EL-VPD: A High-Resolution Globally Validated Vapor Pressure Deficit Record from 1951 to near present

Overview This repository provides the complete MATLAB workflow used to generate, validate, and visualize the EL-VPD dataset, a globally (90°N–90°S, 180°W–180°E) consistent daily vapor pressure deficit (VPD) product spanning 1951-2025 (1951-01-01 to 2025-12-31) at 0.1° spatial resolution. The dataset is derived from ERA5-Land temperature variables, incorporating elevation-dependent pressure corrections to improve the physical consistency of VPD estimates across diverse climatic and topographic regions. VPD is a key atmospheric variable governing plant water stress, land-atmosphere interactions, ecosystem productivity, wildfire risk, and hydroclimatic extremes. The EL-VPD dataset provides a long-term, high-resolution record suitable for climate, ecological, agricultural, and environmental research applications. Repository Structure The repository is organized into three primary components: meteorological input processing, VPD computation, and reproducibility scripts for figure generation. 1. Input Data Processing The following scripts prepare and harmonize meteorological inputs required for VPD estimation: Process_ERA5Land_Temp_Data.m – Processes ERA5-Land air temperature and dew-point temperature fields. Process_ERA5_Pmsl_Data.m – Processes ERA5 mean sea level pressure data. Processing_Met_Inputs_v2.m – Performs data formatting, quality checks, spatial harmonization, and hemispheric alignment of meteorological inputs. 2. VPD Computation The core VPD estimation workflow is implemented through the following scripts: Year_Wise_VPD_Computation.m – Executes the annual processing workflow and generates output files following the EL-VPD archive structure. ComputeVPD.m – Calculates saturation vapor pressure, actual vapor pressure, and VPD using the Magnus formulation with elevation-dependent pressure correction. 3. Reproducibility and Figure Generation All figures presented in the accompanying manuscript can be reproduced using the supplied MATLAB scripts: Fig_1_Map_ERA5Land_VPD_Distn.m, Fig_2_Map_Station_VPD_Distn.m, Fig_3_Drivers_VPD_Station_ERA5Land.m, Fig_4_Grid_Timeseries_Plotting.m, Fig_5_Performance_Summary_Aridity_Class.m, and Fig_6_Seasonality_VPD.m. These scripts recreate the analyses, visualizations, and performance assessments reported in the manuscript, supporting full computational reproducibility. Data Availability Because of the large volume of the complete EL-VPD archive (approximately 1.5 TB), the daily dataset is not distributed directly through this repository. The full dataset can be accessed through: http://140.112.67.11/public_datasets/EL-VPD/. Access credentials are provided as: User ID: user and Password: public_user. Data Organization The archive is organized as monthly NetCDF-4 files stored within year-specific directories. Each file contains a three-dimensional array with dimensions: time × latitude × longitude, where time = number of days in the corresponding month, latitude = 1801 grid cells spanning 90°N to 90°S at 0.1° resolution, longitude = 3600 grid cells spanning 180°W to 180°E at 0.1° resolution. VPD values are stored in hectopascals (hPa). Files follow the naming convention as Daily_VPD_Y<YEAR>_M<MONTH>.nc. For example, daily VPD estimates for January 2024 are located at /daily/Year_2024/Daily_VPD_Y2024_M1.nc. File Contents Each NetCDF-4 file includes daily VPD values, latitude coordinates, longitude coordinates and time coordinates (days since reference epoch). The files also contain comprehensive metadata describing dataset name and version, spatial and temporal resolution, variable definitions and units, valid data ranges, and missing-value indicators. Citation If you use the EL-VPD dataset, or any portion of this repository in your research, please cite the associated dataset and accompanying manuscript. Proper citation helps support continued development, maintenance, and dissemination of open environmental data products. A formal citation for the dataset and manuscript will be provided upon publication.