Trace Gases - The Biogeochemistry of Submerged Soils - page 243

power. Nonetheless the great strength of a mechanism-based modelling approach
is that the sensitivity to different variables can be tested and the effects of time-
dependent variables, feedback mechanisms and the consequences of changes in
the system can be explored.

Down-scaling using Inverse Modelling

The spatial and temporal variation in sources and sinks of a trace gas are reflected
in the spatial and temporal variation of its mixing ratio in the atmosphere. In
inverse modelling, observed net emissions over a region are apportioned to known
sources and sinks according to apriori assumptions about their relative impor-
tance. The resulting magnitudes of the sources and sinks and their distributions
are then used to calculate the net flux a posteriori using models of atmospheric
transport and chemistry. The agreement between the apriori and a posteriori
values indicates the accuracy of the apriori assumptions.
This approach has been applied to global emissions of CO 2 , CH 4 , N 2 O, halo-
carbons and CO, which have sufficiently long lifetimes and well understood atmo-
spheric chemistries (Heimann and Kaminski, 1999). Methane emissions have
been studied by Hein et al . (1977) and Houweling et al . (1999). van der Gon
et al . (2000) used inverse modelling to test the effects of different apriori
assumptions about the magnitude of CH 4 emissions from rice. In Scenario A a
widely accepted standard range of 50-80 Tg CH 4 year − 1 (Lelieveld et al ., 1998)
is used, and in Scenario B their own best estimate of 15-30 Tg CH 4 year − 1 .The
same total emission is assumed for the two scenarios, and the same combined
flux from wetlands and ricelands. The global emission is apportioned to rice and
other sources according to these assumptions, and calculations made for the whole
globe, the northern and southern hemispheres, and an area roughly correspond-
ing to the part of Asia where rice is most important. The results, summarized
in Table 8.3, show that the assumed and calculated results for the rice area are
much closer for the lower emission scenario, indicating that it is more realistic.

Table 8.3 Global distributions of CH 4 emissions ( Tg CH 4 year − 1 ) calculated using
inverse modelling. In Scenario A rice contributes 50-80 Tg year − 1 and in B
15-30Tg year − 1 ; the net contribution of natural wetlands and ricelands is constant
Globe Asia a

Northern
hemisphere

Southern
hemisphere

Scenario A
assumed 528 ± 90 405 ± 81 123 ± 40 111 ± 56

calculated 505 ± 24 340 ± 19 165 ± 18 77 ± 23

Scenario A
assumed 528 ± 24 384 ± 66 143 ± 38 74 ± 31

calculated 508 ± 24 342 ± 16 166 ± 17 66 ± 18

a 10 ◦ N, 75 ◦ Wto40 ◦ N, 135 ◦ W.
Source : van der Gon et al . (2000). Reproduced with kind permission of Kluwer Academic Publishers.