Matching Items (8)
Filtering by
- All Subjects: Geography
- All Subjects: spatial statistics
- Genre: Academic theses
- Creators: Brazel, Anthony J.
Description
This dissertation research investigates both spatial and temporal aspects of Bronze Age land use and land cover in the Eastern Mediterranean using botanical macrofossils of charcoal and charred seeds as sources of proxy data. Comparisons through time and over space using seed and charcoal densities, seed to charcoal ratios, and seed and charcoal identifications provide a comprehensive view of island vs. mainland vegetative trajectories through the critical 1000 year time period from 2500 BC to 1500 BC of both climatic fluctuation and significant anthropogenic forces. This research focuses particularly on the Mediterranean island of Cyprus during this crucial interface of climatic and human impacts on the landscape. Macrobotanical data often are interpreted locally in reference to a specific site, whereas this research draws spatial comparisons between contemporaneous archaeological sites as well as temporal comparisons between non-contemporaneous sites. This larger perspective is particularly crucial on Cyprus, where field scientists commonly assume that botanical macrofossils are poorly preserved, thus unnecessarily limiting their use as an interpretive proxy. These data reveal very minor anthropogenic landscape changes on the island of Cyprus compared to those associated with contemporaneous mainland sites. These data also reveal that climatic forces influenced land use decisions on the mainland sites, and provides crucial evidence pertaining to the rise of early anthropogenic landscapes and urbanized civilization.
ContributorsKlinge, JoAnna M (Author) / Fall, Patricia L. (Thesis advisor) / Falconer, Steven E. (Committee member) / Brazel, Anthony J. (Committee member) / Pigg, Kathleen B (Committee member) / Arizona State University (Publisher)
Created2013
Description
Droughts are a common phenomenon of the arid South-west USA climate. Despite water limitations, the region has been substantially transformed by agriculture and urbanization. The water requirements to support these human activities along with the projected increase in droughts intensity and frequency challenge long term sustainability and water security, thus the need to spatially and temporally characterize land use/land cover response to drought and quantify water consumption is crucial. This dissertation evaluates changes in `undisturbed' desert vegetation in response to water availability to characterize climate-driven variability. A new model coupling phenology and spectral unmixing was applied to Landsat time series (1987-2010) in order to derive fractional cover (FC) maps of annuals, perennials, and evergreen vegetation. Results show that annuals FC is controlled by short term water availability and antecedent soil moisture. Perennials FC follow wet-dry multi-year regime shifts, while evergreen is completely decoupled from short term changes in water availability. Trend analysis suggests that different processes operate at the local scale. Regionally, evergreen cover increased while perennials and annuals cover decreased. Subsequently, urban land cover was compared with its surrounding desert. A distinct signal of rain use efficiency and aridity index was documented from remote sensing and a soil-water-balance model. It was estimated that a total of 295 mm of water input is needed to sustain current greenness. Finally, an energy balance model was developed to spatio-temporally estimate evapotranspiration (ET) as a proxy for water consumption, and evaluate land use/land cover types in response to drought. Agricultural fields show an average ET of 9.3 mm/day with no significant difference between drought and wet conditions, implying similar level of water usage regardless of climatic conditions. Xeric neighborhoods show significant variability between dry and wet conditions, while mesic neighborhoods retain high ET of 400-500 mm during drought due to irrigation. Considering the potentially limited water availability, land use/land cover changes due to population increases, and the threat of a warming and drying climate, maintaining large water-consuming, irrigated landscapes challenges sustainable practices of water conservation and the need to provide amenities of this desert area for enhancing quality of life.
