Show simple item record

dc.contributor.advisorJacob, Daniel J.
dc.contributor.authorYu, Karen
dc.date.accessioned2019-05-17T14:17:09Z
dc.date.created2017-11
dc.date.issued2017-07-12
dc.date.submitted2017
dc.identifier.urihttp://nrs.harvard.edu/urn-3:HUL.InstRepos:39987868*
dc.description.abstractAccurate simulation of the atmosphere has important implications for climate, air quality, and human health. With increasing computational resources, the complexity of atmospheric models have also increased, resolving more physical and chemical processes that happen on smaller temporal and spatial scales. In this work, I address some of the issues associated with modeling interacting physical and chemical processes that occur on widely different spatiotemporal scales, the tradeoffs between model complexity computational expense, and methods of alleviating some of these errors. Chemical production of ozone and organic aerosols is highly non-linear, which can lead to errors when the subgrid scale variablity in their precursors are not captured. Using aircraft observations of isoprene and NOx over the Southeast US, we find that high-resolution models better capture the spatial segregation of emissions and the subsequent chemical oxidation pathways of these emitted chemicals. However, the cumulative probability distribution functions (CDFs) of NOx, isoprene, and ozone concentrations show little difference across model resolutions and good agreement with observations, suggesting that smaller-scale non-linearities are not important on the regional scale. Off-line chemical transport models (CTMs) use archived meteorological data from a general circulation model (GCM) to drive transport. To reduce computational load, the CTM simulation is often conducted on a coarser grid than the GCM, requiring both spatial and temporal averaging of the meteorological data. This results in significantly reduced vertical transport in the CTM, due to transient organized vertical motions in the GCM (resolved convection) being averaged out. Additional post-processing of the meteorological archive, in particular remapping from cubed-sphere to rectilinear grid also leads to significant errors. Re-computing the convective mass fluxes at the resolution of the CTM and adjusting for bias in the averaged boundary layer heights can partly restore the lost vertical transport in the coarse-resolution CTM simulation.
dc.description.sponsorshipEngineering and Applied Sciences - Engineering Sciences
dc.format.mimetypeapplication/pdf
dc.language.isoen
dash.licenseLAA
dc.subjectAtmospheric Sciences
dc.titleScale issues in atmospheric chemistry modeling
dc.typeThesis or Dissertation
dash.depositing.authorYu, Karen
dc.date.available2019-05-17T14:17:09Z
thesis.degree.date2017
thesis.degree.grantorGraduate School of Arts & Sciences
thesis.degree.levelDoctoral
thesis.degree.nameDoctor of Philosophy
dc.contributor.committeeMemberWofsy, Steven C.
dc.contributor.committeeMemberKuang, Zhiming
dc.type.materialtext
thesis.degree.departmentEngineering and Applied Sciences - Engineering Sciences
dash.identifier.vireohttp://etds.lib.harvard.edu/gsas/admin/view/1759
dc.description.keywordsatmospheric chemistry modeling; chemical transport modeling; atmospheric chemistry; model resolution; scale issues; grid resolution; southeast us; air quality; offline model
dc.identifier.orcid0000-0003-1307-3738
dash.author.emailkyu0110@gmail.com


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record