Data from: Asymmetrical predation intensity produces divergent antipredator behaviors in primary and secondary prey
Data files
Aug 02, 2024 version files 213.21 KB
-
All_GUD_Data.csv
26 KB
-
Behavioral_Video_Data.csv
30.91 KB
-
Harvest_Rate_Curve_Data.csv
14.88 KB
-
Harvest_Rate_Curve_Peromyscus_Data.csv
5.50 KB
-
Peromyscus_Behavioral_Data.csv
15.30 KB
-
README.md
48.67 KB
-
Visitation_Data.csv
35.66 KB
Abstract
It is widely recognized that predators can influence prey through both direct consumption and by inducing costly antipredator behaviors, the latter of which can produce nonconsumptive effects that cascade through trophic systems. Yet, determining how particular prey manage risk in natural settings remains challenging as empirical studies disproportionately focus on single predator-prey dyads. Here, we contrast foraging strategies within the context of a primary and secondary prey to explore how antipredator behaviors emerge as a product of predation intensity as well as the setting in which an encounter takes place. We studied the effects of spotted owls (Strix occidentalis) on two species experiencing asymmetrical risk: dusky-footed woodrats (Neotoma fuscipes; primary prey) and deer mice (Peromyscus spp.; alternative prey). Woodrats are most abundant within young forest, but predominantly captured by owls foraging within mature forest; in contrast, deer mice occur in high densities across forest types and seral stages and are consumed at lower per-capita rates overall. We deployed experimental foraging patches within areas of high and low spotted owl activity, created artificial risky and safe refuge treatments, and monitored behavior throughout the entirety of prey foraging bouts. Woodrats were more vigilant and foraged less within mature forest and at riskier patches, although the effect of refuge treatment was contingent upon forest type. In contrast, deer mice only demonstrated consistent behavioral responses to riskier refuge treatments; forest type had little effect on perceived risk or the relative importance of refuge treatment. Thus, habitat can interact with predator activity to structure antipredator responses differently for primary versus secondary prey. Our findings show that asymmetrical predation can modulate both the magnitude of perceived risk and the strategies used to manage it, thus highlighting an important and understudied contingency in risk effects research. Evaluating the direct and indirect effects of predation through the paradigm of primary and secondary prey may improve our understanding of how nonconsumptive effects can extend to population- and community-level responses.
https://doi.org/10.5061/dryad.qjq2bvqqp
Description of the data and file structure
These datasets were collected in the central Sierra Nevada, California as part of an experimental study of foraging behaviors. The data consist of response variables from a combination of video footage, values of food in the beginning and remaining in a patch after a foraging night, and predictor variables including age, sex, weight, and body condition for woodrats from direct captures, and habitat data collected from both vegetative surveys on the ground and variables derived from ArcGIS. They have been processed in BORIS, R, and ArcGIS.
This dataset consists of 6 files for datasets and 3 files for R code. They are listed below, followed by individual tables for each .csv file that better describes the individual data files and variables within.
1. File List:
A) Behavioral_Video_Data.csv
B) All_GUD_Data.csv
C) Harvest_Rate_Curve_Data.csv
D) Harvest_Rate_Curve_Peromyscus_Data.csv
E) Peromyscus_Behavioral_Data.csv
F) Visitation_Data.csv
G) Loading_And_Preparing_Data.r
H) Top_Models_And_Figure_Construction.r
I) Harvest_Rate_Curves.r
Descriptions of variables within datafiles:
A) Behavioral_Video_Data.csv
| Variable Name | Variable description |
|---|---|
| Observations id | This is a unique identifier for this night of data, reflecting the camera name and date. |
| Tag | A unique identifying number for the specific indvidual and/or foraging station (pair of foraging trays). Used as a random effect in model construction. |
| Target_visit | Binary variable reflecting whether a patch was visited (1) or not (0) on a given experimental night. For this dataset all values should be 'yes'. |
| Nontarget_visit | Presence/absence of additional woodrats beyond the target individual (1, 0; woodrat only) |
| Peromyscus_visit | Presence/absence of deer mice (1, 0; woodrat only) |
| Any_nontarget_forager | Presence/absence of deer mice or a nontarget woodrat (1, 0; woodrat only) |
| Target_wr_total_tray_time | Cumulative patch residence time (PRT; s) by dusky-footed woodrats during an experimental night (Response Variable) |
| Nontarget_wr_total_tray_time | Cumulative time spent in patch (s) in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
| Nontarget_wr_and_peromyscus_duration | Cumulative time spent in patch (s) in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
| Total_wr_tray_time | Cumulative time spent in patch (s) in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
| Proportion_of_time_vigilant | Percentage of nightly patch residence time allocated to vigilance behavior |
