Figure 3
From: Harnessing deep learning to forecast local microclimate using global climate data
Outputs of the physical model. The following figures have been produced using the global climate data collected in the ERA5 database as inputs for the physical model. (a) The figure displays an example of the local variation of the short-wave radiation according to the local terrain conformation of the study area for February 2nd,2023, at 1 p.m. The variation of the short-wave radiation has been modelled as described in the methods section by Eq. (1). (b) The figure displays an example of the local variation of the long-wave radiation according to the local terrain conformation of the study area for February 2nd, 2023, at 1 p.m. The variation of the long-wave radiation has been modelled as described in the methods section by Eq. (9). (c) The figure displays an example of the local variation of the wind speed according to the local terrain conformation of the study area for February 2nd, 2023, at 1 p.m. The variation of the wind speed has been modelled as described in the methods section by Eq. (13). (d) The figure displays an example of the local variation of the reference temperature, corrected by the moist lapse rate according to the local terrain conformation of the study area, for February 2nd, 2023, at 1 p.m. The moist lapse rate correction has been modelled as described in the supplementary materials.
