Horrell, Robin Scott (1991) Aspects of atmospheric transport and dispersion within an air basin. Dissertation (Ph.D.), California Institute of Technology. http://resolver.caltech.edu/CaltechETD:etd-07092007-080911
Atmospheric transport and dispersion and their relationship to observed pollution levels are analyzed using three methods. Two traditional methods, (1) dimensional analysis and (2) atmospheric sulfur hexafluoride tracer experiments are applied to new problems; the third, (3) neural computation is a new method of looking at incomplete ozone data reconstruction, and short term ozone forecasting.
A comprehensive dimensional analysis of buoyantly driven slope wind is presented to unify the subject. The functional relationship between thickness, velocity, Reynolds, Froude, Rossby and Richardson numbers of the buoyant layer and the characteristic slope parameters is established. A detailed study of the velocity and temperature profiles is done using the Von Karman-Pohlhausen (KP) integral technique. Generated profiles are in agreement with field measurements. The technique is extended to handle transient slope flows using a combination of the KP method and the method of characteristics.
The transport and dispersion of airborn emissions from Los Angeles are studied using tracer during the Southern California Air Quality Study of 1987. Emphasis is placed on emissions generated during the morning land breeze/sea breeze transition periods. Analysis of tracer transport patterns indicates emission transport from downtown Los Angeles was by two routes, and the relative importance of each route was determined by a diurnally occurring divergence zone. This zone, resulting from geography and pressure gradients, was shown to migrate east to west in the air space above Los Angeles. The zone routes a majority of morning emissions northwest of Los Angeles, contrary to the commonly held belief that morning emissions were routed east. Results were verified using surface wind trajectories constructed from telemetry data.
Artificial neural computing strategies are applied to two air quality data sets for the purpose of ozone data reconstruction and forecasting at Newhall, California. The Madaline III algorithm is used to reconstruct data at a single site. The reconstruction is done to add missing data. This method is advantageous when peaks in the data are missing, and traditional interpolation methods inapplicable. Williams and Zipser's recurrent algorithm is used to make short-term ozone forecasts at a single location based on past history of ozone observations at that site. The results are systematically better than regression methods.
|Item Type:||Thesis (Dissertation (Ph.D.))|
|Degree Grantor:||California Institute of Technology|
|Division:||Chemistry and Chemical Engineering|
|Major Option:||Chemical Engineering|
|Thesis Availability:||Restricted to Caltech community only|
|Defense Date:||10 January 1991|
|Default Usage Policy:||No commercial reproduction, distribution, display or performance rights in this work are provided.|
|Deposited By:||Imported from ETD-db|
|Deposited On:||23 Jul 2007|
|Last Modified:||26 Dec 2012 02:54|
- Final Version
Restricted to Caltech community only
See Usage Policy.
Repository Staff Only: item control page