Global Fusarium wilt suitability and environmental predictors derived from remote sensing and Earth system models

This repository provides global raster datasets and modelling outputs used to assess the suitability and potential establishment of Fusarium wilt (FW) under current and future climate conditions.
It includes:
Occurrence records of Fusarium oxysporum.Environmental and host-related variables derived from remote sensing (RS) data.An ensemble species distribution model (SDM) output based solely on RS data for the present.Suitability and range projections derived from four Coupled Model Intercomparison Project Phase 6 (CMIP6) Earth System Models (ESMs) and their multi-model ensemble.Repository contents1) Occurrences.csvThis dataset contains georeferenced literature records of Fusarium oxysporum from 1990 to 2022. Each row corresponds to a single occurrence and reports species and forma specialis, host plant, country and administrative subdivision, year and reference, and geographic coordinates (longitude x, latitude y). Spatial accuracy is indicated by Level of precision: 1 = country, 2 = subdivision, 3 = municipality, 4 = exact coordinates.
2) RS_variables.7zThis file contains the global environmental and host-related predictor variables derived from remote sensing data that were used as inputs for the species distribution modelling. Each raster layer was harmonized to 1 km resolution and represents long-term bioclimatic, topographic, edaphic, or vegetation conditions.
The dataset includes the following variables and their data sources:
Bio17 – Precipitation of the driest quarter (kg·m⁻²): Computed from GloH2O MSWEP v2 monthly precipitation (0.1°; 1990–2020).Bio18 – Precipitation of the warmest quarter (kg·m⁻²): Derived from MSWEP v2 precipitation (0.1°; 1990–2020) combined with CHELSA v2.1 monthly temperature to identify the warmest quarter.Bio24 – Annual mean soil moisture (m³·m⁻³): Derived from ESA-CCI Surface Soil Moisture v06.1 product (1 km; 2000–2020).Bio27 – Soil moisture seasonality (%): Derived from ESA-CCI Surface Soil Moisture v06.1 product (1 km; 2000–2020).Bio32 – Annual mean land surface temperature (K): Derived from MODIS Terra/Aqua monthly land surface temperature products (MOD11C3 and MYD11C3, Collection 6.1; 0.05°; 2003–2022).Bio33 – Annual mean land surface temperature diurnal range (K): Computed from MODIS Terra/Aqua land surface temperature products (MOD11C3 and MYD11C3, Collection 6.1; 0.05°; 2003–2022).Bio46 – Evapotranspiration seasonality (%): Calculated from the MODIS/Terra 8-day gap-filled evapotranspiration product (MOD16A2GF, Collection 6.1; 500 m; 2000–2022).Bio60 – Highest monthly wind speed (m·s⁻¹): Derived from CHELSA v2.1 monthly near-surface wind speed data (1 km; 1990–2018).Clay – Clay content (g·kg⁻¹): Extracted from SoilGrids v2.0 at 250 m resolution, aggregated for 0–30 cm soil depth using weighted means.Cropland – Fraction of cropland per pixel (%): Obtained from Landsat-based 4-year average global cropland fraction maps (3 km; 2003–2019).Mean SIF of the driest quarter – Solar-Induced Chlorophyll Fluorescence (SIF): Derived from harmonized and downscaled SCIAMACHY and GOME-2 datasets (0.05°; 2002–2018), averaged for the driest quarter based on MSWEP v2 precipitation data.Silt – Silt content (g·kg⁻¹): Extracted from SoilGrids v2.0 at 250 m resolution, aggregated for 0–30 cm soil depth using weighted means.Slope – Slope (°): Derived from GMTED2010 global elevation dataset (1 km resolution).These data serve as global explanatory layers for developing SDM of FW risk and are intended to support future research on plant disease modelling, environmental suitability analyses, and climate change impact assessments on soilborne pathogens.
3) RS_modelling_output.7zThis file contains the global suitability map for FW derived solely from the RS-based SDM using the remote sensing variables described above. The raster represents suitability values (0–1) for the present period, providing an empirically derived baseline without climate model projections.
4–7) CMIP6 ESM-based projectionsThese four files contain the outputs of the ensemble SDM applied to environmental and host-related variables simulated by individual CMIP6 ESMs:
GFDL-ESM4_modelling_output.7zIPSL-CM6A-LR_modelling_output.7zMPI-ESM1-2-HR_modelling_output.7zUKESM1-0-LL_modelling_output.7zEach archive includes:
Global suitability rasters for the historical baseline and three future periods (2015–2040, 2041–2070, and 2071–2100), each for three climate change scenarios:SSP126 (green development, low forcing),SSP370 (regional rivalry, moderate forcing),SSP585 (fossil-fueled development, high forcing).Categorical range-change maps showing areas of suitability gain or loss based on binary presence–absence transitions.In these range-change maps, each pixel value represents the level of agreement among scenarios on the direction of change:
−6, −5, −4: Areas that lose suitability in three, two, or one scenario(s).−3: Areas of persistent presence, where all scenarios agree on maintaining suitability.0: Areas of persistent absence, where all scenarios agree on unsuitability.1, 2, 3: Areas that gain suitability in one, two, or three scenarios.8) ESMs_ensemble_modelling_output.7zThis file contains the multi-model ensemble that integrates outputs from the four individual ESMs.
For each climate scenario and time window, suitability values were averaged across the four models, producing a comprehensive ensemble prediction.
The file also includes the corresponding range-change maps derived from the combined ensemble, indicating global consensus on projected gains and losses in FW suitability.
Data format and metadataFormat: GeoTIFF, compressed in .7z archives.Coordinate Reference System: EPSG:4326 (WGS84 — World Geodetic System 1984).Spatial resolution: 1 km.