One of our primary research topics is the use of remote sensing data to estimate the terrestrial component of net primary production (NPP). Our objective is to produce accurate estimates of terrestrial NPP over the entire globe, with sub-monthly temporal resolution. This research activity is a major part of our contribution to MODIS Land Science.
Background
Our approach to remote sensing estimates of NPP is based on the concept of radiation
use efficiency (RUE). Solar radiation in the wavelength band from 0.4 - 0.7
micrometers is potentially useful for photosynthesis, and radiation in this
band is called photosynthetically active radiation (PAR). On average about 45%
of the shortwave radiation reaching Earth's surface from the Sun is in the PAR
spectral region. Some of this PAR is absorbed by plants (APAR), and some of
this absorbed energy is used to fix CO2 from the atmosphere as carbohydrate
for growth and respiration. The amount of new growth that is produced from a
given amount of APAR is a measure of the plant's RUE. It depends on the physiological
characteristics of the plant, and also on the environmental conditions at any
particular time. For example, a fast growing forb may have an intrinsically
higher RUE than a slow-growing pine tree, but both plants might have a wide
range of RUE depending on factors such as air temperature and the availability
of water and essential nutrients in the soil.
Implementation
Although the use of remote sensing data can provide global coverage on a frequent
update cycle, the RUE approach to estimating NPP involves many simplifications
and approximations, ignoring much of the ecophysiological complexity of the
real terrestrial primary production processes. In order to capture some of that
complexity in the simple remote sensing approach, we use results from a sophisticated
ecosystem process model (Biome-BGC)
to set the parameter values for our simpler remote sensing algorithm. Graphics
in the following list illustrate the connection between the sophisticated and
simple models, and also show the essential details of the remote sensing NPP
algorithm as it is implemented for the MODIS project.
