This preprint is currently under review for the journal ESSD., Systematic monitoring and assessment of vegetation dynamics and changes are essential for informing
environmental management and conservation strategies. Addressing this need, our study introduces a pioneering procedure to generate a database of vegetation indices that provides semi-monthly updates from 1981 to the present at a 1.1 km spatial resolution, focusing on the Iberian Peninsula and the Balearic Islands. This database enables near-real-time monitoring and analysis of vegetation anomalies. The methodology developed combines harmonized historical satellite imagery from AVHRR, MODIS, and VIIRS sensors. The database's performance was assessed, demonstrating highly accurate and consistent harmonization of NDVI data over time. Notably, the database is adept at identifying temporal variability and trends in vegetation activity and detecting disturbances caused by fire and other phenomena. This work not only advances our understanding of vegetation dynamics in the region but also serves as a crucial tool for policymakers, environmental managers, and agricultural stakeholders. By providing near-real-time updates and using indices to monitor vegetation anomalies, the data allows for comparisons across seasons and vegetation types. The database, which includes the NDVI and kNDVI vegetation indices as well as their standardized versions, SNDVI and SkNDVI, is accessible via https://doi.org/10.20350/digitalCSIC/16201 (Franquesa et al., 2024)., This research has been supported by the Grant JDC2022-048710-I funded by MCIN/AEI/10.13039/501100011033 and by the European Union NextGenerationEU/PRTR., Peer reviewed

This article belongs to the Topic AI for Natural Disasters Detection, Prediction and Modeling, A substantial proportion of wildfires in Mediterranean regions continue to be recorded without information about the cause or source of ignition, limiting our ability to understand ignition drivers and design effective prevention strategies. In this study, we develop a spatially harmonised wildfire database for mainland Spain by integrating ignition records from the Spanish General Fire Statistics (EGIF) with fire perimeters generated from satellite images. We then apply a Random Forest classifier to infer ignition causes for events lacking cause attribution. To interpret model behaviour, we use Shapley Additive Explanation (SHAP) values at both global and local scales. Results indicate that human-caused ignitions are dominant, with intentional and negligence-related fires accounting for 52.13% of all known events, although they are associated with contrasting climatic and land-use settings. Negligence-related fires tend to occur under hot, dry and windy conditions, often in agricultural interfaces, whereas intentional fires are more frequent under cooler and wetter conditions and in areas with higher population density and land-use change. Lightning-caused fires represent a small fraction of total ignitions (3%) but exhibit a distinct climatic signature, occurring primarily in sparsely populated areas, under intermediate moisture conditions, and often leading to larger burned areas. Despite strong overall model performance (F1-score = 0.82), minority classes (e.g., lightning and fire rekindling, 0.17%) remain challenging to classify, reflecting both data imbalance and uncertainty in causal attribution. Overall, the combined use of machine learning and explainable AI provides a coherent spatial characterisation of wildfire ignition drivers across mainland Spain, highlights systematic differences among ignition causes, and identifies key limitations in existing fire cause records. This framework represents a practical step towards improving fire cause information by integrating remote sensing products with field-based fire reports, thereby supporting more targeted and evidence-based fire risk management., This paper is funded by the H2020 Project FirEUrisk (Grant Agreement Number: 101003890). The FirEUrisk project contributions reported in this paper were carried out by Rodrigues, M., Ochoa, C. and Chuvieco, E. Franquesa, M. was supported by the Grant JDC2022-048710-I, funded by MCIN/AEI/10.13039/501100011033, and by the European Union NextGenerationEU/PRTR., Peer reviewed