ContributorsKaplan, Shai (Author) / Myint, Soe Win (Thesis advisor) / Brazel, Anthony J. (Committee member) / Georgescu, Matei (Committee member) / Arizona State University (Publisher)
Created2014
Description
Accurate characterization of forest canopy cover from satellite imagery hinges on the development of a model that considers the level of detail achieved by field methods. With the improved precision of both optical sensors and various spatial techniques, models built to extract forest structure attributes have become increasingly robust, yet many still fail to address some of the most important characteristics of a forest stand's intricate make-up. The objective of this study, therefore, was to address canopy cover from the ground, up. To assess canopy cover in the field, a vertical densitometer was used to acquire a total of 2,160 percent-cover readings from 30 randomly located triangular plots within a 6.94 km2 study area in the central highlands of the Bradshaw Ranger District, Prescott National Forest, Arizona. Categorized by species with the largest overall percentage of cover observations (Pinus ponderosa, Populus tremuloides, and Quercus gambelii), three datasets were created to assess the predictability of coniferous, deciduous, and mixed (coniferous and deciduous) canopies. Landsat-TM 5 imagery was processed using six spectral enhancement algorithms (PCA, TCT, NDVI, EVI, RVI, SAVI) and three local windows (3x3, 5x5, 7x7) to extract and assess the various ways in which these data were expressed in the imagery, and from those expressions, develop a model that predicted percent-cover for the entire study area. Generally, modeled cover estimates exceeded actual cover, over predicting percent-cover by a margin of 9-13%. Models predicted percent-cover more accurately when treated with a 3x3 local window than those treated with 5x5 and 7x7 local windows. In addition, the performance of models defined by the principal components of three vegetation indices (NDVI, EVI, RVI) were superior to those defined by the principal components of all four (NDVI, EVI, RVI, SAVI), as well as the principal and tasseled cap components of all multispectral bands (bands 123457). Models designed to predict mixed and coniferous percent-cover were more accurate than deciduous models.
ContributorsSchirmang, Tracy Lynn (Author) / Myint, Soe W (Thesis advisor) / Fall, Patricia L. (Thesis advisor) / Brazel, Anthony J. (Committee member) / Arizona State University (Publisher)
Created2012
Description
This research examines lateral separation zones and sand bar slope stability using two methods: a parallelized turbulence resolving model and full-scale laboratory experiments. Lateral flow separation occurs in rivers where banks exhibit strong curvature, for instance canyon rivers, sharp meanders and river confluences. In the Colorado River, downstream Glen Canyon Dam, lateral separation zones are the principal storage of sandbars. Maximum ramp rates have been imposed to Glen Canyon Dam operation to minimize mass loss of sandbars. Assessment of the effect of restricting maximum ramp rates in bar stability is conducted using multiple laboratory experiments. Results reveal that steep sandbar faces would rapidly erode by mass failure and seepage erosion to stable slopes, regardless of dam discharge ramp rates. Thus, continued erosion of sand bars depends primarily of turbulent flow and waves. A parallelized, three-dimensional, turbulence resolving model is developed to study flow structures in two lateral separation zones located along the Colorado River in Grand Canyon. The model employs a Detached Eddy Simulation (DES) technique where variables larger than the grid scale are fully resolved, while Sub-Grid-Scale (SGS) variables are modeled. The DES-3D model is validated using ADCP flow measurements and skill metric scores show predictive capabilities of simulated flow. The model reproduces the patterns and magnitudes of flow velocity in lateral recirculation zones, including size and position of primary and secondary eddy cells and return current. Turbulence structures with a predominately vertical axis of vorticity are observed in the shear layer, becoming three-dimensional without preferred orientation downstream. The DES-3D model is coupled with a sediment advection-diffusion formulation, wherein advection is provided by the DES velocity field minus particles settling velocity, and diffusion is provided by the SGS. Results show a lateral recirculation zone having a continuous export and import of sediment from and to the main channel following a pattern of high frequency pulsations of positive deposition fluxes. These high frequency pulsations play an important role to prevent an oversupply of sediment within the lateral separation zones. Improved predictive capabilities are achieved with this model when compared with previous two- and three-dimensional quasi steady and steady models.
