Data from: Wood hardness drives nest site selection in woodpeckers of the humid Chaco
Data files
Nov 21, 2024 version files 29.79 KB
-
Exc_input_Chaco.csv
9.45 KB
-
Exc_input_foraging_Chaco.txt
6.44 KB
-
README.md
13.89 KB
Abstract
Avian excavators (woodpeckers and other species) select nest sites based on characteristics of the nest patch, nest tree and substrate. These characteristics could increase foraging opportunities or reduce the risk of predation, but there is also a potentially important role for wood hardness in restricting nest site selection, a role that has been little explored and is expected to vary among species according to their ability to excavate. We examined patterns of nest site selection in eight woodpecker species in the humid Chaco of South America, where the dominant trees have extremely hard wood. We hypothesized that 1) wood hardness is the main factor driving selection of excavation sites, and 2) interspecific variation in body size and foraging behavior (traits frequently related to excavation ability) explain interspecific variation in the wood hardness of excavated nest substrates. From 2016 to 2019 in well-preserved forests of the Argentine Chaco, we compared nesting cavities excavated in wood (cases; n = 42) vs. potential wood substrates without cavities (matched controls) and made 187 focal observations of foraging woodpeckers. Woodpeckers selected nesting substrates with softer heartwood than potential substrates, regardless of any other characteristics of the tree or nest patch. Wood hardness around nest cavities increased with body size and the prevalence of chiseling during foraging, traits that were positively correlated. Woodpeckers often excavated in Prosopis spp. (Algarrobos) but rarely in Schinopsis balansae (Quebracho), a tree with exceptionally hard internal wood, in which cavity adopters frequently nest in non-excavated cavities. Wood hardness is critical to assessing the role of woodpeckers as cavity facilitators, understanding the costs and benefits of excavation, and interpreting excavation patterns across species and forests globally.
README: Wood hardness drives nest site selection in woodpeckers of the humid Chaco
https://doi.org/10.5061/dryad.1jwstqk4q
Description of the data and file structure
We studied excavators, their nests and foraging behaviors during four breeding seasons (August to January) from 2016 to 2019 in Chaco National Park (28.8204°S, 59.6204°W, 14981 ha) and during two breeding seasons (2016 and 2017) in Pampa del Indio Provincial Park (26.2767°S, 59.9702°W, 8366 ha) in the humid Chaco, Chaco province, Argentina.
We conducted nest searches in quebrachales (about 700 hours/year) and gallery forests (about 250 hours/year).
Files and variables
File: Exc_input_Chaco.csv
Description: Case-control study. Measurements of excavated nest cavities (case) by eight woodpecker species and randomly selected unexcavated substrates (control).
Variables
- NID: pair case/control code.
- SpeciesID: code for the scientific name of the bird that excavated the cavity. The first 4 letters of the genus and the first 4 letters of the specific epithet are conjugated.
- Spgraph: code for the scientific name of the bird that excavated the cavity. The first 4 letters of the genus and the first 4 letters of the specific epithet are conjugated. In control the code is "azar".
- Clase: woodpecker or azar
- Nest: case/control code. 1 to case, 0 to control
- plotcover: the mean percent canopy cover from four measurements taken with a densiometer (3% resolution) at 10 m from the central tree in the four cardinal directions. Unit in percent
- cavitycover: percent canopy cover, measurements taken with a densiometer (3% resolution). Unit in percent
- treesp: code for scientific name of the tree species containing the cavity (case) or potential substrate (control). The first 4 letters of the genus and the first 4 letters of the specific epithet are conjugated.
- livetree: if the tree was live o dead.
- treeclass: if the tree was live (A, B) o dead (C, D). A = healthy living tree, B = unhealthy living tree, C = recently dead tree, D = dead tree trunk only.
- dbh: tree diameter at breast height. Unit in centimetres.
- htree: tree height. Unit in metres.
- canopytouching: the percentage of the crown in contact with neighboring trees (measure related to access of arboreal predators to the nest tree). Unit in percent.
