All meteorological and CO 2 flux datasets used in this case study were downloaded from the AmeriFlux data portal (http://public.ornl.gov/ameriflux/). The
AmeriFlux network is a part of FLUXNET (http://daac.ornl.gov/FLUXNET),
which aims to continuously observe CO 2 , water, and energy fluxes at ecosystem
and landscape levels. It provides meteorological and CO 2 flux datasets at daily,
weekly, and monthly intervals. The weekly gap-filled and CO 2 flux-partitioned
datasets of Level 4 products during 2001–2005 were chosen for this study, as this
time scale matches the MODIS eight-day composite products used here.
The seasonal dynamics of GPP VPM predicted by the VPM were compared with
GPP EC (Fig. 5.3). Both GPP VPM and GPP EC rose rapidly in June, and reached
seasonal peaks in July–August for maize and soybeans. GPP VPM decreased to zero
in September and remained near zero until harvest, which agreed well with GPP EC
dynamics.
The scatterplots between GPP VPM and GPP EC of maize over the crop-growth
period for an individual year and all years (Fig. 5.4) show that GPP VPM is strongly
correlated with GPP EC , the correlation coefficient (R
2
) is 0.92 across multiple
Fig. 5.2 Seasonal dynamics and interannual variations of Normalized Difference Vegetation
Index (NDVI), Enhanced Vegetation Index (EVI) and Land Surface Water Index (LSWI) during
2001–2005 at the Mead site, Nebraska, USA
Fig. 5.3 Seasonal dynamics and interannual variations of the VPM-predicted gross primary
production (GPP VPM , g C m
-2 day
-1
) and the estimated GPP from the CO 2 eddy flux tower data
(GPP EC ) during 2001–2005 at the Mead site, Nebraska, USA
5 Gross Primary Production of Terrestrial Vegetation
139
AmeriFlux network is a part of FLUXNET (http://daac.ornl.gov/FLUXNET),
which aims to continuously observe CO 2 , water, and energy fluxes at ecosystem
and landscape levels. It provides meteorological and CO 2 flux datasets at daily,
weekly, and monthly intervals. The weekly gap-filled and CO 2 flux-partitioned
datasets of Level 4 products during 2001–2005 were chosen for this study, as this
time scale matches the MODIS eight-day composite products used here.
The seasonal dynamics of GPP VPM predicted by the VPM were compared with
GPP EC (Fig. 5.3). Both GPP VPM and GPP EC rose rapidly in June, and reached
seasonal peaks in July–August for maize and soybeans. GPP VPM decreased to zero
in September and remained near zero until harvest, which agreed well with GPP EC
dynamics.
The scatterplots between GPP VPM and GPP EC of maize over the crop-growth
period for an individual year and all years (Fig. 5.4) show that GPP VPM is strongly
correlated with GPP EC , the correlation coefficient (R
2
) is 0.92 across multiple
Fig. 5.2 Seasonal dynamics and interannual variations of Normalized Difference Vegetation
Index (NDVI), Enhanced Vegetation Index (EVI) and Land Surface Water Index (LSWI) during
2001–2005 at the Mead site, Nebraska, USA
Fig. 5.3 Seasonal dynamics and interannual variations of the VPM-predicted gross primary
production (GPP VPM , g C m
-2 day
-1
) and the estimated GPP from the CO 2 eddy flux tower data
(GPP EC ) during 2001–2005 at the Mead site, Nebraska, USA
5 Gross Primary Production of Terrestrial Vegetation
139