Table 1. Runtimes for FEEMS and EEMS on the North American gray wolf dataset.

We show a table of runtimes for FEEMS and EEMS for two different grid densities, a sparse grid with 307 nodes and a dense grid with 1207 nodes. The second row shows the FEEMS run-times for applying leave-one-out cross-validation to select λ. The third row shows the run-times when applying FEEMS at the best λ value selected using cross-validation. FEEMS is orders of magnitude faster than EEMS, even when using cross-validation to select λ. Runtimes are based on computation using Intel Xeon E5-2680v4 2.4 GHz CPUs with 5 Gb RAM reserved using the University of Chicago Midway2 cluster.

Method	Sparse grid (run-time)	Dense grid (run-time)
EEMS	27.43 hr	N/A
FEEMS (Cross-validation)	10 min 32 s	1.03 hr
FEEMS (Best λ)	1.23 s	4.08 s