| Proportion_of_time_vigilant_indiv_over_90sec | Percentage of nightly patch residence time allocated to vigilance behavior - all nights where the total residence time was less than 90 seconds were censored from the dataset (NA values) (Response Variable) |
| Proportion_of_time_vigilant_indiv_over_120sec | Percentage of nightly patch residence time allocated to vigilance behavior - all nights where the total residence time was less than 120 seconds were censored from the dataset (NA values) |
| Average_visit_duration | The average duration of a visit (s) to a foraging patch during a given night. Derived by dividing the value for PRT by the # of visits on a given night. |
| Visits | Number of nightly patch visits by dusky-footed woodrats (Response Variable) |
| Nontarget_wr_visits | Number of visits to a patch in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
| Peromyscus_visits | Number of visits to a patch in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
| Nontarget_wr_plus_peromyscus_visits | Number of visits to a patch in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
| Habitat | Forest structure designation defined by canopy cover and quadratic mean diameter of dominant trees (mature, young) |
| Treatment | Presence/absence of artificial cover structure at foraging patch (risky, refuge) |
| Residuals | Residuals from the regression of body mass (unitless) against hind foot length; positive values represent individuals in better body condition while negative values represent poor body condition |
| Sex | Sex of target individual (female, male) |
| Age_simple | Age class of target individual (adult, subadult) |
| Understory | Proportion of ground obscured by vegetation <2m height within 12.5m |
| Basal_area | Total basal area (m2) of all live and dead standing trees <2m height within 12.5m |
| Cover | Proportion of sky obscured by canopy >2m height within 12.5m |
B) All_GUD_Data.csv
| Variable Name | Variable description |
|---|---|
| Observations id | This is a unique identifier for this night of data, reflecting the camera name and date. |
| Tag | A unique identifying number for the specific indvidual and/or foraging station (pair of foraging trays). Used as a random effect in model construction. |
| Target_wr_GUD_cleaned | Giving-up Density value (g)for a foraging patch visited by woodrats during an experimental night. Neglible to no foraging by deer mice. NA values mean that this night was not included in the dataset because it was either used in the deer mice dataset or was excluded for other reasons (e.g., weather, nontarget visitation). |
| Peromyscus_GUD_cleaned | Giving-up Density value (g) for a foraging patch visited by deer mice during an experimental night. Neglible to no foraging by woodrats. NA values mean that this night was not included in the dataset because it was either used in the woodrat dataset or was excluded for other reasons (e.g., weather, nontarget visitation). |
| Nontarget_woodrat_visitation | Presence/absence of additional woodrats beyond the target individual (yes, no; woodrat only) |
| Habitat | Forest structure designation defined by canopy cover and quadratic mean diameter of dominant trees (mature, young) |
| Treatment | Presence/absence of artificial cover structure at foraging patch (risky, refuge) |
| Residuals | Residuals from the regression of body mass against hind foot length; positive values represent individuals in better body condition while negative values represent poor body condition |
| Sex | Sex of target individual (female, male) |
| Age_simple | Age class of target individual (adult, subadult) |
| Understory | Proportion of ground obscured by vegetation <2m height within 12.5m |
| Cover | Proportion of sky obscured by canopy >2m height within 12.5m |
| Basal_area | Total basal area (m2) of all live and dead standing trees <2m height within 12.5m |
| Nontarget_wr_total_tray_time | Cumulative time spent in patch (s) in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
| Total_wr_tray_time | Cumulative time spent in patch (s) in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
| Visits | Number of visits to a patch in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice. This variable reflects the # of visits by the target (collared) woodrat, and is combined with the 'Nontarget_wr_visits' variable in the R code to create the total number of woodrat visits |
| Nontarget_wr_visits | Number of visits to a patch in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
| Peromyscus_visits | Number of visits to a patch in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
| Peromyscus_foraging_duration | Cumulative time spent in patch (s) in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
C) Harvest_Rate_Curve_Data.csv
| Variable Name | Variable description |
|---|---|
| Observations id | This is a unique identifier for this night of data, reflecting the camera name and date. |
| Tag | A unique identifying number for the specific indvidual and/or foraging station (pair of foraging trays). Used as a random effect in model construction. |
| Handling_time_wr | (h; time [g s-1] required by the forager to process and consume food. A constant value estimated from direct observations of foraging video data. |
| Total_wr_prt | Cumulative patch residence time (PRT) by foraging woodrats during an experimental night (s) |
| Initial_gud | Value reflecting the initial amount of food in the foraging patch. A constant at 20g. |
