4.5. Selection of Model Runs for UV (Chapter 5)
and Climate Impact (Chapter 6) Scenario Studies
In this section, the model results that are used as a guide for the ultraviolet
B (UV-B) impact (Chapter 5) and the estimation of climate
impact (Chapter 6) are selected and uncertainties are assigned
in light of the discussion in Section 4.4. A key uncertainty
in assessing the impact of subsonic and supersonic exhaust emissions is that all
models used to assess the aircraft impact are either tropospheric models with
limited representation of stratospheric chemistry and transport (tropospheric
3-D CTMs), or stratospheric 2-D and 3-D CTMs with limited representation of tropospheric
processes. In addition, neither the 2-D nor the 3-D models have been designed
to deal specifically with transport processes in the tropopause region, where
most aircraft emissions take place.
4.5.1. Model Simulations of Subsonic Aircraft
How well can we calculate the impact from subsonic aircraft in the tropopause
region using the tropospheric 3-D CTMs or the stratospheric 2-D models? For
these aircraft, part of the perturbation occurs in the LS and part in the UT.
As discussed in Chapter 2, based on the limited set of
observational data that could be chosen for model validation, none of the 3-D
CTMs could be picked as the best assessment model for impact studies of future
aircraft emissions. The UiO 3-D model was selected as a representative model
for the UV (Chapter 5) and climate impact studies (Chapter
6), because it gave results in the middle range of model results and it
was easily available for sensitivity studies. Height profiles for 30-60°N zonal
mean O3 perturbations for 1992 subsonic emissions
(Scenario B-A in Table 4-4) obtained with the AER
2-D model and the UiO 3-D model are shown in Figure 4-10a.
There is reasonable agreement between the two models in the 8-12 km region,
near the tropopause. The UiO 3-D model, whose chemistry is most suited to the
troposphere, shows a smaller O3 perturbation
in the middle and lowest troposphere. The AER 2-D model, whose chemistry is
most suited to the stratosphere, shows a larger O3
perturbation in the LS. An uncertainty range of a factor of 2 was adopted for
O3 perturbation from future subsonic aircraft
emissions. This uncertainty range was based on the range of model results obtained
by participating models in basic perturbation studies and results from the limited
number of sensitivity studies. A "fair" confidence is associated with this uncertainty
range for 2015, and a "poor" confidence is associated with this uncertainty
range for 2050.
Figure 4-10b shows zonal mean O3 changes for five
calculations from the UiO 3-D model. As illustrated in the figure, predicted
changes in O3 from subsonic aircraft are comparable to changes from surface
sources in 2015 and 2050.
4.5.2. Model Simulations of Supersonic Aircraft
The basic assumption in market studies that determine the routing and size
of the supersonic fleet is that the supersonic fleet will replace certain routes
of the given subsonic fleet, corresponding to about 10% of the subsonic fuel
burn. The results of these studies were used to generate the fuel burn for a
combined fleet consisting of a supersonic fleet with a modified subsonic fleet.
For this reason, it is more appropriate to compare the effect of the combined
fleet to the subsonic fleet, rather than to look at the supersonic fleet in
isolation. Taking 2015 as an example, numerical results were generated for the
2015 atmosphere without aircraft (scenario C), the 2015 atmosphere with the
standard subsonic fleet (scenario D), and the 2015 atmosphere with the combined
fleet (scenario S1k). Recognizing the uncertainties concerning the tropospheric
response generated by the stratospheric models, the strategy is to use the change
in O3 computed between S1k and D and the results from the tropospheric models
from scenario D minus scenario C for the effect of the subsonic fleet, after
adjusting for the 10% difference in fuel burn.