- trunktouching: the percentage of the trunk in contact with ground vegetation (lianas or shrubs). Unit in percent.
- substrate: if the substrate was live or dead.
- branchorder: order of the excavated or potential substrate.
- branchdiameter: diameter of the substrate at cavity height. Unit in centimetres.
- height: substrate height. Unit in metres.
- orientation: substrate cardinal orientation.
- fungi: presence of fungi (yes/no).
- groundvisivility: percent visibility at ground height, we filmed or photographed the substrate from three points, all at a horizontal distance of 10 m, one directly in front of the substrate, one 45° to the left and one 45° to the right. For each photo or video, we estimated the percentage of the cavity entrance that was visible, or for potential substrates the percentage of an area of similar size to a cavity entrance. For each height, we took the mean of the three values obtained (from the three directions). Unit in percent.
- cavityvisibility: percent visibility at each of cavity height, we filmed or photographed the substrate from three points, all at a horizontal distance of 10 m, one directly in front of the substrate, one 45° to the left and one 45° to the right. For each photo or video, we estimated the percentage of the cavity entrance that was visible, or for potential substrates the percentage of an area of similar size to a cavity entrance. For each height, we took the mean of the three values obtained (from the three directions). Unit in percent.
- ntrees: the number of trees (individuals with DBH > 10 cm) including the central tree in nest patch (11. 3 m around nest or potential tree).
- dbhtrees: DBH mean (individuals with DBH > 10 cm) including the central tree in nest patch (11. 3 m around nest or potential tree). Unit in centimetres.
- ndeadtrees: the number of dead trees (individuals with DBH > 10 cm) including the central tree in nest patch (11. 3 m around nest or potential tree).
- sillhardness: external wood hardness. We used the method developed by Matsuoka (2000) to measure wood hardness. With a drill and a 9 mm drill bit, we drilled a hole radially into the substrate (towards the center of the trunk or branch). Wood hardness is proportional to the torque required to expand this hole with an increment borer (30 cm long and 5.15 mm external diameter). We measured the torque every 1 cm along the perforation by means of a torque meter attached to the increment borer. As suggested by Matsuoka (2000), we use Newton meters (N∙m) as the unit of torque for statistical analyses. Following Matsuoka (2008), we drilled 5 cm above the nest entrance or, if we could not drill above the cavity (e.g., if the substrate was broken or a branch was in the way), we drilled 10 cm below the bottom of the cavity. For nests, we drilled radially until the entire horizontal depth of the cavity was included. In potential substrates, we drilled radially to the center of the tree.
- bodyhardness: internal wood hardness. We used the method developed by Matsuoka (2000) to measure wood hardness. With a drill and a 9 mm drill bit, we drilled a hole radially into the substrate (towards the center of the trunk or branch). Wood hardness is proportional to the torque required to expand this hole with an increment borer (30 cm long and 5.15 mm external diameter). We measured the torque every 1 cm along the perforation by means of a torque meter attached to the increment borer. As suggested by Matsuoka (2000), we use Newton meters (N∙m) as the unit of torque for statistical analyses. Following Matsuoka (2008), we drilled 5 cm above the nest entrance or, if we could not drill above the cavity (e.g., if the substrate was broken or a branch was in the way), we drilled 10 cm below the bottom of the cavity. For nests, we drilled radially until the entire horizontal depth of the cavity was included. In potential substrates, we drilled radially to the center of the tree.
File: Exc_input_foraging_Chaco.txt
Description: Measurements of excavated nest cavities (case) by eight woodpecker species and data on prevalence of foraging techniques in each woodpecker species, we conducted 187 focal observations on individuals found foraging.
Variables
- NID: pair case/control code.
- SpeciesID: code for the scientific name of the bird that excavated the cavity. The first 4 letters of the genus and the first 4 letters of the specific epithet are conjugated.
- weight: mean body weight. The data for that variable is an average using data from different sources (own data, Dunning 2007, Birds of the World platform 2022: https://birdsoftheworld.org). Unit in grams.
