Dataset DOI: 10.5061/dryad.gqnk98t0k

Description of the data and file structure

This dataset was collected to investigate the demographic parameters and population dynamics of Notomabuya frenata, a lizard species from the Cerrado biome in Brazil. The study aimed to assess growth, survival, reproduction, and population viability under different environmental conditions and climate change scenarios.

Field data was collected from December 2005 to January 2021 in the Reserva Ecológica do IBGE (Distrito Federal, Brazil) using standardized pitfall trap sampling. Additional reproductive data was obtained from museum specimens collected during the filling of the Serra da Mesa reservoir (Goiás, Brazil). Environmental predictors (temperature and precipitation) were sourced from the WorldClim database for past conditions (2005–2020) and future projections (2021–2100) under different climate scenarios.

The dataset includes measurements of individual lizards (snout-vent length, tail length, body mass, sex), capture-recapture histories, reproductive parameters, and environmental data. These data were used to model growth (Von Bertalanffy equation), survival (Gompertz mortality model, Cormack-Jolly-Seber model), reproduction (Bayesian regression models), and population dynamics (Integral Projection Model) to estimate population trends over time and construct a demographic distribution model.

The repository is organized into four main categories: (1) Raw Data (capture records and museum measurements); (2) Processed Data (cleaned, imputed, and spatially explicit climatic datasets); (3) Model Outputs (fitted model objects in .rds format); and (4) R Scripts (code for data inspection, analysis, and figure generation). Missing values are coded as NA.

Note on Language and Script Compatibility

The column headers and categorical data values within the files data_tab_crescimento.csv, Fecundidade_N_frenata.txt, Demografia_N_frenata_corrigido_atejaneiro.xls, and nfrenata5.csv have been maintained in their original Portuguese language (e.g., "cauda" instead of "tail", "s/n" instead of "y/n"). This preservation is necessary to ensure full compatibility and reproducibility with the provided R scripts, which explicitly reference these variable names and specific string values.

To facilitate reuse, comprehensive English descriptions and translation keys for all non-English variables and categorical levels are provided in the Variable Descriptions section of this README.

Files and variables

File: 01_InspectData.R

Description: This R script performs the initial inspection, cleaning, and imputation of the raw mark-recapture data for Notomabuya frenata. The workflow includes:

1. Loading raw data and correcting basic formatting (dates, species names).
2. Identifying and correcting inconsistencies in campaign numbers, plot/trap assignments, and biological plausibility of measurements (mass, SVL, tail length).
3. Detecting outliers using Z-distribution and Huber and Van der Veeken (2008) methods.
4. Imputing missing body mass and Snout-Vent Length (SVL) values using the Random Forest algorithm (package missForest) to maximize data usage.
5. Generating the final cleaned dataset (nfrenata5.csv) and a growth increment table (tab.crescimento) for subsequent demographic analyses.

File: 02_PreliminaryAnalyses.R

Description: This script conducts exploratory data analyses and generates visualizations from the cleaned dataset (nfrenata5.csv). The workflow includes:

1. Capture Dynamics: Analysis of temporal variation in capture and recapture rates, including the frequency of captures per individual.
2. Population Structure & Seasonality: Assessment of juvenile and adult detection patterns over months and their relationship with precipitation (seasonality).
3. Body Size Analysis: Investigation of Snout-Vent Length (SVL) variations across sex and time.
4. Figure Generation: Production of Figure 1 (Scatter plot of SVL over time with rainy season overlays and sexual maturity thresholds).
5. Data Preparation: Calculation of the mean SVL for each sampling campaign. It uses the missForest package to impute missing mean SVL values for months with no captures, exporting the file crc_medio_camp.csv for subsequent environment-dependent modeling.

