Hi, I am working on a habitat-based PFT map over Europe. We first map habitats for present and future conditions, and then split them into CLM5 PFTs. For meaningful results with CLM5 in SP mode, I aim to derive fractions that are compatible with the satellite-observed LAI.
If I understand the CLM logic correctly, a sparse forest with 80% canopy cover could be represented in 2 ways:
(m1) 80% tree PFT and 20% bare/herbaceous PFT, with the LAI of the tree PFT representing dense forest
(m2) 100% tree PFT, with LAI of the tree PFT representing sparse (80% canopy cover) forest
According to the CLM tiling approach, m1 should be the desired method. Considering that the PFT fractions are derived by combining MODIS Vegetation Continuous Fields (which has % tree cover, enabling m1) and MODIS LC (which has mixed LC classes, leading to m2), I imagine that m1 works pretty well for forest and mixes of trees and other LC, and that m2 happens only if small bare/herbaceous fractions cannot be resolved. However, for other PFTs like shrub, there is no data source for % shrub cover, so probably m2 happens for sparse shrubland. Is that generally right?
For instance, I saw over Europe that BDS temperate has quite low LAI during the growing season, lower than grass and crop PFTs. I guess shrubs themselves have higher LAI than grasslands and croplands, but because shrublands are mostly sparse, this LAI might represent shrubland with <100% shrub cover, i.e. using m2.
So in summary, is it right that m1 applies to tree PFTs and m2 for other PFTs, especially shrub and possibly herbaceous types, for which we don't have information on % vegetation cover?
If I understand the CLM logic correctly, a sparse forest with 80% canopy cover could be represented in 2 ways:
(m1) 80% tree PFT and 20% bare/herbaceous PFT, with the LAI of the tree PFT representing dense forest
(m2) 100% tree PFT, with LAI of the tree PFT representing sparse (80% canopy cover) forest
According to the CLM tiling approach, m1 should be the desired method. Considering that the PFT fractions are derived by combining MODIS Vegetation Continuous Fields (which has % tree cover, enabling m1) and MODIS LC (which has mixed LC classes, leading to m2), I imagine that m1 works pretty well for forest and mixes of trees and other LC, and that m2 happens only if small bare/herbaceous fractions cannot be resolved. However, for other PFTs like shrub, there is no data source for % shrub cover, so probably m2 happens for sparse shrubland. Is that generally right?
For instance, I saw over Europe that BDS temperate has quite low LAI during the growing season, lower than grass and crop PFTs. I guess shrubs themselves have higher LAI than grasslands and croplands, but because shrublands are mostly sparse, this LAI might represent shrubland with <100% shrub cover, i.e. using m2.
So in summary, is it right that m1 applies to tree PFTs and m2 for other PFTs, especially shrub and possibly herbaceous types, for which we don't have information on % vegetation cover?
