Lemming abundance and demography are monitored using three methods:
Capture by snap trapping is done using Museum Special traps at three sites during the summer from mid-July to early August since 1994. At least 500 trap-nights are cumulated annually at each site. From 1994 to 2006, there were 50 trapping stations with one trap per station and trapping sessions lasted for 10 or 11 consecutive days. Trapping stations were spaced out every 10 m along two or more parallel trapping lines separated by 100 m (Gruyer et al. 2008). The sampling protocol changed from 2007 onward. There are now 80 trapping stations (except in 2007 and 2008 when there were only 68 stations) with three traps per station and trapping sessions last for 3 or 4 consecutive days. Each trapping stations are spaced out every 15 m along two or more parallel trapping lines which are separated by 100 m (Fauteux et al. 2018). All captured animals are identified at the species level.
Capture-mark-recapture by live trapping animals is conducted using Longworth traps on three grids, one of 7 ha (96 to 100 traps depending of the year) and two of 11 ha (144 traps; from 2004 to 2005, only 7 ha and 100 traps)), in wetland (one grid) and mesic (two grids) habitats since 2004. Traps are positioned in a Cartesian plane and spaced out every 30 m. Traps are checked every 12 hrs during trapping sessions that last three consecutive days (4 or 5 days in 2004 to 2007) and which are repeated three times (four times in 2005 to 2007) during the summer from mid-June to mid-August. Since 2015, three grids of 7 ha (96 to 100 traps/grid) were added in mixed habitat (wetland and mesic). Those three grids are monitored for only one trapping session in mi-July. The species, age (juvenile or adult, based on body mass), sex and reproductive condition of captured animals is determined (for more details, see Gruyer et al. 2010, Bilodeau et al. 2013 and Fauteux et al. 2018).
One of the live-trapping grids in mesic habitat (the one of 7 ha) was subjected to various experimental manipulations. From 2007 to 2011, snow fences were used to enhance snow depth to determine its effect on lemming abundance and distribution. Snow fences (1.3 m high) were erected in six rows of fencing 270 m-long each and perpendicular to the prevailing wind (spacing varied between 30 and 50 m). Since 2013, this grid is used for a predator control experiment. All terrestrial and avian predators (except ermine) are excluded by a 1.3 to 2-m high fence made of chicken wire surrounding the grid and covered by criss-crossing fishing line spaced every 50 cm on top.
Lemming winter nests are surveyed after snow-melt in early July using the line transect method since 2007. Each transect is permanent (since 2009), is 500 m long and is randomly located within each habitat. A total of 60 line transects were surveyed in wetlands, mesic tundra and streams in mesic tundra (20 transects/habitat) between 2009 and 2014. In 2007 and 2008, the number of transects were 75 (25 per habitat) and 30 (10 per habitat), respectively. Between 2015 and 2017, 66 line transects were added to increase the spatial coverage of the monitoring for a total of 126 permanent line transects (between 5 and 20 transects/habitat/sector). Between 2015 and 2019, several line transects were added to increase the spatial coverage of the monitoring for a total of 139 permanent line transects (between 6 and 20 transects/habitat/sector). The exact position of each nest found along the transect is recorded. In addition, each spring, all winter nests are systematically counted on three live trapping grids by walking parallel lines 10 m apart. Winter nests found opportunistically are also collected in years of low lemming abundance. All nests are dissected to determine the lemming species using it, reproductive activity and signs of predation (for more details, see Duchesne et al. 2011a, b and Fauteux et al. 2018).
