https://doi.org/10.5061/dryad.31zcrjdxg

Description of the data and file structure

This dataset was collected to investigate niche partitioning among carnivores in the Spiti Valley, northern India, under increasing human pressures. The study focused on four species: snow leopards (Panthera uncia), Himalayan wolves (Canis lupus chanco), red foxes (Vulpes vulpes), and free-ranging dogs (Canis lupus familiaris).

The data includes results from Camera Trap Surveys used to monitor species occupancy, spatio-temporal activity patterns, and interactions among carnivores.

Files and variables

File: WLB-2025-01494.R2.zip

Description: It contains the following three folders:

1. occupancy

The folder contains data and analysis scripts related to the detection and occupancy patterns of carnivore species in the Spiti Valley.

Dataset Files

1. Detection Histories: dog (DH_dog.csv), red fox (DH_redfox.csv), snow leopard (DH_SL.csv) & Himalayan wolf (DH_wolf.csv).
   • These files contain a binary dataset (1 = species detected, 0 = species not detected) collected from camera trap surveys.
   • Columns (o1, o2, o3, o4...) represent a survey occasion, where one occasion corresponds to a one-week sampling period at a given camera station.
   • Rows indicate observations from individual survey locations/camera traps.
2. Scaled Covariates (Scaled_covs.csv): 
   • Includes environmental and anthropogenic covariates (e.g., elevation, prey availability, distance from settlements) scaled for analysis. These covariates are used to model occupancy probabilities.
   • All covariates were standardized (z-scores; mean = 0, SD = 1) prior to modeling.
   • Covariates were calculated as mean values for the grid cell (1 × 1 km) in which each camera trap was located. Each row, therefore, represents a single grid cell containing one camera trap station. Correlated covariates were not used in the final analysis.

Analysis Codes: The folder includes three R scripts used for multi-species occupancy modeling and visualization of results. All scripts are written in R and require the following packages: unmarked, AICcmodavg, ggplot2, MuMIn, and dplyr

• marginal_occupancy.txt – Estimates marginal occupancy probabilities (ψ) for each species independently using site-level covariates, and computes detection probabilities (p) across all survey sites.

• conditional_occupancy.txt – Performs conditional occupancy analyses that account for species co-occurrence, estimating how the occupancy of one species varies with the presence or absence of another (e.g., fox given dog presence).

• marginal_graphs.txt – Generates graphical outputs of the marginal occupancy models, visualizing predicted occupancy probabilities across key environmental gradients such as elevation or distance to settlements.

  2. spatial overlap

This folder contains data and scripts related to the spatial interactions among carnivore species in the Spiti Valley, with a focus on co-occurrence patterns and the relative abundance index (RAI).

Dataset Files

1. Co-occurrence File (All_Co-occurrence.csv): Contains binary data (1 = species detected, 0 = species not detected) indicating the presence of species at different sampling locations/camera traps.
   • Columns:
     • species: Species names (e.g., Canis lupus, Panthera uncia, Feral dog, Vulpes vulpes).
     • A_01, A_02, ...: Camera trap IDs.
   • Rows: Detection results for each species at specific locations.
2. Relative Abundance Index (RAI) File (All_Pianka_RAI.csv): Contains the Relative Abundance Index (RAI) values for each carnivore species across different sampling locations/camera traps. The RAI represents the relative abundance of species based on detection frequency and was calculated as the number of independent detections per 100 trap-nights for each camera location.
   • Columns: Species names (e.g., Canis lupus, Panthera uncia, Feral dog, Vulpes vulpes).
   • Rows: RAI values calculated for each location

Analysis Codes:  Includes R script (spatial_overlap_codes.txt) for assessing spatial overlap and co-occurrence among carnivore species.

• Pianka’s Niche Overlap Index was used (via the spaa package) to measure pairwise overlap in habitat use among species based on RAI values, including bootstrapped estimates of overlap stability.
• Species Co-occurrence Analysis (via the cooccur package) was performed to evaluate non-random patterns of spatial association between species using binary detection data, providing effect sizes, probability tables, and summaries of significant pairwise associations.

3. temporal interactions

The folder contains data and scripts to analyze the temporal activity patterns of carnivore species in the Spiti Valley. The dataset helps assess temporal overlap between species by examining their activity times throughout the 24-hour cycle.

Dataset Files:

1. Camera trap data (CameraTrapProject_CT_data_for_analysis_MASTER.csv): **Contains metadata for all camera trap stations deployed during the survey. Includes information on setup and retrieval dates, session identifiers, and a scaled measure of human influence at each station. This dataset defines the temporal range and operational details for each camera trap, used to align species detections with sampling effort and assess potential human influence.

   • Session: Sampling period (e.g., Summer2022).
   • Site: Camera trap ID.
   • Date_setup / Date_retr: Dates when each camera trap was installed and retrieved.
   • Problem1_from / Problem1_to: Indicates inactive periods, if any, during the sampling session.
   • ghm: Scaled Human Footprint Index (z-score), representing relative anthropogenic disturbance at each camera location. 

2. Species detection data (Species_records.csv): Contains species-level detection records with exact timestamps for each camera trap station. These data form the basis for temporal activity pattern and overlap analyses. The file provides precise detection times used to model daily activity patterns and estimate temporal overlap among species.

   • Station: Camera trap ID.
   • Species: Species name (e.g., Panthera uncia, Vulpes vulpes, Feral dog).
   • DateTimeOriginal: Original timestamp of the detection event.
   • Date / Time: Extracted date and time components for analysis.
   • Session: Sampling period corresponding to the detection.

   Analysis Codes: Includes R scripts to model species-specific diel activity patterns and quantify temporal overlap between carnivore pairs using GLMM-based predictions.

• shape of activity curves.txt – Models diel activity patterns of individual carnivore species. Uses hourly detection data and trigonometric GLMMs (GLMMadaptive package) to classify activity as unimodal, bimodal, or cathemeral based on best-fitting AIC models.
• interaction_code.txt – Estimates temporal overlap between species pairs by comparing predicted activity probabilities from the best-fitting models, visualizing shared activity periods across a 24-hour cycle.