- Nest: case/control code. 1 to case, 0 to control.
- plotcover: the mean percent canopy cover from four measurements taken with a densiometer (3% resolution) at 10 m from the central tree in the four cardinal directions. Unit in percent.
- cavitycover: percent canopy cover, measurements taken with a densiometer (3% resolution). Unit in percent.
- treesp: code for scientific name of the tree species containing the cavity (case) or potential substrate (control). The first 4 letters of the genus and the first 4 letters of the specific epithet are conjugated.
- livetree: if the tree was live o dead.
- treeclass: if the tree was live (A, B) o dead (C, D).A = healthy living tree, B = unhealthy living tree, C = recently dead tree, D = dead tree trunk only.
- dbh: tree diameter at breast height. Unit in centimetres.
- htree: tree height. Unit in metres.
- canopytouching: the percentage of the crown in contact with neighboring trees (measure related to access of arboreal predators to the nest tree). Unit in percent.
- trunktouching: the percentage of the trunk in contact with ground vegetation (lianas or shrubs). Unit in percent.
- substrate: if the substrate was live or dead.
- branchorder: order of the excavated or potential substrate.
- branchdiameter: diameter of the substrate at cavity height. Unit in centimetres.
- height: substrate height. Unit in metres.
- orientation: substrate cardinal orientation.
- fungi: presence of fungi (yes/no).
- groundvisivility: percent visibility at ground height, we filmed or photographed the substrate from three points, all at a horizontal distance of 10 m, one directly in front of the substrate, one 45° to the left and one 45° to the right. For each photo or video, we estimated the percentage of the cavity entrance that was visible, or for potential substrates the percentage of an area of similar size to a cavity entrance. For each height, we took the mean of the three values obtained (from the three directions). Unit in percent.
- cavityvisibility: percent visibility at each of cavity height, we filmed or photographed the substrate from three points, all at a horizontal distance of 10 m, one directly in front of the substrate, one 45° to the left and one 45° to the right. For each photo or video, we estimated the percentage of the cavity entrance that was visible, or for potential substrates the percentage of an area of similar size to a cavity entrance. For each height, we took the mean of the three values obtained (from the three directions). Unit in percent.
- ntrees: the number of trees (individuals with DBH > 10 cm) including the central tree in nest patch (11. 3 m around nest or potential tree)
- dbhtrees: DBH mean (individuals with DBH > 10 cm) including the central tree in nest patch (11. 3 m around nest or potential tree). Unit in centimetres.
- ndeadtrees: the number of dead trees (individuals with DBH > 10 cm) including the central tree in nest patch (11. 3 m around nest or potential tree).
- sillhardness: external wood hardness. We used the method developed by Matsuoka (2000) to measure wood hardness. With a drill and a 9 mm drill bit, we drilled a hole radially into the substrate (towards the center of the trunk or branch). Wood hardness is proportional to the torque required to expand this hole with an increment borer (30 cm long and 5.15 mm external diameter). We measured the torque every 1 cm along the perforation by means of a torque meter attached to the increment borer. As suggested by Matsuoka (2000), we use Newton meters (N∙m) as the unit of torque for statistical analyses. Following Matsuoka (2008), we drilled 5 cm above the nest entrance or, if we could not drill above the cavity (e.g., if the substrate was broken or a branch was in the way), we drilled 10 cm below the bottom of the cavity. For nests, we drilled radially until the entire horizontal depth of the cavity was included. In potential substrates, we drilled radially to the center of the tree.
- bodyhardness: internal wood hardness. We used the method developed by Matsuoka (2000) to measure wood hardness. With a drill and a 9 mm drill bit, we drilled a hole radially into the substrate (towards the center of the trunk or branch). Wood hardness is proportional to the torque required to expand this hole with an increment borer (30 cm long and 5.15 mm external diameter). We measured the torque every 1 cm along the perforation by means of a torque meter attached to the increment borer. As suggested by Matsuoka (2000), we use Newton meters (N∙m) as the unit of torque for statistical analyses. Following Matsuoka (2008), we drilled 5 cm above the nest entrance or, if we could not drill above the cavity (e.g., if the substrate was broken or a branch was in the way), we drilled 10 cm below the bottom of the cavity. For nests, we drilled radially until the entire horizontal depth of the cavity was included. In potential substrates, we drilled radially to the center of the tree.
