Comparison of LUE, LAI and GPP estimation methods in a Southern Indiana broadleaf deciduous forest

1
Comparison of LUE, LAI and GPP estimation methods
in a Southern Indiana broadleaf deciduous
forest V.D. Cordova1 A. F. Rahman1 H.P.
Schmid2 D.A. Sims1. 1 Ball State University.
Muncie, IN 47306 2 Indiana University.
Bloomington, IN 47405
Flux Tower
MODIS Imagery and Products
LAI overestimation was evident for MODIS,
according to which the growing season starts with
a full foliage coverage. Compared to field
measurements, the LAI readings were significantly
different (t-5.33plt0.01). Biome-BGC modeled LAI
appeared to be statistically similar to field
readings (t-1.09pgt0.05).
Ecosystem modeling BIOME-BGC
Historic Meteorological Data Precipitation Max.
Temperature Min. Temperature
Biome Characteristics
BIOME-BGC Simulation
Site Modeled Data Precipitation Max.
temperature Min. temperature Day
temperature Solar radiation Vapor pressure
deficit Day length
LAI predictability using Biome-BGC appeared to be
more feasible based on its high correlation to
field measurements (r0.82plt0.01). Correlation
of MODIS LAI with field measurements were not
significant. (r0.15pgt0.05).
Spatial Information DEM Slope Aspect Horizon
elevation Geog. Coordinates
When compared to Flux tower measurements, MODIS
overestimated LAI showing an almost constant
value throughout the growing season. Biome-BGC
LAI was similar to field LAI, especially at the
beginning and middle of the growing season.
MODIS based Light Use Efficiency estimations
products were lower than the Flux tower based
LUE. Biome-BGC also showed lower LUE when
compared to Flux tower based values. MODIS GPP
estimation was considerably lower than Flux tower
GPP estimates. The difference in GPP estimation
between Biome-BGC and Flux tower measurements was
not significant. The temporal extent of this
study was limited to one year. The use of data
from several years will help to determine if the
observed trends are consistent through time.
References Grimmond C.S.B., Hanson P. J., Schmid
H. P., Wullschleger S.D. and Cropley F. 2000.
Evapotranspiration rates at the Morgan Monroe
State Forest Ameriflux site a comparison of
results from eddy covariance turbulent flux
measurements and sap flow techniques. 15th
Conference on Hydrology, American Meteorological
Society Long Beach, CA . January 2000. Kimball
J.S., White M.A., Running S.W. 1997. BIOME-BGC
simulations of BOREAS stand hydrologic processes.
Journal of Geophysical Research. 102(D24),
29043-29051. Knyazikhin Y., Martonchik J. V.,
Myneni R. B., Diner D. J., and Running S. 1998.
Synergistic algorithm for estimating vegetation
canopy leaf area index and fraction of absorbed
photosynthetically active radiation from MODIS
and MISR data. J. Geophys. Res.,
10332,257-32,275. Running S.W., Nemani R.,
Glassy J., Thornton P. 1999b. MODIS daily
photosynthesis (PSN) and annual net primary
production (NPP) product (MOD17). Algorithm
Theoretical Basis Document Version 3.0. NASA. 59
p Schmid H. P., Grimmond S.B., Cropley F.,
Offerle B., Hong-Bing Su. 2000. Measurements of
C02 and energy fluxes over a mixed hardwood
forest in the mid-western United States.
Department of Geography, Indiana University,
Bloomington, IN 47405, USA. 18 p.
MODIS consistently underestimated GPP compared to
Flux tower based productivity (t3.67 plt0.01).
The accumulated difference is about 400 g C m-2
y-1. Modeled GPP showed statistically similar
readings to Flux tower based data, along the
growing season. The overall difference between
Biome- BGC and Flux tower based GPP is around 90
g C m-2 y-1.
Biome-BGC GPP correlated better to Flux tower
based GPP (r0.88 plt0.01). Modis GPP shows a
weaker but still highly significant correlation
with Flux tower readings (r0.62 plt0.01).