Spatiotemporal Distributions of Threadfin Shad in Tropical Reservoirs

Abstract

AbstractEffective predator management requires accurate knowledge of prey availability and dynamics, yet prey species are rarely studied as intensely as their predatory counterparts. Almost nothing is known about the biology or behavior of threadfin shad Dorosoma petenense in tropical environments, despite that this species is the primary prey in many systems. The goals of this study were to examine spatial and temporal patterns in threadfin shad abundance in tropical reservoirs and to refine protocols for sampling this species. A custom, 3 × 3‐m frame trawl was used to sample four reservoirs in Puerto Rico every 3 months for 1 year. Samples were stratified by longitudinal section (upper and lower lake) and by depth strata (0–3, 3–6, 6–9 m). Threadfin shad densities varied markedly among seasons, reservoirs, longitudinal sections, and depth (overall model: F 36, 188 = 32.0, P < 0.001, coefficient of variation [CV] = 22.1%, r 2 = 0.86). Density estimates from samples were highest in spring and winter when there was substantially more larval production, and density fluctuated as much as 42‐fold among seasons within reservoirs. Differences in mean density were as much as 23 times greater in productive reservoirs than in the least productive reservoir during the same season. Longitudinal variability was substantial between upper and lower reservoir sections (mean percent difference = 70%), but patterns were not consistent among reservoirs or seasons. Threadfin shad showed clear preference for shallow depth strata where 75.4% of fish were collected in the 0–3‐m stratum, and only 2.4% collected in the 6–9‐m stratum. Oxygen profiles suggested that threadfin shad selected shallow depths even when dissolved oxygen levels were suitable in deeper depth strata. Understanding spatiotemporal patterns in threadfin shad distributions will improve sampling programs in terms of timing and frequency of sampling, longitudinal and vertical sampling design, and sample size requirements.Received March 19, 2012; accepted June 2, 2012