File: 03_AmbVarSelection.R

Description: This script processes environmental data and selects the climatic predictors for demographic modeling. The workflow includes:

1. Climate Data Extraction: Loads high-resolution climate rasters (WorldClim 2.1) for Minimum Temperature, Maximum Temperature, and Precipitation, cropping them to the study area extent.
2. Collinearity Analysis: Calculates Variance Inflation Factors (VIF) to detect and remove highly correlated environmental variables (e.g., removing Minimum Temperature to avoid multicollinearity).
3. Data Integration & Imputation: Merges the climate data with the mean Snout-Vent Length (SVL) data and uses the Amelia package to impute any remaining missing values in the time series.
4. Variable Selection: Performs variable importance analyses using Random Forest (Boruta), stepwise regression, and AIC-based model selection (MuMIn) to identify which environmental variables best explain variations in mean body size.
5. Output Generation: Exports the final dataset crc_amb_worldclim.csv containing the selected environmental covariates and mean body size for each time step.
   Note: Users must update the file paths for the climatic raster files (TIFFs) to match their local directory structure.

File: crc_amb_worldclim.csv

Description: A dataset combining the time series of mean body size (SVL) of the population with the selected local climatic variables for each month

Variable List:

• Row ID: Row number (automatically generated).
• year: Year of the record.
• month.n: Sequential month number (1 to 182, representing the timeline from Dec 2005 to Jan 2021).
• crc: Mean Snout-Vent Length (SVL) of the population for that month. Unit: millimeters (mm).
• tmax: Mean Maximum Temperature for the study area. Unit: degrees Celsius (°C).
• prec: Mean Precipitation for the study area. Unit: millimeters (mm).

File: 04_WorldClimData.R

Description: This script extracts and processes spatially explicit climate data (Maximum Temperature) required for the demographic projections. The workflow includes:

1. Occurrence Mapping: Visualizes species occurrence points (N_frenata_occurclean.xlsx) on a map of South America to define the study area extent.
2. Climate Data Retrieval:
   • Loads historical WorldClim data (2005-2020) from local TIFF files.
   • Downloads future climate projections (CMIP6, model MPI-ESM1-2-HR) for two Shared Socioeconomic Pathways: SSP2-4.5 (optimistic) and SSP5-8.5 (pessimistic) covering the period 2021-2100.
3. Data Processing: Crops the climate layers to the species' distribution buffer, extracts pixel-level temperature values, and organizes them into long-format data frames containing coordinates (lat/lon), year, month, and maximum temperature.
4. Output: Saves the processed spatiotemporal climate datasets as .rds files (amb_worldclim_245_20052100_BufferCrop.rdsand amb_worldclim_585_20052100_BufferCrop.rds) for use in spatial IPMs.

File: 05_Growth_Sex_FabensVB.R

Description: This script analyzes the growth patterns of Notomabuya frenata considering sexual dimorphism. It fits sex-specific Von Bertalanffy Growth Models (VBGM) using Fabens' method implemented via Non-linear Mixed-Effects (NLME). The workflow includes:

1. Data Classification: Categorizes individuals as Male, Female, or Juvenile based on their entire capture history to maximize sample size for each sex.
2. Statistical Testing: Performs T-tests, Shapiro-Wilk, and Wilcoxon tests to assess differences in body size between sexes.
3. Growth Modeling: Fits separate NLME models for females and males to estimate growth parameters (K and L∞​) while accounting for individual heterogeneity.
4. Visualization: Generates Figure 2, plotting the estimated growth curves and confidence intervals for both sexes against age.
5. Output: Exports the growth increment dataset (data_tab_crescimento.csv) and saves the fitted model objects (e.g., growth_fem.rds).

File: 06_Survival_BaSTA.R

Description: This script estimates age-specific survival and mortality rates for Notomabuya frenata using the BaSTA(Bayesian Survival Trajectory Analysis) package. The workflow includes:

1. Data Filtering: Selects adult individuals (Males and Females) and removes outliers, similar to the growth analysis scripts.
2. Input Preparation:
   • Converts capture histories into the specific matrix format required by BaSTA.
   • Estimates birth times (Tbirth​) by back-calculating from the estimated age at first capture (derived from the growth models in 05_Growth_Sex_FabensVB.R).
   • Constructs a covariate matrix to test the effect of Sex (Female vs. Male) on mortality.
3. Bayesian Modeling: Runs the BaSTA model (CMR configuration) with 50,000 MCMC iterations to estimate Gompertz mortality parameters.
4. Visualization:
   • Generates diagnostic plots (trace plots, posterior densities) to check convergence.
   • Produces Figure 3, which compares the age-specific mortality and survival curves for Males and Females.
5. Output: Saves the BaSTA model output (e.g., outBaSTA_sex.rds and outBaSTA_Fem.rds).