- Pctpeaking: percent the time, in seconds, that the bird spent in pecking: the bird intermittently (< 4 pecks) taps its beak against the substrate to remove external parts of the substrate (bark, mosses, lichens). In this case, when the head is tilted backwards, the head does not exceed the vertical axis of the body. Unit in percent.
- Pctcinc: percent the time, in seconds, that the bird spent in chiseling: the bird removes bark or layers of wood with blows at oblique angles. When it tilts its head back to make the blows, the head exceeds the vertical axis of the body. Unit in percent.
- Pcthumm: percent the time, in seconds, that the bird spent in hammering: the bird performs a series of sustained pecks (> 4 blows) producing deep holes. In these blows, too, the head exceeds the vertical axis of the body. Unit in percent.
- PctGolpes: percent the time, in seconds, that the bird spent in chiseling, hammering and pecking. Unit in percent.
- Pctprobing: percent the time, in seconds, that the bird spent in probing: the bird inserts the tip of its bill into pre-existing cracks or holes. Unit in percent.
- Pctseek: percent the time, in seconds, that the bird spent in searching/gleaning: the bird searches for and captures prey as it climbs up the trunk or branch. Unit in percent.
- PctNoGolpes: percent the time, in seconds, that the bird spent in probing, scanning, eating and searching/gleaning. Unit in percent.
- Robustas: percent the time, in seconds, that the bird spent in chiseling and hammering. Unit in percent.
- Graciles: percent the time, in seconds, that the bird spent in probing and searching/gleaning. Unit in percent.
Code/software
Microsoft Excel or R Studio.
Methods
We studied excavators and their nests during four breeding seasons (August to January) from 2016 to 2019 in Chaco National Park (28.8204°S, 59.6204°W, 14981 ha) and during two breeding seasons (2016 and 2017) in Pampa del Indio Provincial Park (26.2767°S, 59.9702°W, 8366 ha) in the humid Chaco, Chaco province, Argentina. Our data set includes 8 cavity-excavating bird species: Dryocopus schulzii (Black-bodied Woodpecker, 200 g), Picumnus cirratus (White-barred Piculet, 10 g), Melanerpes cactorum (White-fronted Woodpecker, 40 g), Celeus lugubris (Pale-crested Woodpecker, 125 g), Dryobates mixtus (Checkered Woodpecker, 30 g), Dryobates passerinus (Little Woodpecker, 37 g), Colaptes melanochloros (Green-barred Woodpecker, 135 g), and Campephilus leucopogon (Cream-backed Woodpecker, 230 g).
To obtain sufficient nest observations for analysis, we employed a case-control study design (Keating and Cherry 2004), comparing each cavity excavated in wood (nest) with a nearby potential substrate (trunk or branch that looked similar but was not excavated). To locate nests, we inspected areas where we heard continuous tapping or observed recurrent flights. To find a potential substrate (paired control for each nest cavity excavated in the corresponding breeding season), we first determined the decay class of the nest tree (live unhealthy vs. dead). Second, we searched for a potential excavation substrate in a random direction and at a minimum distance of 25 m from the nest tree so as not to overlap potential sites with nest sites (see following paragraph). Our potential (unused) tree was the first tree that met the following conditions: 1) same decay class as the paired nest tree, 2) same species as the paired nest tree, and 3) no previous excavations.
