• Hasler, Caleb T.
  • Suski, Cory D.
  • Hanson, Kyle C.
  • Cooke, Steven J.
  • Tufts, Bruce L.


In this study, field biotelemetry and laboratory physiology approaches were coupled to allow understanding of the behavioral and physiological responses of fish to winter hypoxia. The biotelemetry study compared dissolved oxygen levels measured throughout the winter period with continually tracked locations of nine adult largemouth bass obtained from a whole‐lake submerged telemetry array. Fish habitat usage was compared with habitat availability to assess whether fish were selecting for specific dissolved oxygen concentrations. The laboratory study examined behavioral and physiological responses to progressive hypoxia in juvenile largemouth bass acclimated to winter temperatures. Results from the dissolved oxygen measurements made during the biotelemetry study showed high variance in under‐ice dissolved oxygen levels. Avoidance of water with dissolved oxygen <2.0 mg/L by telemetered fish was demonstrated, but significant use of water with intermediate dissolved oxygen levels was also found. Results from the lab experiments showed marked changes in behavior (i.e., yawning and vertical movement) at <2.0 mg/L of dissolved oxygen but no change in tissue lactate, an indicator of anaerobic metabolism. Combined results of the biotelemetry and laboratory studies demonstrate that a dissolved oxygen content of 2.0 mg/L may be a critical threshold that induces behavioral responses by largemouth bass during the winter. In addition, the use by fish of areas with intermediate levels of dissolved oxygen suggests that there are multiple environmental factors influencing winter behavior.


The biotelemetry study was carried out in Warner Lake, eastern Ontario, Canada. Warner Lake, located entirely within the property of the Queen’s University Biological Station (44°31'N, 76°22'W; Fig. 1), is an 8.3-ha freshwater lake with a naturally self-sustaining population of largemouth bass. There is little flow of water into and out of the lake, resulting in virtually no immigration or emigration. Dissolved oxygen and temperature were sampled at 22 sites throughout Warner Lake on seven different occasions from November 2, 2005, to April 12, 2006, using a submersible dissolved oxygen and temperature probe (model 55, YSI, Yellow Springs, OH; Fig. 1). At each of the 22 sites, dissolved oxygen was measured at 1-m increments starting at the surface, and the number of measurements made at each site varied as a result of depth differences; dissolved oxygen concentrations for all depths were combined to generate a single mean oxygen concentration for each site. Dissolved oxygen sampling dates were chosen to ensure that the lake was sampled during the pre–ice cover period (November 2, 2005; December 1, 2005), during the period of time when ice was present across the entire surface of the lake (January 22, 2006; February 14, 2006; February 24, 2006; March 21, 2006), and during the post–ice cover period (April 12, 2006). Water temperature was varied between 4.0 and 5.5C during periods of ice cover and between 4.0 and 6.5C during non–ice cover periods. For dates when ice was covering the lake, an 18-cm auger was used to drill through the ice to access the water. To quantify the movement of largemouth bass in Warner Lake in relation to dissolved oxygen distributions, a stationary acoustic telemetry array consisting of a code division multiple access (CDMA)-based telemetry system (Lotek MAP_600, Lotek Wireless, Newmarket, Ontario) was used. The system and its accuracy are summarized by Cooke et al. (2005) and Hanson et al. (2007). Briefly, the array consists of 13 hydrophones distributed to allow for submeter position solutions to be calculated during periods when tagged fish were within the footprint of the array (Niezgoda et al. 2002). Position solutions were recorded on flash cards that were regularly collected and downloaded to personal laptop computers for processing and filtering in BioMAP (ver. 2.1; Lotek Wireless), which uses a wavelet analysis to remove uncorrelated noise from time series (HessNielsen and Wickerhauser 1996; Akay and Mello 1997).