Browsing by Subject "statistics"
Now showing 1 - 5 of 5
Results Per Page
Sort Options
Item A comparison of clovis caches(Texas A&M University, 2006-04-12) Lassen, Robert DetlefThe Clovis caches in this study consist of assemblages of tools left behind in an area either for future use or as ritual offerings. Clovis caches are the earliest of such assemblages known in North America. This research specifically examines a sample of four caches: East Wenatchee from Douglas County, Washington; Anzick from Park County, Montana; Simon from Camas County, Idaho; and Fenn, inferred to be from Sweetwater County, Wyoming. The artifact types in this study include fluted points, bifaces, blades, flakes, bone rods, and miscellaneous. The variables used in this study include maximum length, mid-length and maximum width, thickness, (length*width*thickness)/1000, length/width, and width/thickness; using millimeters as the basic measurement unit. This study utilizes five methods in the study of the caches: descriptive statistics, factor analysis, cluster analysis, correspondence analysis, and geoarchaeology. The descriptive statistics reveal the most prominent trends that become more apparent in the subsequent statistical analyses. Such trends include East Wenatchee containing the largest points but the smallest bifaces, Anzick and Simon having significant biface variation, Fenn tending to be average in most respects, and bone rods being larger in East Wenatchee than they are in Anzick. The factor analysis explores the relationships between the variables and assigns them to larger components. Length, width, thickness, and length*width*thickness comprise the size component, and length/width and width/thickness make up the shape component. The cluster analysis examines the artifacts within each site and between all sites to identify the most appropriate grouping arrangements based on similarities in artifact measurements. The general results show that fluted points form three clusters according to size more than shape, bifaces are highly variable but have no obvious clusters, and bone rods form three clusters with the first two being strictly divided by site. The correspondence analysis shows that the differences in count data between caches appear to relate to the geographic distances between them. Finally, geoarchaeological analysis posits that East Wenatchee has no discernable pit feature, Anzick contains only one human burial, Simon was not deposited in a pluvial lake, and Fenn would have been shallowly buried but was probably disturbed by erosion.Item An Investigation of Linked Physical And Biogeochemical Processes In Heterogeneous Soils In The Vadose Zone(2012-10-19) Hansen, David JosephChemical dynamics in the vadose zone are poorly understood due to the transient nature of chemical and hydrologic conditions, but are nonetheless critical to understanding contaminant fate and transport. This work explored the effects of soil structure (i.e. layers, lenses) on linked geochemical, hydrological, and microbiological processes under changing hydrologic conditions (e.g. rainfall, introduction of groundwater, and fluctuating water table heights). A homogenized medium-grained sand, homogenized organic-rich loam and a sand-over-loam layered column were constructed for the first series of experiments. The second series of experiments employed two soil columns with lenses that were packed identically with sterilized and untreated sediments. Each consisted of two lenses of organic-rich loam in a medium-grained sand matrix. Lenses were located at different vertical depths and were horizontally offset. In-situ collocated probes collected soil hydrologic and chemical data. In the layered column, enhanced biogeochemical cycling was observed over the texturally homogeneous soil columns. Enumerations of Fe(III) and SO42- reducing microorganisms also show 1-2 orders of magnitude greater community numbers in the layered column. The greatest concentrations of aqueous FeS clusters (FeSaq) were observed in close proximity to the soil interface. To our knowledge, this was the first documentation of FeSaq in partially saturated sediments. Mineral and soil aggregate composite layers were also most abundant near the soil layer interface; the presence of which, likely contributed to an order of magnitude decrease of hydraulic conductivity. In the live lens column, Fe-oxide bands formed at the fringes of the lenses that retarded water flow rates by an order of magnitude compared to the sterilized column. Microbial activity also produced insoluble gases and that led to the creation of a separate gas phase that reduced hydraulic conductivity. This limited the interaction between groundwater with soil-pore waters that led to the formation of geochemically distinct water masses in relatively close proximity to one another. No such changes were observed in the sterilized column. When compared to homogenous columns, the presence of soil heterogeneities altered biogeochemical and hydrologic processes considerably which highlights the need to consider soil heterogeneity in contaminant fate and transport models. These findings suggest that quantifying coupled hydrologic-biogeochemical processes occurring at small scale soil interfaces is critical to accurately describing and predicting chemical changes at the larger system scale.Item Microbial Risk Perspective on the Temporal and Spatial Variability of Indicator Bacteria in Texas Urban and Rural Watersheds(2012-07-16) Srinivasan Ravichandran, SriambharrishThe high incidence of pathogens is one of the main causes for impaired surface water quality designations in the United States. Pathogen presence in fresh water is monitored through the detection of indicator bacteria. Indicator bacteria concentrations, spatial and temporal variability, and microbial risks were evaluated in two rural watersheds, the Bosque and Leon Rivers, and one predominantly urban watershed, the San Jacinto River, all in Texas. Human health risk was predicted from contaminated waters as indicated by ingestion of Escherichia coli found in surface water for contact recreation scenarios. The watersheds were chosen because many segments were previously placed on the 303 (d) list (published by the TCEQ) for failing the indicator bacteria standards. Predominantly