For each excavated cavity and each potential substrate, after breeding was complete, we used a 7 m ladder or a rope with harness equipment to take measurements of fifteen characteristics at the scale of the patch, tree, and excavation substrate. As a nest patch we considered a circular area of 11.3 m radius with the nest tree or potential tree in the center (central tree). The minimum distance of 25 m between nest tree and potential tree ensured that the patches did not overlap. For each nest patch we measured (1) the mean percent canopy cover from four measurements taken with a densiometer (3% resolution) at 10 m from the central tree in the four cardinal directions. We recorded (2) the number of trees (individuals with DBH > 10 cm) including the central tree. For each tree in the patch with DBH > 10 cm, we used a tape (0.1 cm) to measure (3) DBH, determined (4) the decay class (dead or live), and identified (5) the species. At the level of the central tree (nest or potential), we recorded (6) DBH, and (7) the tree species. We estimated (8) the percentage of the crown in contact with neighboring trees (measure related to access of arboreal predators to the nest tree). At substrate level (excavated or potential) we measured (9) the diameter of the substrate at cavity height (DCH), (10) internal and (11) external wood hardness (see below), and (12) height above ground. We used a densiometer at substrate height to estimate (13) percent cover. We determined percent visibility of the substrate from two heights: (14) at ground level (0.5 m) and (15) at substrate height.
We used the method developed by Matsuoka (2000) to measure wood hardness. With a drill and a 9 mm drill bit, we drilled a hole radially into the substrate (towards the center of the trunk or branch). Wood hardness is proportional to the torque required to expand this hole with an increment borer (30 cm long and 5.15 mm external diameter). We measured the torque every 1 cm along the perforation by means of a torque meter attached to the increment borer. As suggested by Matsuoka (2000), we use Newton meters (N∙m) as the unit of torque for statistical analyses. Following Matsuoka (2008), we drilled 5 cm above the nest entrance or, if we could not drill above the cavity (e.g., if the substrate was broken or a branch was in the way), we drilled 10 cm below the bottom of the cavity. For nests, we drilled radially until the entire horizontal depth of the cavity was included. In potential substrates, we drilled radially to the center of the tree. We divided the perforation into two sections: external wood (corresponding to the cavity wall) and internal wood (corresponding to the nest chamber). For potential substrates, we considered the first 3 cm to be external wood and the following 12 cm to be internal wood. For each substrate, we took the mean of torque measurements (every 1 cm) within the corresponding section of the perforation to obtain the final values for external and internal wood hardness.
We found foraging adults opportunistically as we searched for nests. For each focal observation, the first author used binoculars to observe the adult foraging, for 60 seconds beginning with the first visual contact (n = 187 focal observations). We recorded the time, in seconds, that the bird spent in each of the following foraging activities adapted from Remsen and Robinson (1990) and Fernández et al. (2020). (1) Chiseling: the bird removes bark or layers of wood with blows at oblique angles. When it tilts its head back to make the blows, the head exceeds the vertical axis of the body. (2) Hammering: the bird performs a series of sustained pecks (> 4 blows) producing deep holes. In these blows, too, the head exceeds the vertical axis of the body. (3) Pecking: the bird intermittently (< 4 pecks) taps its beak against the substrate to remove external parts of the substrate (bark, mosses, lichens). In this case, when the head is tilted backwards, the head does not exceed the vertical axis of the body. (4) Probing: the bird inserts the tip of its bill into pre-existing cracks or holes. (5) Searching/gleaning: the bird searches for and captures prey as it climbs up the trunk or branch. (6) Scanning: the bird looks around. (7) Eating: the bird consumes prey. We consider chiseling and hammering as foraging techniques related to robust anatomies and searching/gleaning as a foraging technique associated with less robust anatomies (Kirby 1980, Bock 1999, Donatelli et al. 2014). To include variability among individuals and to promote independence among focal observations, we made no more than three focal observations of the same species within the same 30 ha area. The eight woodpecker species we studied represent a diversity of foraging guilds (Fig. 1).
We report mean ± standard deviation, except where indicated. All statistical analyses were performed in R software, version 4.3.3 (R Core Team 2024). We analyzed the correlation between independent variables and excluded variables that were correlated (Pearson's r > 0.7, Dormann et al. 2013). To improve the interpretation of interaction terms, we standardized all continuous predictor variables to each have a mean of zero.