Evaluation and application of pectoral spine microchemistry to identify stocked Channel Catfish and infer fish size at stocking
Morgan Winstead, Allison M Asher, Gregory W WhitledgeAbstract
Objective
Channel Catfish Ictalurus punctatus populations are commonly supported by stocking, and knowledge of stocked fish contribution is helpful to inform decisions regarding fish stocking locations, rates, and sizes. Calcified structure microchemistry can be useful for identifying stocked fish but has not been applied to distinguish fish stocked at different sizes. We compared pectoral spine microchemistry of Channel Catfish from hatchery and rearing facilities to four lakes in Arkansas to evaluate this technique for identifying stocked fish and estimated contributions of fish stocked as yearlings or at catchable size in three lakes.
Methods
Sectioned pectoral spines from Channel Catfish sampled from four lakes, one hatchery, and two rearing facilities were analyzed for strontium : calcium and barium : calcium along a transect from the spine core to the edge using laser ablation-inductively coupled plasma mass spectrometry. Random forest models were used to assess classification accuracy of assigning individual fish to collection locations using pectoral spine edge strontium : calcium and barium : calcium. Lake-specific random forest models were applied to classify individual Channel Catfish sampled from each of the three primary study lakes as stocked or wild fish. For Channel Catfish identified as stocked individuals, the distance that the hatchery or rearing facility microchemical signature extended outward from the pectoral spine nucleus was used as an indicator of whether each individual was stocked as a yearling or catchable-size fish.
Results
Lake-specific random forest models generally yielded high accuracy (88–100%) for assigning individual Channel Catfish to their collection location using pectoral spine edge microchemistry data when lakes and hatchery or rearing facilities were in different ecoregions, although fish from two of the lakes were much more difficult to distinguish from hatchery residents (43–50% assignment accuracy). However, Channel Catfish from both of these lakes were distinguishable (88–100% accuracy) from rearing facility residents. All individuals sampled from rearing facilities showed expected changes in pectoral spine microchemistry reflective of transfer from the hatchery to the rearing facility. Estimated contributions of stocked fish ranged from 0% to 83% among lakes. Mean distances that hatchery signatures extended outward from the spine core were 549 ± 92 μm (mean ± SD) for yearlings and 1,242 ± 290 μm (mean ± SD) for catchable-size fish; 56–68% of stocked fish were identified as having been stocked at catchable size.
Conclusions
Pectoral spine microchemistry provides a nonlethal approach to identify stocked catfish and infer fish size at stocking when hatchery or rearing facilities and stocking locations are chemically distinct. Differences in estimated contributions of stocked Channel Catfish to study lakes may have been influenced by differences in stocking or harvest rates. We expect that the approach we used to infer fish size at stocking from microchemistry data will be applicable to other fish species and other types of calcified structures when different sizes of fish are stocked into a water body from the same hatchery sources.