to replanting, this site was a mix of bare soil and ungrazed pasture. On the other
hand, vegetation at EF-21 is characterized by mixed deciduous forest.
Scatterplots in Figs. 15.8a and 15.9a indicate the validation between observed LE
and estimated LE from MODIS-Terra at EF-21 and EF-14, respectively. Similarly,
comparison of LE calculated from Aqua-Ts products are shown in Figs. 15.8b and
15.9b for EF-21 and EF-14 sites, respectively. Results suggested that similar to
US-Skr, the afternoon T s (i.e., Aqua-based) produced smaller errors in LE than
morning T s (Terra-based) at both sites. The trapezoid model on average yielded
50.674 Wm
À2 and 49.886 Wm
À2 RMSEs in morning and afternoon T s -based LE at
EF-21, respectively. In contrast, RMSEs in LE estimates at EF-14 relatively lower
(39.304 Wm
À2 for Terra and 34.902 Wm
À2 for Aqua) than those at EF-21 due to the
fact that the average observed LE is usually higher at forests than croplands and
Fig. 15.8 Validation of the Terra-based (a) and Aqua-based (b) modeled Latent Heat (LE; energy
equivalent of LE in Wm
À2 units) against diurnal-averaged tower-based LE at EF-21 in 2011. LE is
in Wm
À2 units
Fig. 15.9 Validation of the Terra-based (a) and Aqua-based (b) modeled LE against diurnalaveraged tower-based LE at EF-14 in 2011. LE is in Wm
À2 units
15 Mapping and Monitoring of Soil Moisture, Evapotranspiration, and Agricultural. . .
311