ContributorsAlvarez Rueda, Laura Verónica (Author) / Schmeeckle, Mark W. (Thesis advisor) / Dorn, Ronald I. (Committee member) / Brazel, Anthony J. (Committee member) / Grams, Paul E. (Committee member) / Topping, David J. (Committee member) / Arizona State University (Publisher)
Created2015
Description
Metropolitan Phoenix, Arizona, is one of the most rapidly urbanizing areas in the U.S., which has resulted in an urban heat island (UHI) of substantial size and intensity. Several detrimental biophysical and social impacts arising from the large UHI has posed, and continues to pose, a challenge to stakeholders actively engaging in discussion and policy formulation for a sustainable desert city. There is a need to mitigate some of its detrimental effects through sustainable methods, such as through the application of low-water, desert-adapted low-water use trees within residential yards (i.e. urban xeriscaping). This has the potential to sustainably reduce urban temperatures and outdoor thermal discomfort in Phoenix, but evaluating its effectiveness has not been widely researched in this city or elsewhere. Hence, this dissertation first evaluated peer-reviewed literature on UHI research within metropolitan Phoenix and discerned several major themes and factors that drove existing research trajectories. Subsequently, the nocturnal cooling influence of an urban green-space was examined through direct observations and simulations from a microscale climate model (ENVI-Met 3.1) with an improved vegetation parameterization scheme. A distinct park cool island (PCI) of 0.7-3.6 °C was documented from traverse and model data with larger magnitudes closer to the surface. A key factor in the spatial expansion of PCI was advection of cooler air towards adjacent urban surfaces, especially at 0-1 m heights. Modeled results also possessed varying but reasonable accuracy in simulating temperature data, although some systematic errors remained. Finally, ENVI-Met generated xeriscaping scenarios in two residential areas with different surface vegetation cover (mesic vs. xeric), and examined resulting impacts on near-surface temperatures and outdoor thermal comfort. Desert-adapted low-water use shade trees may have strong UHI mitigation potential in xeric residential areas, with greater cooling occurring at (i.) microscales (~2.5 °C) vs. local-scales (~1.1 °C), and during (ii.) nocturnal (0500 h) vs. daytime periods (1700 h) under high xeriscaping scenarios. Conversely, net warming from increased xeriscaping occurred over mesic residential neighborhoods over all spatial scales and temporal periods. These varying results therefore must be considered by stakeholders when considering residential xeriscaping as a UHI mitigation method.
ContributorsChow, Winston T. L (Author) / Brazel, Anthony J. (Thesis advisor) / Grossman-Clarke, Susanne (Committee member) / Martin, Chris A (Committee member) / Arizona State University (Publisher)
Created2011
Description
Spatial regression is one of the central topics in spatial statistics. Based on the goals, interpretation or prediction, spatial regression models can be classified into two categories, linear mixed regression models and nonlinear regression models. This dissertation explored these models and their real world applications. New methods and models were proposed to overcome the challenges in practice. There are three major parts in the dissertation.
In the first part, nonlinear regression models were embedded into a multistage workflow to predict the spatial abundance of reef fish species in the Gulf of Mexico. There were two challenges, zero-inflated data and out of sample prediction. The methods and models in the workflow could effectively handle the zero-inflated sampling data without strong assumptions. Three strategies were proposed to solve the out of sample prediction problem. The results and discussions showed that the nonlinear prediction had the advantages of high accuracy, low bias and well-performed in multi-resolution.
In the second part, a two-stage spatial regression model was proposed for analyzing soil carbon stock (SOC) data. In the first stage, there was a spatial linear mixed model that captured the linear and stationary effects. In the second stage, a generalized additive model was used to explain the nonlinear and nonstationary effects. The results illustrated that the two-stage model had good interpretability in understanding the effect of covariates, meanwhile, it kept high prediction accuracy which is competitive to the popular machine learning models, like, random forest, xgboost and support vector machine.
A new nonlinear regression model, Gaussian process BART (Bayesian additive regression tree), was proposed in the third part. Combining advantages in both BART and Gaussian process, the model could capture the nonlinear effects of both observed and latent covariates. To develop the model, first, the traditional BART was generalized to accommodate correlated errors. Then, the failure of likelihood based Markov chain Monte Carlo (MCMC) in parameter estimating was discussed. Based on the idea of analysis of variation, back comparing and tuning range, were proposed to tackle this failure. Finally, effectiveness of the new model was examined by experiments on both simulation and real data.
In the first part, nonlinear regression models were embedded into a multistage workflow to predict the spatial abundance of reef fish species in the Gulf of Mexico. There were two challenges, zero-inflated data and out of sample prediction. The methods and models in the workflow could effectively handle the zero-inflated sampling data without strong assumptions. Three strategies were proposed to solve the out of sample prediction problem. The results and discussions showed that the nonlinear prediction had the advantages of high accuracy, low bias and well-performed in multi-resolution.
