In this paper we combine modeling and remote sensing to
produce maps of forest productivity over a large area, using
data from 10 widely dispersed locations in the Pacific and
Inland Northwest for calibration. For these calibration sites
we obtained the expected linear relationship between max L
and max PAI, calculated using a process-based stand growth
model (3-PG) constrained to reproduce stand properties
generated by the best available empirical growth mode (forest projection and planning system, FPS developed by
Arney et al., 2004). Next we established that the satellitederived
EVI values acquired in mid-summer were also
linearly related to site index; this relationship was expected
since the EVI values depend to a great extent on L. The third
step involves testing the ability of mid summer EVI to predict
Douglas-fir (Pseudotsuga menziesii (Mirb.) Franco) site index
values derived from >5000 federal inventory and analysis
(FIA) field measurements in Oregon. Finally, we generate a
1 km resolution map that predicts site indices across all
forested areas in the Pacific and Inland Northwest using the
correlation with EVI.