Territory mapping and densities

Early in the breeding season, from April to early June, we colour-banded male wrens by catching them in mist nets with targeted playback of their song to aid in individual identification of territorial behaviour. To determine territory size we used spot-mapping, a well-established method for determining habitat use of breeding birds [41,44,45]. We initiated spot-mapping one hour after sunrise each day and continued for 1.5 hours. Observers moved through the plot along the geolocated points at a consistent walking pace, recording the locations of singing, countersinging, calling, and visual observations of male wrens against known plot locations. Each stream was visited eight times each year, from late April until July, in rotating order (one site per day) and for each visit, we began spot mapping at either a different corner of the plot, or by moving in a different direction to ensure adequate coverage throughout the morning chorus.

Territories were delineated in ArcGIS (Version 10.1) by geo-referencing each song or visual observation against known site points and then overlaying each visit to create an overall pattern, or cluster, of territorial observations throughout the season (Fig 1). Territory boundaries were mapped according to countersinging locations between male wrens. Countersinging occurs when birds sing in response to their neighbour’s song by matching or overlaying songs [46]. Pacific wrens will often countersing at the edges of defended habitat, allowing us to mark the outer borders of territories.

Representation of the territories of male Pacific wrens at different streams with different numbers of spawning salmon (A-D). Each polygon represents an individual wren territory along a stream: (a) Ripley with no salmon, (b) Fancy Right with ~2,500 kg, (c) Fannie Left with 32,000 kg, and (d) Clatse with 47,000 kg average annual spawning pink and chum salmon biomass. Blue lines represent the respective streams and the salmon drawing represents amount of salmon biomass.

We measured territory size using 95% minimum convex polygons, with the adehabitat package in R [47], using 95% of all singing observations for each wren. Territories that had more than half of their observations outside the surveyed plot area were excluded from the study as their size may have been underestimated. Similarly, male wrens that were detected in fewer than three survey visits or that had fewer than three total observations throughout the breeding season were excluded from the analysis as we categorized them as floaters unable to establish a breeding territory.

To calculate absolute male wren densities at each stream we determined the proportion of each male’s territory within the surveyed stream-site, summed the proportions of the territories in the surveyed area, and divided the sum by the total surveyed stream-site area. This gave us male wren densities as number of territorial males per hectare.