Date of Award

2021-05-01

Degree Name

Doctor of Philosophy

Department

Environmental Science and Engineering

Advisor(s)

Thomas E. Gill

Abstract

In this Dissertation, the impact of drought at different spatiotemporal domains on dust events in the Southern Great Plains and Chihuahuan Desert is investigated. The Moderate Resolution Imaging Spectroradiometer (MODIS) Aerosol Optical Depth (AOD) was evaluated against AOD from AERONET program covering the western USA. The fractional abundance of endmembers, hydrological changes, soil loss, and PM10 dispersion are estimated across the Lordsburg Playa, a dust hot spot in the Chihuahuan Desert.

The dust source points in the Chihuahuan Desert were highly concentrated in areas of severe to extreme drought, while those in the Southern Great Plains span a wider range of drought conditions. There were an increased number of dust events in those regions which were experiencing or recovering from severe to exceptional drought conditions, especially in the southwestern Southern High Plains, west and south of Lubbock, which is a region dominated by non-irrigated cropland. At lower temporal averaging domains, drought showed a strong connection with the dust initiation points in the desert at severe to exceptional drought conditions irrespective of different spatial scenarios, however, dust source points in the plains were linked to lower drought intensities of no drought to abnormally dry. Increasing temporal averaging to 2-years and 5-years, dust events in the desert showed strong connection with moderate to severe drought, while in the plains they were clearly linked to moderate drought at all spatial settings.

The number of AOD retrievals in the southwest USA was on average greater than 14.8 days per 30 days for almost all months except summer, likely triggered by its arid and semi-arid climate. The majority of collocating domains of AOD from the MODIS on Aqua showed a better correlation with AERONET AOD than did AOD from MODIS Terra, and the correlation coefficients exhibited large regional variability across the study area. In general, the majority of the stations revealed significant correlation between MODIS and AERONET AOD at all spatiotemporal aggregating domains, although MODIS generally overestimated AOD compared to AERONET. The correlation coefficient in the southwest United States was the lowest and in the stations from the high latitude was the highest.

A remote sensing study of the surface conditions of Lordsburg Playa also demonstrated that the combination of optical and radar images significantly improved the effort to identify long-term changes of the playa and locations within the Lordsburg Playa susceptible to hydrological and LULC changes. Applying the linear spectral unmixing addressed the limitation of moderate spatial resolution sensors of Landsat and Sentinel-2. Google Earth Engine facilitated the research by minimizing the time required for acquisition, processing, and analysis of images, and the storage required for the larger data size. Modeling wind erosion and dust dispersion from the Lordsburg Playa addressed the limitation of the SWEEP model in simulating soil loss from irregular fields by breaking the playa fields into strips of rectangular subfields and aligning the longest sides with the dominant wind direction. As a result, the North Playa fields emitted 1245.92 and 1664.73 metric tons of PM10 during the 3 February and 5 June 2020 dust days, respectively. The AERMOD model characterized the PM10 dispersion successfully, and identified the timings and locations of dust clouds affecting the major transportation systems crossing the playa during the two dusty days.

Language

en

Provenance

Received from ProQuest

File Size

239 p.

File Format

application/pdf

Rights Holder

Iyasu Gebrezgabher Eibedingil

Download

Share

COinS