The wash zone and habitat use among three benthic fish species in stratified lakes

Mixing processes in lakes are important in determining sedimentation zones and in setting the so-called “wash zone”, the area of lake bottom in contact with an oscillating thermocline during wind driven internal seiche events. The wash zone also aligns with a sharp change in sediment roughness and hardness. Taken together these rapid changes in temperature and sediment indicate that the wash zone is a distinctive ecotone in stratified lakes.  Depth stratified randomized netting was used to develop count-based habitat use models for three common benthic fish species as a function of depth or temperature covariates.  Using data from two lakes with quite different wash zone depths, we show the wash zone to describe fish habitat for two of three benthic fish species by utilizing the top 50% of estimated fish abundance as an indicator of habitat use. White sucker (Catostomus commersoni) habitat use was fully within the boundaries of the wash zone. Lake whitefish (Coregonus clupeaformis) habitat was adjacent and within the wash zone. Longnose sucker (C. catostomus) habitat use was in the deep areas of lakes dominated by sediment focusing and did not overlap white sucker. Lake whitefish habitat use overlapped both catostomids, but peak abundance of both lake whitefish and white sucker overlapped pointing to potential interactions between these species. Smaller lakes have less vigorous mixing processes and a narrower wash zone, so with a decline in lake size the likely area of the wash zone as habitat for benthic feeding fish would become smaller.

Methods
The study focused on Lakes Manitou (46.015; -78.992) and Kioshkokwi (46.078; -78.885) in the northwest corner of Algonquin Park, Ontario, Canada (see Fig. S1 in Supplement). Both are large oligotrophic lakes (Manitou, secchi depth = 2.75 m, TP = 6.6 μg/L; Kioshkokwi, secchi depth = 3.7 m, TP = 7.5 μg/L) with similar morphometries (Manitou, SA = 1,382 ha, max. depth = 38.4 m,  depth = 13.6m; Kioshkokwi, SA = 1,070 ha, max. depth = 47.4 m, depth = 13.9 m). Manitou drains to Kioshkokwi.

We evaluate habitat use hypotheses based on mapping of habitat use models stemming from catch data following depth-stratified randomized netting (2 hour soak time per net) in each lake. The netting details are as follows: 

Gill nets were 64x2 m in total length and height, respectively, composed of 8 single mesh panels (8 m each) sewn together. The mesh series included 2.25 inches (57mm), 2.5” (64mm), 2.75” (70mm), 3.0” (76mm), 3.5” (89mm), 4.0” (102mm), 4.5” (114mm), and 5.0” (127mm) stretch mesh with monofilament diameter for each mesh being 0.23, 0.23, 0.23, 0.28, 0.28, 0.33, 0.33, and 0.33 mm, respectively.

Net sites were depth stratified and randomly allocated within strata proportional to stratum surface area. Depth strata were 2-10 m, 10-20, 20-30, 30-40 and 40-60 m. The total number of net sets (N=65 for Kioskokwi; N=78 for Manitou) was determined by the formula: total net sets = 0.0184*(lake surface area) + 24.  Nets were set for 2 hours during the day (08:00-16:00), and net sites were sampled once in each survey. Lake surveys were conducted in the summer of 2015 (Lake Kioshkokwi July 21-23; Lake Manitou Aug 11-13).

Habitat use models were based on catch counts for the three species using either count regression or hurdle models incorporating a negative binomial assumption in data distribution.