Dataset DOI: 10.5061/dryad.h18932016

Description of the data and file structure

This dataset compiles outputs from Lagrangian dispersal simulations performed with the ICHTHYOP model (Lett et al., 2008), using hydrodynamic fields from the Regional Ocean Modeling System (ROMS) to investigate larval transport processes in the Western Tropical Atlantic between 1992 and 2015. The dataset represents the trajectories and survival of virtual larvae constrained by biological and environmental factors, including pelagic larval duration (PLD), diel vertical migration (DVM), and tolerance limits for temperature and salinity. Simulations were conducted under two scenarios—with and without DVM—across 24 years and two hydrological seasons (wet and dry), totaling 288 experiments.

A total of 614 spawning habitats, each covering approximately 500 km², were defined along five coastal ecoregions: Northeast Brazil, São Marcos Bay, Pará-Maranhão, the Mouth of the Amazon River, and Amapá. In each simulation, 70,000 particles were released within the upper 50 m of the water column, totaling 420,000 particles per year. The dataset includes particle trajectories, mortality causes, environmental conditions, and metadata describing the spatial and temporal configuration of each release.

Files and variables

File: Data_Table_ARM.xlsx

Description:  Tabulated data for the multivariate regression tree.

File: Data_Table_GAM.xlsx

Description:  Tabulated data for the Generalized Additive Model.

The tabular data files (Data_Table_GAM.xlsx and Data_Table_ARM.xlsx) contain the following variables:

• Zone – Geographic location identifier of the 614 individual spawning habitats considered in the Lagrangian particle dispersion simulations. Each record represents a spawning habitat used as a source or destination in connectivity analyses. Spatial locations and associated habitat polygons can be visualized in Ecoregions.zip, which contains the shapefiles of spawning zones and marine ecoregions used in the simulations. No units apply.
• Season – Seasonal oceanographic scenario used in the simulations. Two categorical classes were considered:
  • Wet: February–April, period associated with the northwestward flow of the North Brazil Current (NBC).
  • Dry: August–October, period during which approximately 70% of plume waters follow the NBC retroflection and feed the North Equatorial Countercurrent (NECC).
    No units apply.
• Diel vertical migration (DVM) – Indicates whether diel vertical migration behavior was included in the particle simulations. DVM refers to the synchronized vertical movement of marine organisms within the water column over a 24-hour cycle. This variable is categorical with two possible values:
  • With – simulations including vertical migration behavior.
  • Without – simulations assuming passive particles without vertical migration.
    No units apply.
• Ecoregion – Marine ecoregion classification assigned to each spawning habitat. For analytical purposes, spawning habitats were grouped into five ecoregions along the northern Brazilian continental shelf, from east to west: 1. Northeast Brazil; 2. São Marcos Bay; 3. Pará–Maranhão; 4. Mouth of the Amazon River; 5. Amapá
  This is a categorical variable with no units.
• Distance(km) – Geographic distance between spawning habitats or connectivity nodes, measured in kilometers (km). Continuous numerical variable.
• Pelagic larval duration(days) – Duration of the pelagic larval stage used to define the dispersal time window in the simulations. Four PLD scenarios were analyzed: 5 days; 15 days; 30 days; 45 days
  Units: days.
• Depth(m) – Bathymetric depth of the spawning habitat location. Continuous numerical variable measured in meters (m).

File: R_ARM.R

Description: R script for a multivariate regression tree.

File: R_GAM.R

Description:  R script for a multivariate regression tree. 

File: DistanceTraveled_withDVM_EN.pkl

Description: Processed results from the dispersal simulation performed in Ichthyop (a Lagrangian tool for simulating ichthyoplankton dynamics) containing the cumulative horizontal distance traveled by each particle throughout its pelagic duration. Scenario with daily vertical migration. The .pkl file can be opened in Python 3.x using the built-in pickle library or the pandas package. Users can load the file with pickle.load() or pandas.read_pickle(), depending on the object structure stored in the file. The file contains serialized Python objects generated during the post-processing of Ichthyop simulation outputs, and requires a Python environment with the appropriate libraries installed for access and analysis.

File: DistanceTraveled_withoutDVM_EN.pkl

Description:  Processed results from the dispersal simulation performed in Ichthyop (a Lagrangian tool for simulating ichthyoplankton dynamics) containing the cumulative horizontal distance traveled by each particle throughout its pelagic duration. Scenario without daily vertical migration. The .pkl file can be opened in Python 3.x using the built-in pickle library or the pandas package. Users can load the file with pickle.load() or pandas.read_pickle(), depending on the object structure stored in the file. The file contains serialized Python objects generated during the post-processing of Ichthyop simulation outputs, and requires a Python environment with the appropriate libraries installed for access and analysis.

File: Ecoregions.zip

Description: Shapefile containing the spawning zones and ecoregions considered in the particle dispersion simulation. The Shapefile can be opened in Geographic Information System (GIS) software such as QGIS or ArcGIS Pro. After extracting the contents of the .zip file, users should load the .shp file into the GIS environment using the “Add Layer” or “Add Data” function. The associated files (.dbf, .shx, .prj, and others) must remain in the same directory as the .shp file to ensure correct visualization and attribute access. Once loaded, the shapefile can be visualized as georeferenced vector data representing the spawning zones and ecoregions used in the particle dispersion simulations.

Code/software

The data can be opened and viewed in R and/or Python.

Access information

Other publicly accessible locations of the data:

• N/A

Data was derived from the following sources:

• N/A