P44 Sensitivity experiments of urban canopy parameterizations using ensemble WRF simulations over the Chicago metropolitan.
Sever, Gökhan, Rob Jacob, Rajeev Jain, Aleks Obabko, Rao Kotamarthi, and Charlie Catlett, Argonne National Laboratory
Forecasting the
weather at urban-scales requires the development of models that are capable of
running at a few cubic meters resolution as well as physical representations of
the atmosphere that are verified at scales relevant to buildings. The
Multiscale Coupled Urban Systems exascale computing project has designed
possible pathways to handle the large computational and data requirements of
such future modeling infrastructure. While similar efforts have shown rapid
progression in recent years, observational data that captures the high
spatio-temporal pulse of weather that models are designed upon and verified
against have not scaled up with the computational advancements. Synergistic
efforts between the computational and observational fronts are essential to
making the forecasting of weather a reality at building scales.
We performed High-Resolution Rapid Refresh (HRRR) initialized Weather Research
and Forecasting (WRF) model simulations over the Chicago region for the
entirety of 2018. Three kilometers to a few hundreds of meters horizontal
resolution configurations allow the model to capture meso and micro scale
weather evolution from hours to seconds. The impact of buildings are treated
with multi-layer urban canopy parameterizations with different land-surface and
planetary boundary layer schemes. Basic atmospheric measurements obtained with
the Array-of-Things sensors will be used to validate model results within the
metropolitan area. In addition, we plan to verify parameterized urban effects
using an explicit building resolving model. The ensemble simulations will
provide a basis to better understand mechanical and thermal processes within
complex atmospheric boundary layer of Chicago which are locally affected by
buildings, urban heat island, and land-sea breeze forcings.