc/FSA.