| Final_gud | Value reflecting the Giving-Up Density (GUD), or final amount of food (g) in the foraging patch after an experimental night. |
| Nontarget_woodrat_visitation | Presence/absence of additional woodrats beyond the target individual (1, 0; woodrat only) |
| Habitat | Forest structure designation defined by canopy cover and quadratic mean diameter of dominant trees (mature, young) |
| Treatment | Presence/absence of artificial cover structure at foraging patch (risky, refuge) |
| Habitat_by_treatment | A new variable representing the combination of habitat (forest type) and treatment (refuge treatment) variables that reflects the background and immediate risk that may be perceived in an area. Has four forms (Mature-Risky, Mature-Safe, Young-Risky, Young-Safe). |
| Residuals | Residuals from the regression of body mass against hind foot length; positive values represent individuals in better body condition while negative values represent poor body condition |
| Sex | Sex of target individual (female, male) |
| Age_simple | Age class of target individual (adult, subadult) |
| Understory | Proportion of ground obscured by vegetation <2m height within 12.5m |
| Cover | Proportion of sky obscured by canopy >2m height within 12.5m |
| Basal_area | Total basal area (m2) of all live and dead standing trees <2m height within 12.5m |
| Visits | Number of nightly patch visits by dusky-footed woodrats |
| Nontarget_wr_visits | Number of visits to a patch by nontarget (non-collared) woodrats in a given experimental night |
D) Harvest_Rate_Curve_Peromyscus_Data.csv
| Variable Name | Variable description |
|---|---|
| Observations id | This is a unique identifier for this night of data, reflecting the camera name and date. |
| Tag | A unique identifying number for the specific indvidual and/or foraging station (pair of foraging trays). Used as a random effect in model construction. |
| Handling_time_peromyscus | (h; time [g s-1] required by the forager to process and consume food. A constant value estimated from direct observations of foraging video data. |
| Peromyscus_prt | Cumulative patch residence time (PRT) by deer mice during an experimental night (g) |
| Initial_gud | Value reflecting the initial amount of food in the foraging patch. A constant at 20g. |
| Final_gud | Value reflecting the Giving-Up Density (GUD), or final amount of food (g) in the foraging patch after an experimental night. |
| Peromyscus_visitation | Presence/absence of deer mice (1, 0). Should always be 1 for this dataset |
| Habitat | Forest structure designation defined by canopy cover and quadratic mean diameter of dominant trees (mature, young) |
| Treatment | Presence/absence of artificial cover structure at foraging patch (risky, refuge) |
| Habitat_by_treatment | A new variable representing the combination of habitat (forest type) and treatment (refuge treatment) variables that reflects the background and immediate risk that may be perceived in an area. Has four forms (Mature-Risky, Mature-Safe, Young-Risky, Young-Safe). |
| Understory | Proportion of ground obscured by vegetation <2m height within 12.5m |
| Cover | Proportion of sky obscured by canopy >2m height within 12.5m |
| Basal_area | Total basal area (m2) of all live and dead standing trees <2m height within 12.5m |
| Peromyscus_visits | Number of visits to a patch in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
E) Peromyscus_Behavioral_Data.csv
| Variable Name | Variable description |
|---|---|
| Observations id | This is a unique identifier for this night of data, reflecting the camera name and date. |
| Tag | A unique identifying number for the specific indvidual and/or foraging station (pair of foraging trays). Used as a random effect in model construction. |
| Any_woodrat_visitation | Presence/absence of any woodrat during an experimental night (1, 0; deer mice only) |
| Peromyscus_visitation | Presence/absence of deer mice (1, 0; woodrat only) |
| Any_nontarget_forager | Presence/absence of deer mice or a nontarget woodrat (1, 0; woodrat only) |
| Habitat | Forest structure designation defined by canopy cover and quadratic mean diameter of dominant trees (mature, young) |
| Treatment | Presence/absence of artificial cover structure at foraging patch (risky, refuge) |
| Understory | Proportion of ground obscured by vegetation <2m height within 12.5m |
| Cover | Proportion of sky obscured by canopy >2m height within 12.5m |
| Basal_area | Total basal area (m2) of all live and dead standing trees <2m height within 12.5m |
| Total_wr_tray_time | Cumulative time spent in patch (s) in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
| Average_visit_duration | The average duration of a visit (s) to a foraging patch (by deer mice) during a given night. Derived by dividing the value for PRT by the # of visits on a given night. |
| Visits | Number of visits to a patch in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice. This variable reflects the # of visits by the target (collared) woodrat, and is combined with the 'Nontarget_wr_visits' variable in the R code to create the total number of woodrat visits as a predictor variable for deer mice (A) |
| Nontarget_wr_visits | Number of visits to a patch in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
