 |
| Titel |
Parametric uncertainties in global model simulations of black carbon column mass concentration |
| VerfasserIn |
Hana Pearce, Lindsay Lee, Carly Reddington, Ken Carslaw, Graham Mann |
| Konferenz |
EGU General Assembly 2016
|
| Medientyp |
Artikel
|
| Sprache |
en
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 18 (2016) |
| Datensatznummer |
250132955
|
| Publikation (Nr.) |
 EGU/EGU2016-13510.pdf |
|
|
|
|
|
| Zusammenfassung |
| Previous studies have deduced that the annual mean direct radiative forcing from black
carbon (BC) aerosol may regionally be up to 5 W m−2 larger than expected due to
underestimation of global atmospheric BC absorption in models. We have identified the
magnitude and important sources of parametric uncertainty in simulations of BC column
mass concentration from a global aerosol microphysics model (GLOMAP-Mode). A
variance-based uncertainty analysis of 28 parameters has been performed, based on statistical
emulators trained on model output from GLOMAP-Mode. This is the largest number of
uncertain model parameters to be considered in a BC uncertainty analysis to date and covers
primary aerosol emissions, microphysical processes and structural parameters related to the
aerosol size distribution.
We will present several recommendations for further research to improve the fidelity of
simulated BC. In brief, we find that the standard deviation around the simulated
mean annual BC column mass concentration varies globally between 2.5 x 10−9 g
cm−2 in remote marine regions and 1.25 x 10−6 g cm−2 near emission sources
due to parameter uncertainty Between 60 and 90% of the variance over source
regions is due to uncertainty associated with primary BC emission fluxes, including
biomass burning, fossil fuel and biofuel emissions. While the contributions to BC
column uncertainty from microphysical processes, for example those related to
dry and wet deposition, are increased over remote regions, we find that emissions
still make an important contribution in these areas. It is likely, however, that the
importance of structural model error, i.e. differences between models, is greater than
parametric uncertainty. We have extended our analysis to emulate vertical BC profiles at
several locations in the mid-Pacific Ocean and identify the parameters contributing
to uncertainty in the vertical distribution of black carbon at these locations. We
will present preliminary comparisons of emulated BC vertical profiles from the
AeroCom multi-model ensemble and Hiaper Pole-to-Pole (HIPPO) observations. |
|
|
|
|
|
|