Note: This script assumes the objects data_validF and data_validM from the growth analysis script are available in the R environment.

File: 07_Survival_CJS.R

Description: This script fits Cormack-Jolly-Seber (CJS) models to estimate survival (Φ) and recapture (p) probabilities as functions of environmental and biological covariates. The analysis is performed using the marked package. The workflow includes:

1. Capture History Construction: Converts the individual capture data into a monthly capture history matrix (0/1 format), filling in gaps for months without sampling campaigns to ensure a continuous time series (182 months).
2. Covariate Integration: Linkages the capture history with time-varying covariates:
   • Environmental: Mean Maximum Temperature (tmax) and Precipitation (prec) from crc_amb_worldclim.csv.
   • Biological: Mean population SVL (crc) from crc_medio_camp.csv.
3. Model Selection: Defines and runs a set of candidate CJS models testing different combinations of covariates on Φand p.
4. Visualization:
   • Figure 4a: Plots the estimated survival probability over time alongside maximum temperature.
   • Figure 4b: Visualizes the correlation between maximum temperature and survival.
   • Figure S1: Decomposes the survival time series into seasonal, trend, and residual components.
5. Output: Saves the CJS model results object (e.g., surv_VarAmb_WC.rds).

File: 08_FecundityCRC.R

Description: This script analyzes the relationship between female body size (SVL) and fecundity (litter size) using Bayesian Generalized Linear Models (GLM) with a Poisson distribution (implemented via brms). The workflow includes:

1. Data Loading: Imports the fecundity dataset containing measurements of female SVL and their corresponding litter sizes.
2. Model Fitting: Fits and compares four candidate models to describe the size-fecundity relationship:
   • Null model (Intercept only).
   • Linear model (neggs ~ crc).
   • Non-linear model with smoothed splines (neggs ~ s(crc)).
   • Quadratic model (neggs ~ crc + crc^2).
3. Model Selection: Uses Leave-One-Out Cross-Validation (loo) to select the model with the best predictive performance.
4. Visualization: Generates Figure S2, plotting the observed data against the predicted relationship (with credible intervals) derived from the selected Bayesian model.
5. Output: Saves the selected fecundity model object (e.g., mnb_Nfrenata.rds).

File: 09_Occurrence_data.R

Description: This script processes raw species occurrence records to define the spatial extent of the study. The workflow includes:

1. Data Cleaning: Imports raw occurrence data, corrects coordinate formatting (decimal separators), standardizes scientific names, and filters records for Notomabuya frenata.
2. Duplicate Removal: Identifies and removes duplicate geographical coordinates to create a unique set of occurrence points.
3. Data Export: Saves the cleaned occurrence dataset as N_frenata_occurclean.xlsx.
4. Spatial Visualization:
   • Loads the South American continent shapefile.
   • Generates a Minimum Convex Polygon (MCP) representing the species' extent of occurrence (Note: The MCP generation code is present but commented out; the script reads the resulting shapefile n_frenata_mcp_sf.shpfor plotting).
   • Produces Figure S1 (referred to as "Figure 0" in the code), calculating the Minimum Convex Polygon (MCP) and mapping the occurrence points.