| Peromyscus_visits | Number of nightly patch visits by deer mice (Response Variable) |
| Nontarget_wr_plus_peromyscus_visits | Number of visits to a patch in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
| Peromyscus_foraging_duration | Cumulative patch residence time (PRT; s) by deer mice during an experimental night (Response Variable) |
| Total_wr_visits | Number of visits to a patch in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
F) Visitation_Data.csv
| Variable Name | Variable description |
|---|---|
| Observations id | This is a unique identifier for this night of data, reflecting the camera name and date. |
| Tag | A unique identifying number for the specific indvidual and/or foraging station (pair of foraging trays). Used as a random effect in model construction. |
| Target_woodrat_visitation | Binary variable reflecting whether a patch was visited (0) or not (1) on a given experimental night. |
| Nontarget_woodrat_visitation | Presence/absence of additional woodrats beyond the target individual (1, 0; woodrat only) |
| Any_woodrat_visitation | Presence/absence of any woodrat during an experimental night (1, 0; deer mice only) |
| Peromyscus_visitation | Presence/absence of deer mice (1, 0; woodrat only) |
| Any_nontarget_forager | Presence/absence of deer mice or a nontarget woodrat (1, 0; woodrat only) |
| Habitat | Forest structure designation defined by canopy cover and quadratic mean diameter of dominant trees (mature, young) |
| Treatment | Presence/absence of artificial cover structure at foraging patch (risky, refuge) |
| Residuals | Residuals from the regression of body mass against hind foot length (unitless); positive values represent individuals in better body condition while negative values represent poor body condition |
| Sex | Sex of target individual (female, male) |
| Age_simple | Age class of target individual (adult, subadult) |
| Understory | Proportion of ground obscured by vegetation <2m height within 12.5m |
| Cover | Proportion of sky obscured by canopy >2m height within 12.5m |
| Basal_area | Total basal area (m2) of all live and dead standing trees <2m height within 12.5m |
| Target_wr_total_tray_time | Cumulative patch residence time (PRT) by dusky-footed woodrats during an experimental night |
| Nontarget_wr_total_tray_time | Cumulative time spent in patch (s) in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
| Nontarget_wr_and_peromyscus_duration | Cumulative time spent in patch (s) in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
| Total_wr_tray_time | Cumulative time spent in patch (s) in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
| Visits | Number of visits to a patch in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice. This variable reflects the # of visits by the target (collared) woodrat, and is combined with the 'Nontarget_wr_visits' variable in the R code to create the total number of woodrat visits as a predictor variable for deer mice (A) |
| Nontarget_wr_visits | Number of visits to a patch in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
| Peromyscus_visits | Number of visits to a patch in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
| Nontarget_wr_plus_peromyscus_visits | Number of visits to a patch in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
| Peromyscus_foraging_duration | Cumulative time spent in patch (s) in one of four forms: A) all woodrats, B) nontarget woodrats, C) deer mice, or D) nontarget woodrats and deer mice |
2. Relationship between files:
Files A, B, C, D, E, and F represent datasets that are all used at various points within the R code in files G, H, and I. File G represents the R code that loads in all of the relevant data and packages, and formats, renames, and scales the data - as well as checks for correlated variables - before running file H, which is used in all analyses except for the harvest rate curve section and produces the values from the top models for each stage of the foraging process, as well as figures 2, 3, and 4, as well as figure S1, in the text. Finally, file I is standalone with its own data, and is used to produce values for Quitting Harvest Rates and Attack rates, and to graph the harvest rate curves in figure 5.
3. Additional related data collected that was not included in the current data package:
Figure 1 in the main text is a map of the study area produced in ArcGIS. This has no relevant data for Dryad, however.
4. Are there multiple versions of the dataset?
No
Sharing/Access information
Data was derived from the following sources:
- Video footage (processed in BORIS)
- Values of food (g) in the beginning and remaining in a patch after a foraging night
- Predictor variables including age, sex, weight, and body condition for woodrats from direct captures
- Habitat data from vegetative surveys
- Forest type classifications from ArcGIS
Code/Software
Microsoft Excel and R are needed to open these data files. For R, there are a number of packages utilized in the analysis, but these are all publicly available and noted within the R code itself.
This dataset was collected during behavioral experiments of foraging behavior in the field in the Sierra Nevada mountains in summer 2021. The video data was processed using the BORIS software, while the rest of the analysis was done in R and data was curated in Excel.