CAFRI Labs: Landsat AGB 1.0.0: One Ensemble to Rule Them All

Mike Mahoney

Evaluation Results

Using version of FIA models 0.0.99 Final Countdown

Using version of LiDAR models 0.0.99 Final Countdown

Change Summary

Ensembling five component models: three Landsat:FIA based, two Landsat:LiDAR based
All models were used to generate predictions against data clipped to FIA plots:
- FIA models were run through 10-fold CV, and predictions were made against the hold-out fold
- LiDAR models, which have never “seen” FIA plots before except indirectly through the LiDAR-AGB surface, predicted the FIA training data directly without any CV
RMSE weighting and LINMOD fitting were performed as normal against these predictions
This test set maps directly to the Landsat-FIA model test set

	FIA-based			LiDAR-based		Ensemble
	RF	GBM	SVM	RF	GBM	RMSE	LINMOD
RMSE	59.403	60.827	57.355	58.275	57.415	56.637	56.626
RMSE %	0.435	0.446	0.420	0.427	0.421	0.415	0.415
MBE	0.762	-0.754	-4.398	1.268	1.352	-0.366	-0.967
R2	0.302	0.271	0.347	0.323	0.345	0.363	0.359

AGB Distribution

summary(bind_rows(training, testing)$agb_mgha)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
   0.00   89.95  134.63  136.85  179.18  425.00

Corrections

If you see mistakes or want to suggest changes, please create an issue on the source repository.

Citation

For attribution, please cite this work as

Mahoney (2022, April 13). CAFRI Labs: Landsat AGB 1.0.0: One Ensemble to Rule Them All. Retrieved from https://cafri-labs.github.io/acceptable-growing-stock/posts/landsat-agb-100-one-ensemble-to-rule-them-all/

BibTeX citation

@misc{mahoney2022landsat,
  author = {Mahoney, Mike},
  title = {CAFRI Labs: Landsat AGB 1.0.0: One Ensemble to Rule Them All},
  url = {https://cafri-labs.github.io/acceptable-growing-stock/posts/landsat-agb-100-one-ensemble-to-rule-them-all/},
  year = {2022}
}