File: 10_IPM_2005_2020.R

Description: This script builds spatially explicit Integral Projection Models (IPMs) to estimate population growth rates (λ) across the species' distribution for the historical period (2005–2020). It serves as the core integration step, linking vital rates to environmental drivers. The workflow includes:

1. Vital Rate Integration: Loads parameters from the previously fitted growth (growth_fem.rds), age-dependent survival (outBaSTA_sex.rds), and environment-dependent survival (surv_VarAmb_WC.rds) models.
2. Kernel Construction: Defines the projection kernels (P for survival/growth and F for fecundity) and constructs specific kernels for each grid cell and month based on local climatic conditions (Temperature).
3. Population Analysis:
   • Calculates the asymptotic population growth rate (λ) for each location and time step.
   • Performs perturbation analyses (Sensitivity and Elasticity) to quantify the relative importance of different demographic processes.
   • Computes the geometric mean of λ to assess overall population stability.
4. Visualization:
   • Figures 5a & 5b: Maps the projected mean and standard deviation of population growth (λ) for the 2005-2020 period.
   • Figure S3: Plots the time series of population growth (λ) with seasonal overlays.
   • Figure S4: Decomposes the λ time series into seasonal, trend, and residual components.
   • Generates kernel visualizations (P, F, and K matrices).
5. Output: Saves the computed metrics and kernel objects (e.g., lambda_2005.rds to lambda_2020.rds, MeanKernel_20052020.rds).

File: 10_IPM_ssp245_2021_2100.R

Description: This script projects the population dynamics of Notomabuya frenata under the SSP2-4.5 climate change scenario (an "optimistic" or intermediate greenhouse gas emission pathway) for the period 2021–2100. The workflow essentially mirrors the historical analysis but applies future climate projections:

1. Model Initialization: Loads the fitted vital rate models (Growth, Survival, Fecundity) and defines the integration parameters (kernels).
2. Future Climate Integration: Loads the spatially explicit climate projections (amb_worldclim_245_20052100_BufferCrop.rds) and filters for the future time slices.
3. Kernel Construction & Projection:
   • Constructs IPM kernels for every grid cell across the species' distribution for four future periods: 2021–2040, 2041–2060, 2061–2080, and 2081–2100.
   • Calculates the asymptotic population growth rate (λ) for each location and time step.
4. Visualization:
   • Figures 6a & 6b: Generates maps showing the projected mean λ and its standard deviation across the distribution for the four future periods.
5. Output: Saves the computed λ metrics for each future time slice (e.g., lambda_ssp245_2021_2040.rds, etc.).

File: 10_IPM_ssp585_2021_2100.R

Description: This script projects the population dynamics of Notomabuya frenata under the SSP5-8.5 climate change scenario (a "pessimistic" or high greenhouse gas emission pathway) for the period 2021–2100. The workflow is identical to the SSP2-4.5 analysis but uses the high-emission climate projections:

1. Model Initialization: Loads the fitted vital rate models (Growth, Survival, Fecundity) and integration parameters.
2. Future Climate Integration: Loads the spatially explicit climate projections (amb_worldclim_585_20052100_BufferCrop.rds) corresponding to the SSP5-8.5 scenario.
3. Kernel Construction & Projection:
   • Constructs IPM kernels for every grid cell across the species' distribution for four future periods: 2021–2040, 2041–2060, 2061–2080, and 2081–2100.
   • Calculates the asymptotic population growth rate (λ) for each location and time step.
4. Visualization:
   • Figures 7a & 7b: Generates maps showing the projected mean λ and its standard deviation across the distribution for the four future periods under this high-emission scenario.
5. Output: Saves the computed λ metrics for each future time slice (e.g., lambda_ssp585_2021_2040.rds, etc.).

File: 11_LambdaBoxplots.R

Description: This script performs a comparative analysis of population growth rates (λ) across different time periods and climate scenarios. It uses the outputs generated by the IPM scripts (10_IPM...) to visualize demographic trends. The workflow includes:

1. Data Integration: Combines the historical population growth metrics (2005–2020) with projections from the SSP2-4.5 (optimistic) and SSP5-8.5 (pessimistic) scenarios (2021–2100).
2. Visualization of Trends:
   • Figure 8a: Generates a boxplot comparing the distribution of λ in the historical period versus future periods under the SSP2-4.5 scenario.
   • Figure 8b: Generates a boxplot comparing the distribution of λ in the historical period versus future periods under the SSP5-8.5 scenario.
   • Figure 9: Compares the two future scenarios side-by-side to highlight the potential impact of higher greenhouse gas emissions on population viability.
3. Output: Exports the final publication-ready figures in PDF format.

