Toy Dataset for Emulating the Fire Weather Index (FWI) Using Deep Learning Techniques

This dataset includes .nc files designed to replicate a toy example for emulating the Fire Weather Index (FWI) using deep learning techniques. Additionally, .h5 files containing 40 years of pre-trained models are provided for FWI emulation. Predictands: yTrain.nc: A 5-year ERA5-Land computed FWI dataset used to train a toy example deep learning model. yTest.nc: A 3-year ERA5-Land computed FWI dataset used for prediction. Predictors: t2m_Train.nc: A 5-year ERA5-Land daily mean surface temperature dataset used as a predictor to train a deep learning model. hurs_Train.nc: A 5-year ERA5-Land daily mean relative humidity dataset used as a predictor to train a deep learning model. sfcwind_Train.nc: A 5-year ERA5-Land daily mean wind speed dataset used as a predictor to train a deep learning model. Preprocessed Predictors: xTest_stand.Rdata: A dataset containing pre-prepared predictor sets for use with a 40-year pre-trained deep learning model. Deep Learning Models: dense_P1.h5: A 40-year pre-trained Fully Connected Dense (FCD) model, stored in HDF5 format. unet_P1.h5: A 40-year pre-trained U-Net model, stored in HDF5 format.