HR: 0800h
AN: A11B-0872 [Abstracts]
TI: Influence of Initial and Boundary Conditions on Results of Air Quality Model Simulation
AU: * Liu, T
EM: tliu@ncu.edu.tw
AU: Li, J
EM: ljing@mail.iap.ac.cn
AU: Chang, J S
EM: julius@ncu.edu.tw
AB:
Three-dimensional regional air quality model is mathematically an initial-boundary-value problem and its solution is very
dependent on assumed initial condition (IC) and boundary condition (BC). Since IC and BC are often estimated from few
observations and extrapolations, the inherent uncertainty will influence all model simulation results. It is very important
to know the extent of such artificial uncertainty in influencing model simulated physical realities of interest. We have
systematically conducted a series of numerical experiments to explore this issue. The air quality models applied in this
study were TAQM (Taiwan Air Quality Model, originated from SARMAP Air Quality Model) and its tracer version. Three domains of
different size were designed to illustrate the different distances from the boundaries to interested site. The prevailing
wind of the simulated episode was northeasterly and was simulated by MM5. Influences of IC and BC were estimated by the
difference between the base case and perturbed runs. Arrival times of tracer were estimated by tracer runs and then compared
to those of CO and O3 in runs with full chemistry. Depending on the domain size and distance of the boundary to the site or
area of interest, the arrival time and influence of BC are different. For example, we've found the influence of BC could be
as large as 60% even though the boundary is 3500 kilometers to the north and 2500 kilometers to the east during the
northeasterly episode. These model studies allow us to select an optimal simulation window with the minimum influence of IC
and BC. Such window could be as long as 1 to 2 days but certainly not longer than 3 days for modeling domain considered. It
is recommended such consideration must be applied to all uses of regional air quality model in interpreting observation or in
assessing the air quality scenario.
DE: 3355 Regional modeling
SC: Atmospheric Sciences [A]
MN: Fall Meeting 2005