You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The addition of skyview factor as a new input feature for training the models could better account for the local patterns related to shortwave radiation. This would imply:
Find a way to generate the raw data. This means running an available algorithm to compute this for a globally available DEM (e.g. SRTM pulled from OGGM).
Add this data in the DataProcessing module as part of the Dataset class. This would imply finding the coordinates of each stake/data point and pulling the right skyview factor value for each point.
Adding this in the training dataset and quantify if it has an added value.
And idea for a qualitative validation (beyond the quantitative ones based on training metrics) would be to use optical data of glaciers to see if shadow patters on glaciers are well captured by this new input feature.
The text was updated successfully, but these errors were encountered:
The addition of skyview factor as a new input feature for training the models could better account for the local patterns related to shortwave radiation. This would imply:
DataProcessing
module as part of theDataset
class. This would imply finding the coordinates of each stake/data point and pulling the right skyview factor value for each point.And idea for a qualitative validation (beyond the quantitative ones based on training metrics) would be to use optical data of glaciers to see if shadow patters on glaciers are well captured by this new input feature.
The text was updated successfully, but these errors were encountered: