Long-term global vegetation and climate index datasets

NDVI Data Set (1. NDVI.nc)

Global Vegetation Greenness (NDVI) from AVHRR GIMMS-3G+, 1981-2022

Variable: Normalized Difference Vegetation Index (NDVI)

Area: Global (60°S ~ 70°N, -180°W ~ 180°E)

Period: 1982-01-01 ~ 2022-12-31

Horizontal resolution: 0.25° × 0.25° (Regridded from original 0.0833° × 0.0833°)

Temporal resolution: Bi-monthly (1st–15th and 16th–end of each month)

Source: https://daac.ornl.gov/cgi-bin/dsviewer.pl?ds_id=2187

Meteorological Data Set (2.Temperature.nc, ... , 6.Cloud_cover.nc)

Agrometeorological indicators from 1979 to present derived from reanalysis, Copernicus Climate Change Service

Variable: Normalized Difference Vegetation Index (NDVI)

Area: Global (60°S ~ 70°N, -180°W ~ 180°E)

Period: 1982-01-01 ~ 2022-12-31

Horizontal resolution: 0.25° × 0.25° (Regridded from original 0.1° × 0.1°)

Temporal resolution: Bi-monthly (1st–15th and 16th–end of each month) 

The meteorological data were converted from daily values to bi-monthly average values.

Variables: 2m temperature (K), 2m relative humidity (%), 10m wind speed (m s⁻¹), Precipitation flux (mm day⁻¹), Solar radiation flux (J m⁻² day⁻¹), Cloud cover (dimensionless)

Source: https://cds.climate.copernicus.eu/datasets/sis-agrometeorological-indicators?tab=overview

Pre-processing code (Set_data_1~3)

Set_data_1 : Combining raw data for NDVI and checking for missing values in the original data

Set_data_2 : Combining annual data, calculating semi-monthly averages, and setting the latitude and longitude ranges for meteorological data.

Set_data_3 : Synchronization of latitude, longitude, and resolution between NDVI and meteorological data.

Analysis code (code1~5)

code_1 : This script processes climate data for variables by calculating their seasonal anomalies and time-averaged values. It performs the following steps: 

Monthly Mean Calculation: The script first calculates the monthly mean for each variable over a period of 41 years.

Semi-Monthly Mean Calculation: It then computes the semi-monthly mean by combining the monthly mean data.

Anomaly Calculation: The script calculates the anomaly by subtracting the semi-monthly mean from the original data.

Time Mean Calculation: Finally, the time-mean for the entire time period is calculated for each variable.

code_2 : This script calculates the linear regression slope, intercept, correlation coefficient (r-value), p-value, and standard error for a given climate variable (in this case, temperature anomaly) over time at each latitude and longitude point. The steps involved are:

Load Data: The script loads the input NetCDF file and extracts the time and temperature anomaly (or other climate data) values.

Linear Regression: For each spatial point (latitude, longitude), the script performs a linear regression between time and the corresponding climate data to determine the slope, intercept, r-value, p-value, and standard error.

Save Results: The regression results are saved into a new NetCDF file with variables for slope, intercept, r-value, p-value, and standard error for each latitude and longitude point.

code_3 : This script processes NDVI (Normalized Difference Vegetation Index) data by performing the following steps:

Prepare Heatmap Data: It reshapes the NDVI data into a 4D array of the shape (latitude, longitude, years, 24 months), where each year contains 24 months of data.

Compute NDVI Histograms: It computes histograms of the NDVI data for each latitude, longitude, and year, adjusting the NDVI values into 20 bins for analysis.

Save Histogram Data: The histogram data is then saved to a .npy file, which stores the data for further analysis.

code_4 : This script performs k-means clustering on NDVI data, based on histograms of NDVI values:

Load Data: It loads NDVI data and histogram data (NDVI values in bins) from files.

Filter Data: It filters out regions with zero values to focus on areas of interest.

Reshape Data: The data is reshaped into a 2D format to prepare for clustering.

K-Means Clustering: The script applies k-means clustering to the reshaped histogram data.

Mean NDVI Calculation: It calculates the mean NDVI value for each cluster by extracting values from the NDVI data.

Reordering Clusters: The clusters are reordered based on their mean NDVI values.

Save Results: Finally, the script saves the cluster labels and non-zero indices into separate files.

code_5 : This script processes NDVI (Normalized Difference Vegetation Index) data by clustering and saving the data for each cluster.

Load Data

Loads NDVI slope data (slope) from a NetCDF file.

Loads precomputed cluster labels (cluster_labels_8.npy) and valid data locations (non_zero_indices_8.npy).

Save NDVI Data by Cluster

Categorizes NDVI data based on clusters.

Creates a 2D array for each cluster and assigns NDVI data to the corresponding cluster coordinates.

Saves the clustered NDVI data as  .npy files for further analysis.

Create Directory and Execute

Creates the output directory (if it does not exist).

Runs the main function to save the clustered NDVI data.

Acknowledgments

This work was also supported by Global - Learning & Academic research institution for Master’s·PhD students, and Postdocs (LAMP) Program of the National Research Foundation of Korea (NRF) grant funded by the Ministry of Education (No. RS-2023-00301914).