Example code to perform data assimilation (DA) of sentinel 2 and process-based model (PREBAS) predictions of forest variables (Minunno et al. 2023)

DA is used to combine multiple predictions of forest structural variables (FSV). FSV were predicted using sentinel 2 data for two different years (2016 and 2019). PREBAS is a forest growth and ecosystem carbon balance model, that is used to forecast 2016 estimates to 2019. PREBAS emulators were calibrated and used to reduce the compuational load of DA.

Data processing

Raster files of forest structural variables are read and pre-processed to perform the DA:

library(devtools)
source_url("https://raw.githubusercontent.com/ForModLabUHel/DAexampleCode/master/Rsrc/1_procData_DA.r")

Data assimilation

Data assimilation is performed running the following script:

source_url("https://raw.githubusercontent.com/ForModLabUHel/DAexampleCode/master/Rsrc/2_DA_SFC_FSV.r")

DA raster

New rasters based on DA results can be produced running the script:

source_url("https://raw.githubusercontent.com/ForModLabUHel/DAexampleCode/master/Rsrc/3_DTtoRast.r.r")

References

Minunno, F., et al.. 2023. “Data assimilation of forest status using Sentinel-2 data and a process-based model.” Remote sensing of environment. https://doi.org/....

Acknowledgements:

....