Environmental Engineering Reference
In-Depth Information
Falkowski, Barber, and Smetacek 1998; Behrenfeld et al.
2005).
The results of 16 models of global NPP were com-
pared using standardized input variables (Cramer et al.
1999). Most of these models simulate carbon fluxes
using a prescribed vegetation structure; some use data
from the AVHRR sensor as their major input. After
excluding the two extreme values, the results of 16 mod-
els that were reviewed in the end-of-the-century system-
atic comparison ranged from 44.3 Gt C to 66.3 Gt C,
with a mean of 54.1 Gt C (Cramer et al. 1999). Assum-
ing an average of 45% C, this translates to roughly 98-
147 Gt (mean of 120 Gt) of dry phytomass, or assuming
15 GJ/t, to 1.8 ZJ. Two satellite-based models of marine
NPP came up with very similar results: 37-46 Gt C/a
(Antoine, Andr´, and Morel 1996), and 47.5 Gt C/a
(Behrenfeld and Falkowski 1997). The Pacific Ocean
accounts for slightly more than 40% of the total, and the
marine NPP of the two hemispheres is roughly equal be-
cause of much higher output (about 60%) per unit area
in northern oceans.
The most likely range of global NPP at the end of the
twentieth century was thus 100-110 Gt C, that is, 220-
245 Gt of dry phytomass, or 3.3-3.6 ZJ, and annual flux
of 105-115 TW. Two conceptually similar models of ter-
restrial and marine NPP, both with an emphasis on inte-
grating large-scale satellite observations, combine to yield
a global total of 104.9 Gt C, with 56.4 Gt C on land and
48.5 Gt C in the ocean (Field et al. 1998; Geider et al.
2001). Extrapolation of these totals results in power den-
sities of about 450 mW/m
2
of ice-free land and 130
mW/m
2
of ocean. The NPP total of 105 Gt C also
means that the global GPP during the late 1990s (using
the standard assumption NPP
¼
R
A
) was on the order of
210 Gt C, that slightly more than a one-quarter of the
atmospheric CO
2
(at 370 ppm in 2000
¼
787 Gt C)
was drawn into photosynthesis, and that about 13% of it
was annually incorporated into new phytomass. The car-
bon flux in NPP was thus roughly 16 times as large at the
beginning of the twenty-first century as the annual emis-
sions of carbon from the combustion of fossil fuels (6.67
Gt C in 2000).
The Global Primary Production Data Initiative,
launched in 1994, made NPP measurements readily
available in a standardized format. By the year 2005 the
NPP database at the Oak Ridge National Laboratory
contained data for 65 intensively studied sites (mainly
grasslands and tropical and boreal forests) with geo-
referenced climate and site characteristics data (ORNL
2006). But the greatest advance in productivity studies
was the introduction of continuous satellite-derived mea-
sures of terrestrial GPP (issued weekly). This was made
possible by combining the measurements of canopy
reflectance by the Moderate Resolution Imaging Spec-
troradiometer (MODIS) on the Terra satellite, launched
in 1999, with the information on biome type, fraction of
PAR absorbed by vegetation (changing with growth and
senescence), and daily surface climate conditions (Run-
ning et al. 2004). GPP and NPP totals were, respectively,
108.42 Gt C and 56.06 Gt C in the year 2000, and
107.5 Gt C and 54.8 Gt C in 2003 (NTSG 2006).
Global NPP based on MODIS data shows the expected
maxima (
>
1 kg C/m
2
a) in the Congo Basin, parts of
the Amazon, the highlands of Central America, and the
wettest and warmest parts of monsoonal Southeast Asia
(fig. 3.6).
But even these advances do not allow us to determine
global NPP with a high degree of accuracy. Measure-
ments of large-scale CO
2
flux between vegetation and
the atmosphere indicate (after being corrected for R
H