File: crc_medio_camp.csv

Description: Monthly mean snout-vent length

File: data_tab_crescimento.csv

Description: A derived dataset containing growth increments for recaptured individuals, used as input for the growth models.

Variable List:

• identidade: Unique identifier for the individual.
• delta.camp: Time interval between the initial capture and the recapture (in months).
• crc.t0: Snout-Vent Length (SVL) at the initial capture (Lt​). Unit: mm.
• crc: Snout-Vent Length (SVL) at the recapture (Lt+Δt​). Unit: mm.
• delta.crc: Change in SVL during the interval (ΔL). Unit: mm.
• sexo: Sex of the individual assigned for the growth analysis.
  • Key: F = Female, M = Male, J = Juvenile (or unsexed).

File: Fecundidade_N_frenata.txt

Description: Raw dataset containing morphological and reproductive data obtained from the dissection of preserved museum specimens (Herpetological Collection of the University of Brasília - CHUNB). These specimens were collected during the fauna rescue of the Serra da Mesa hydroelectric power plant reservoir and were used to parameterize the size-fecundity relationship.

Variable List:

• numero: Specimen voucher number/ID in the scientific collection.
• data: Date of collection (Format: DD.MM.YY).
• dia: Day of collection.
• mes: Month of collection.
• ano: Year of collection.
• local: Sampling locality (e.g., "S._Mesa" = Serra da Mesa).
• especie: Scientific name of the species.
• sexo: Sex of the specimen (f = Female).
• condicao_reprodutiva: Text description of the reproductive status (e.g., "5_embrioes" = 5 embryos, "4_fetos" = 4 fetuses).
• embrioes: Count of embryos found.
• fetos_filhotes: Count of fetuses or near-term offspring found.
• tamanho_ninhada: Total litter size (number of eggs/embryos). Unit: Count.
• condicao: Qualitative condition of the specimen or reproductive tract (e.g., "b", "c"). Key: "Internal reference code".
• crc: Snout-Vent Length (SVL). Unit: millimeters (mm)

File: Demografia_N_frenata_corrigido_atejaneiro.xls

Description: The raw, long-term mark-recapture dataset collected from December 2005 to January 2021. It contains individual capture records, morphological measurements, and microsite data for Notomabuya frenata.

Variable List:

• campanha: Sequential campaign number.
• mês: Month of capture.
• data: Date of capture. Note: In the raw CSV/Excel, this may appear as a serial number (e.g., 38688) or formatted date.
• ano: Year of capture.
• parcela: ID of the sampling plot (e.g., "BT", "Q", "BP", "C").
• armadilha: Trap number within the plot.
• dedos: Raw toe-clipping code representing the marked digits (e.g., "40_20_6_5").
• identidade: Unique alphanumeric identifier assigned to the individual.
• massa: Body mass. Unit: grams (g).
• crc: Snout-Vent Length (SVL). Unit: millimeters (mm).
• recaptura: Recapture status for that event. Key: "s" = Yes (Sim), "n" = No (Não).
• cc: Tail length (Comprimento da Cauda). Unit: millimeters (mm).
• bc: Length of the tail base (measured from the cloaca to the proximal limit of the regenerated portion). Unit: millimeters (mm).
• cauda: Tail condition. Key: "inteira" = Intact, "regenerada" = Regenerated, "quebrada" = Broken.
• sexo: Sex of the individual. Key: "F" = Female, "M" = Male, "J" = Juvenile.
• morto: Mortality status during capture. Key: "s" = Dead (Sim), "n" = Alive (Não).
• observacao: Field notes and observations (e.g., "sem_pata_anterior" = missing forelimb).
• month.n: Sequential month number calculated from the start of the study.