[EN] It contains 4 files, one for each vegetation index—NDVI, kNDVI, SNDVI, and SkNDVI—in netcdf format. Multiple options exist for reading and manipulating netCDF files (.nc), with the most common including GIS applications like QGIS and ArcMap, specialized applications such as Panoply (https://www.giss.nasa.gov/tools/panoply/), and dedicated libraries such as ncdf4, raster, or terra in R, as well as netCDF4 or xarray in Python, among others.
[ES] Contiene 4 ficheros, uno por cada índice de vegetación—NDVI, kNDVI, SNDVI y SkNDVI—, en formato netcdf. Existen multiples opciones para leer y manipular ficheros NetCDF (.nc), entre las más habituales encontramos aplicaciones de SIG como Qgis o ArcMap, aplicaciones específicas como Panoply (https://www.giss.nasa.gov/tools/panoply/), o mediante librerías específicas como ncdf4, raster o terra en R o netCDF4 o xarray en Python, entre otras., NetCDF files in this repository correspond to the fixed period from 1981 to 2024. Access to regular database updates is available at https://vi-anomalies.csic.es, [Spatial resolution] 1.1 Km., [Temporal resolution] Bi-weekly., [Geographic extent] Iberian Peninsula and Balearic Islands (Spain, Portugal, and southern France)., [Projected Coordinate Reference System (CRS)] ED50 UTM Z30N (EPSG:23030)., [EN] The Vegetation Indices for the Iberian Peninsula and Balearic Islands (VIIB) Database offers comprehensive long-term time series of vegetation indices—NDVI, kNDVI, SNDVI, and SkNDVI—with a bi-weekly temporal resolution and a spatial resolution of 1.1 km, spanning from 1981 to the present day. Specially designed to encompass the Iberian Peninsula and Balearic Islands, this database facilitates the exploration of both historical and contemporary vegetation patterns and anomalies throughout the region over more than forty years. With bi-weekly updates, it provides a continuous and up-to-date resource for understanding vegetation changes and trends. The database integrates NDVI datasets— Sp_1km_NDVI, MYD13A2 and VNP13A2—derived from AVHRR, MODIS and VIIRS satellite sensors, respectively. These NDVI products have undergone a rigorous harmonization process, ensuring the temporal consistency of the time-series., [ES] La Base de Datos de Índices de Vegetación para la Península Ibérica y las Islas Baleares (VIIB) proporciona series temporal extensas y detalladas de índices de vegetación—NDVI, kNDVI, SNDVI y SkNDVI— con una resolución temporal quincenal y una resolución espacial de 1.1 km, cubriendo el periodo desde 1981 hasta la actualidad. Específicamente diseñada para la Península Ibérica y las Islas Baleares, esta base de datos permite la exploración de patrones de vegetación tanto históricos como actuales, así como la detección y estudio de anomalías a lo largo de la región durante más de cuatro décadas. Con actualizaciones quincenales, proporciona un recurso continuo y actualizado para comprender los cambios y tendencias en la vegetación. La base de datos integra conjuntos de datos de NDVI—Sp_1km_NDVI, MYD13A2 y VNP13A2—derivados de los sensores satelitales AVHRR, MODIS y VIIRS, respectivamente. Estos productos NDVI han sido sometidos a un proceso de armonización riguroso, asegurando la consistencia temporal de las series temporales., Ayuda JDC2022-048710-I financiada por MCIN/AEI /10.13039/501100011033 y por la Unión Europea NextGenerationEU/PRTR. Grant JDC2022-048710-I funded by MCIN/AEI/ 10.13039/501100011033 and by the European Union NextGenerationEU/PRTR., ndvi.nc kndvi.nc sndvi.nc skndvi.nc, No

A substantial proportion of wildfires in Mediterranean regions continue to be recorded without information about the cause or source of ignition, limiting our ability to understand ignition drivers and design effective prevention strategies. In this study, we develop a spatially harmonised wildfire database for mainland Spain by integrating ignition records from the Spanish General Fire Statistics (EGIF) with fire perimeters generated from satellite images. We then apply a Random Forest classifier to infer ignition causes for events lacking cause attribution. To interpret model behaviour, we use Shapley Additive Explanation (SHAP) values at both global and local scales. Results indicate that human-caused ignitions are dominant, with intentional and negligence-related fires accounting for 52.13% of all known events, although they are associated with contrasting climatic and land-use settings. Negligence-related fires tend to occur under hot, dry and windy conditions, often in agricultural interfaces, whereas intentional fires are more frequent under cooler and wetter conditions and in areas with higher population density and land-use change. Lightning-caused fires represent a small fraction of total ignitions (3%) but exhibit a distinct climatic signature, occurring primarily in sparsely populated areas, under intermediate moisture conditions, and often leading to larger burned areas. Despite strong overall model performance (F1-score = 0.82), minority classes (e.g., lightning and fire rekindling, 0.17%) remain challenging to classify, reflecting both data imbalance and uncertainty in causal attribution. Overall, the combined use of machine learning and explainable AI provides a coherent spatial characterisation of wildfire ignition drivers across mainland Spain, highlights systematic differences among ignition causes, and identifies key limitations in existing fire cause records. This framework represents a practical step towards improving fire cause information by integrating remote sensing products with field-based fire reports, thereby supporting more targeted and evidence-based fire risk management.