Hindcast of daily dynamic wildfire probabilities – Trentino and South Tyrol, 2022

Field name Description Science Case Name Hindcast of daily dynamic wildfire probabilities for Trentino and South Tyrol Dataset Name/Title Wildfire predictions Dataset Description Hindcast of daily dynamic wildfire probabilities for the period from 01-07-2022 to 15-07-2022. The predictions illustrate the critical conditions where wildfires are more likely to occur based on static, dynamic, and seasonal controls. Static predictors statistically significant, and therefore considered in the analysis, are landcover, tree density, topographic light, distance to buildings/roads. Dynamic predictors are mean annual precipitation, mean annual temperature and day of the year, and have been combinded dynamically to find the optimal time window to describe the wildfire occurrence i.e., the temperature on the observed day and the cumulative precipitation of 30 days before observation. Direct anthropogenic factors are not considered in the analysis. Key Methodologies Generalized Additive Models (GAMs) Temporal Domain 01-07-2022 to 15-07-2022 for the prediction on daily resolution. Dataset for training and validation: years from 2000-2024 Spatial Domain Italian Provinces of Trentino and South Tyrol; spatial resolition: 50x50 m; EPSG: 32632 Key Variables/Indicators Static predictors: landcover, tree density, topographic light, distance to buildings/roads. Dynamic predictors: mean annual precipitation, mean annual temperature and day of the year Data Format GeoTIFF Source Data Copernicus Land Cover, Precipitation, Temperature, Tree density, Wildfire occurrences Accessibility https://doi.org/10.5281/zenodo.13865655 Stakeholder Relevance Identifying critical conditions that make wildfires more likely to occur Limitations/Assumptions No direct anthropogenic factors considered in wildfire predictions; data before 2000 not considered because of lack of data reliability. Additional information Contact information Mateo Moreno Zapata (editor)