File: N_frenata_occurclean.xlsx

Description: A processed dataset containing unique, georeferenced occurrence records of Notomabuya frenata. This file is the output of the data cleaning script (09_Occurrence_data.R) and serves as the input for defining the study area and extracting climatic variables in the WorldClim processing script (04_WorldClimData.R).

Variable List:

• lon: Longitude of the occurrence record. Unit: Decimal degrees (WGS84).
• lat: Latitude of the occurrence record. Unit: Decimal degrees (WGS84).

File: nfrenata5.csv

Description: The cleaned and processed mark-recapture dataset used for all demographic analyses. This dataset is derived from the raw data (Demografia_N_frenata_corrigido_atejaneiro.xls) after cleaning, error correction, and imputation of missing values (mass and SVL) as performed by the script 01_InspectData.R.

Variable List:

• nrow: Row index.
• campanha: Sequential sampling campaign number.
• mês: Month of capture (Numeric, 1-12).
• data: Date of capture (Format: YYYY-MM-DD).
• ano: Year of capture.
• parcela: Sampling plot ID (e.g., BT, Q, BP, BM, C).
• armadilha: Trap number within the plot.
• dedos: Toe-clipping code representing the marked digits.
• identidade: Unique alphanumeric identifier for each individual.
• massa: Body mass. Contains imputed values for missing entries. Unit: grams (g).
• crc: Snout-Vent Length (SVL). Contains imputed values for missing entries. Unit: millimeters (mm).
• recaptura: Recapture status. Key: "s" = Yes (recapture), "n" = No (first capture).
• cc: Tail length. Unit: millimeters (mm).
• bc: Length of the tail base (cloaca to regeneration start). Unit: millimeters (mm).
• cauda: Tail condition. Key: "inteira" = Intact, "regenerada" = Regenerated.
• sexo: Sex of the individual. Key: "F" = Female, "M" = Male, "J" = Juvenile/Unsexed.
• morto: Mortality status during capture. Key: "s" = Dead, "n" = Alive.
• observacao: Field observations (e.g., physical injuries).
• month.n: Sequential month number from the start of the study (1 to 182).
• especie: Scientific name (Notomabuya frenata).
• year, month, day: Date components extracted from the 'data' column.

File: growth_fem.rds

• Description: The Non-linear Mixed-Effects (NLME) model object describing the Von Bertalanffy growth curve for females. Used to estimate body size transitions.

File: MeanKernel_20052020.rds

• Description: The mean Integral Projection Model (IPM) kernel representing the average demographic transitions over the historical period (2005-2020).

File: mnb_Nfrenata.rds

• Description: The Bayesian Generalized Linear Model (Poisson) object describing the relationship between female body size (SVL) and fecundity (litter size).

File: outBaSTA_sex.rds

• Description: The Bayesian Survival Trajectory Analysis (BaSTA) model output containing age-specific mortality estimates for males and females.

File: surv_VarAmb_WC.rds

• Description: The Cormack-Jolly-Seber (CJS) model object estimating monthly survival probabilities as a function of maximum temperature.

File: WC_2005_2020.zip

Description: A compressed archive containing the raw raster files (GeoTIFF format) of the climatic variables used to characterize the environmental conditions of the study area during the historical period. These data were sourced from WorldClim v2.1 (downscaled CRU-TS 4.06) at a spatial resolution of 10 minutes (~340 km²).

Contents: The archive includes monthly regional layers for the years 2005 to 2020 for the following variables (referenced in script 03_AmbVarSelection.R ):

• Maximum Temperature (tmax): Monthly averages of maximum air temperature (°C).
• Minimum Temperature (tmin): Monthly averages of minimum air temperature (°C).
• Precipitation (prec): Total monthly precipitation (mm).

Note: These files are required to run the scripts 03_AmbVarSelection.R and 04_WorldClimData.R, which extract local climate values for the species' occurrence points and study area buffer.

Code/software

The dataset can be processed and analyzed using R (v4.3.3).

Access information

Climatic data were derived from the WorldClim database.