For supersonic aircraft influences on UV effects (Chapter
5) and radiative forcing and climate change (Chapter 6),
we have chosen a central or most probable emission scenario (S1k-2015 and S9h-2050),
along with a representative assessment model (AER). This emission scenario assumes
a SO2 gas-to-particle conversion of 10%. The AER model was selected because
it was the model that calculated and supplied the enhanced gas-to-particle sulfate
aerosol SAD that was used in all participating assessment models. The middle
plot in Figure 4-11 shows the percentage change in
column O3 for the AER model using the S1k scenario. At all latitudes and months,
the AER model derives a reduction in total column O3.
Because of coupling between chemistry and transport, there is no simple way
to scale O3 change profiles based on a priori estimates of uncertainties. It
is possible to get an idea of the uncertainties by looking at the assembly of
results from various scenarios performed by different models. Concentrating
on the first group of scenarios in Table 4-11 (which
represent a 500-plane Mach 2.4 fleet with EI(NOx)=5 flying in a clean sulfate
background), we picked S1c from UNIVAQ and S1h from GSFC as representative of
the range of possible results. The differences in computed O3 changes by the
AER, UNIVAQ, and GSFC models in northern mid-latitudes are illustrated in Figure
4-12a for 2015. The UNIVAQ and GSFC models were picked because these models
represent different ways of treating transport and PSCs that consistently produce
the smallest and largest O3 depletion in most scenarios. The sampling of scenarios
also covers the possible range of effects from different assumptions of gas-to-particle
conversion in plume processing of SO2 emission. For 2050, we focus on a 1,000-plane
Mach 2.4 fleet with EI(NOx)=5. The AER S9h scenario is taken as the central
case, and the lower and upper extremes were taken to be S9d for UNIVAQ and S9f
for GSFC. The differences in O3 changes computed by the AER, UNIVAQ, and GSFC
models in northern mid-latitudes are illustrated in Figure
4-12b for 2050.
It is also important to note which uncertainties, among those discussed in
Chapters 2 and 3, are not included
in this range. All of the models used rate data from DeMore et al. (1997). A
previous study (described in Stolarski et al., 1995) showed that uncertainties
in rate data could lead to an uncertainty in NH O3 column change of ±1%. Current
studies indicate that most models underestimate the mean age of air as defined
by inert tracers that enter the stratosphere via the tropical tropopause. Given
that there appears to be a positive correlation between calculated increases
in NOy and H2O from HSCT and calculated mean age, it has been suggested that
models that underestimate age will also underestimate NOy and H2O increases
from HSCT. It is difficult to quantify the uncertainty given current information.
We did not consider the effect of plume processing and possible changes (in
temperature and transport circulation) in the future background atmosphere.
Finally, the range cited does not include different technology options for different
EI(NOx), different cruise altitudes, and different fleet sizes.
It is possible to arrive at a subjective estimate for uncertainty estimates
for changes in column O3. This value can be used to calculate changes in UV
because that depends mostly on the changes in column in the lower stratosphere.
It is less obvious whether this value can be used to estimate uncertainties
in radiative forcing. Restricting ourselves to the 1,000-plane Mach 2.4 fleet
with EI(NOx)=5, the range of model results for annual averaged Northern Hemisphere
column O3 depletion is -0.1% to -1.4% (see Table 4-12).
Other model studies lead us to believe that reasonable changes in the background
atmosphere (including background sulfate surface area) would not change this
range in a significant way. Uncertainties in rate data would expand the range
to about +1 to -2.5%. One may also argue that the inability of the models to
simulate the correct mean age may add to uncertainties on the negative side
by another 1%. Thus, a subjective estimate is that actual atmospheric response
would likely lie between +1 and -3.5%.
A 500-plane Mach 2.4 fleet with EI(NOx)=5 would have a similar range of uncertainty.
Here, the range of model results for annual averaged Northern Hemisphere column
O3 depletion is 0 to -1.3% (see Table 4.11)-very
close to the -0.1 to -1.4% model range for 1,000 planes. Applying the same arguments
for this 500-plane fleet, a subjective estimate for the actual atmospheric response
would likely be in the range of +1 to -3.5%. We have "fair" confidence in this
range for the 500-plane and 1,000-plane fleets.