urban areas of the San Jacinto River and rural portions of the Bosque and Leon Rivers, where Concentrated Animal Feeding Operations (CAFOs) are numerous, were compared to relatively pristine rural watersheds. Spatial analysis of the watersheds with E.coli concentrations exceeding the single sample (394 MPN/100mL) and the geometric mean standards (126 MPN/100mL) indicated that land use is a significant factor influencing the incidence of bacterial concentrations. Non-agricultural rural areas of the watersheds, such as forests and rangelands, had significantly lower E.coli concentrations compared to the agricultural areas and urban land uses. Human health risk due to ingestion of E.coli as an indicator organism indicated a similar pattern to that of their concentrations in that urban and agricultural areas had a greater risk compared to the other rural areas of the watersheds. The risk estimate for urban and agricultural areas exceeded the acceptable limit of one in ten thousand (10^-4), indicating a potential for adverse health effects to humans. Temporal variability in the watersheds as a function of streamflow, rainfall, and temperature indicated a positive correlation between bacterial concentration and high streamflow, rainfall and temperature. The positive correlation for these effects was greater in the rural areas compared to urban areas, indicating the presence of multiple factors responsible for E.coli concentrations in urban areas. Thus, land use was confirmed to be a major factor contributing to the presence of indicator bacteria in surface waters.Item Radar Nowcasting of Total Lightning over the Kennedy Space Center(2011-08-08) Seroka, Gregory NicholasThe NASA Kennedy Space Center (KSC) is situated along the east coast of central Florida, where a high frequency of lightning occurs annually. Although cloud-to-ground (CG) lightning forecasting using radar echoes has been thoroughly analyzed, few studies have examined intracloud (IC) and/or total (IC CG) lightning. In addition to CG lightning, IC flashes are of great concern to KSC launch operations. Four years (2006-2009) of summer (June, July, August) daytime (about 14-00 Z) Weather Surveillance Radar ? 1988 Doppler data for Melbourne, FL were analyzed. Convective cells were tracked using a modified version of the Storm Cell Identification and Tracking (SCIT) algorithm and then correlated to CG lightning data from the National Lightning Detection Network (NLDN), as well as grouped IC flash data acquired from the KSC Lightning Detection and Ranging (LDAR) networks I and II. Pairs of reflectivity values (30, 35, and 40 dBZ) at isothermal levels (-10, -15, -20 and updraft -10 degrees C), as well as a vertically integrated ice (VII) product were used to optimize criteria for radar-based forecasting of both IC and CG lightning within storms. Results indicate that the best radar-derived predictor of CG lightning according to CSI was 25 dBZ at -20 degrees C, while the best reflectivity at isothermal predictor for IC was 25 dBZ at -15 degrees C. Meanwhile, the best VII predictor of CG lightning was the 30th percentile (0.840 kg m-2), while the best VII predictor of IC was the 5th percentile (0.143 kg m-2), or nearly 6 times lower than for CG! VII at both CG and IC initiation was higher than at both CG and IC cessation. VII was also found to be lower at IC occurrence, including at initiation, than at CG occurrence. Seventy-six percent of cells had IC initiation before CG initiation; using the first IC flash as a predictor of CG occurrence also statistically outperformed other predictors of CG lightning. Even though average lead time for using IC as a predictor of CG was only 2.4 minutes, when taking into account automation processing and radar scan time for the other methods, lead times are much more comparable.Item Sociospatial Inequality: A Multilevel and Geo-Spatial Analysis of Latino Poverty(2012-02-14) Siordia, CarlosSociology at its core has always been interested in understanding how society works. Previous studies on social stratification have sought to outline who gets what, when, and why. This project introduces the where element to advance our understanding of how resource distribution affects life chances. The research question is: Does the percent of Latinos in the area of residence have an influence on Latino?s individual poverty over and above the influence on poverty of the person characteristics? The study ascertains how micro-level inequality is influenced by macro-level attributes and explores how spatial non-stationarity plays a role in these mechanics. This sociospatial inequality investigation will delineate how individual-level stratifying mechanisms are influenced by context-level structural attributes and how sociospatial non-stationary processes play a role in these mechanics. The dissertation is conceptually driven by Hubert M. Blalock's 1970 theory on minority relationships. Blalock posited the testable hypothesis that discrimination against oppressed groups increases when their population rises. Using theoretical propositions inspired by Blalock leads to the testing of the following two formal hypothesis: the multilevel hypothesis (H1) focuses on macro-level effects, I hypothesize that as the percent of Latinos/as in the area of residence increases, the odds of being in poverty will increase for Latinas/os; on the spatial hypothesis (H2), I hypothesize that the statistical association between percent Latina/o and percent poverty is spatially nonstationary. I find that H1 cannot be falsified. The models reveal, as Blalock predicted, that as the percent of Latinos/as in the area of residence increases, the odds of being in poverty increase for Latinas/os (even after controlling for various level-1, level-2, and GWR-level-2 factors). I also find that H2 could not be falsified. I find that the statistical association between percent Latina/o and percent poverty is spatially nonstationary. My multilevel and spatial modeling investigation was unable to falsify Blalock's minority group threat theory. Hierarchical models indicate that as the percent of Latino/a increases, the likelihood of being in poverty for Latinas/os increases. This statically significant relationship holds constant even after spatial nonstationarity level-2 control factors are introduced.