In the second part, a two-stage spatial regression model was proposed for analyzing soil carbon stock (SOC) data. In the first stage, there was a spatial linear mixed model that captured the linear and stationary effects. In the second stage, a generalized additive model was used to explain the nonlinear and nonstationary effects. The results illustrated that the two-stage model had good interpretability in understanding the effect of covariates, meanwhile, it kept high prediction accuracy which is competitive to the popular machine learning models, like, random forest, xgboost and support vector machine.
A new nonlinear regression model, Gaussian process BART (Bayesian additive regression tree), was proposed in the third part. Combining advantages in both BART and Gaussian process, the model could capture the nonlinear effects of both observed and latent covariates. To develop the model, first, the traditional BART was generalized to accommodate correlated errors. Then, the failure of likelihood based Markov chain Monte Carlo (MCMC) in parameter estimating was discussed. Based on the idea of analysis of variation, back comparing and tuning range, were proposed to tackle this failure. Finally, effectiveness of the new model was examined by experiments on both simulation and real data.
ContributorsLu, Xuetao (Author) / McCulloch, Robert (Thesis advisor) / Hahn, Paul (Committee member) / Lan, Shiwei (Committee member) / Zhou, Shuang (Committee member) / Saul, Steven (Committee member) / Arizona State University (Publisher)
Created2020
Description
Geographically Weighted Regression (GWR) has been broadly used in various fields to
model spatially non-stationary relationships. Classic GWR is considered as a single-scale model that is based on one bandwidth parameter which controls the amount of distance-decay in weighting neighboring data around each location. The single bandwidth in GWR assumes that processes (relationships between the response variable and the predictor variables) all operate at the same scale. However, this posits a limitation in modeling potentially multi-scale processes which are more often seen in the real world. For example, the measured ambient temperature of a location is affected by the built environment, regional weather and global warming, all of which operate at different scales. A recent advancement to GWR termed Multiscale GWR (MGWR) removes the single bandwidth assumption and allows the bandwidths for each covariate to vary. This results in each parameter surface being allowed to have a different degree of spatial variation, reflecting variation across covariate-specific processes. In this way, MGWR has the capability to differentiate local, regional and global processes by using varying bandwidths for covariates. Additionally, bandwidths in MGWR become explicit indicators of the scale at various processes operate. The proposed dissertation covers three perspectives centering on MGWR: Computation; Inference; and Application. The first component focuses on addressing computational issues in MGWR to allow MGWR models to be calibrated more efficiently and to be applied on large datasets. The second component aims to statistically differentiate the spatial scales at which different processes operate by quantifying the uncertainty associated with each bandwidth obtained from MGWR. In the third component, an empirical study will be conducted to model the changing relationships between county-level socio-economic factors and voter preferences in the 2008-2016 United States presidential elections using MGWR.
model spatially non-stationary relationships. Classic GWR is considered as a single-scale model that is based on one bandwidth parameter which controls the amount of distance-decay in weighting neighboring data around each location. The single bandwidth in GWR assumes that processes (relationships between the response variable and the predictor variables) all operate at the same scale. However, this posits a limitation in modeling potentially multi-scale processes which are more often seen in the real world. For example, the measured ambient temperature of a location is affected by the built environment, regional weather and global warming, all of which operate at different scales. A recent advancement to GWR termed Multiscale GWR (MGWR) removes the single bandwidth assumption and allows the bandwidths for each covariate to vary. This results in each parameter surface being allowed to have a different degree of spatial variation, reflecting variation across covariate-specific processes. In this way, MGWR has the capability to differentiate local, regional and global processes by using varying bandwidths for covariates. Additionally, bandwidths in MGWR become explicit indicators of the scale at various processes operate. The proposed dissertation covers three perspectives centering on MGWR: Computation; Inference; and Application. The first component focuses on addressing computational issues in MGWR to allow MGWR models to be calibrated more efficiently and to be applied on large datasets. The second component aims to statistically differentiate the spatial scales at which different processes operate by quantifying the uncertainty associated with each bandwidth obtained from MGWR. In the third component, an empirical study will be conducted to model the changing relationships between county-level socio-economic factors and voter preferences in the 2008-2016 United States presidential elections using MGWR.
ContributorsLi, Ziqi (Author) / Fotheringham, A. Stewart (Thesis advisor) / Goodchild, Michael F. (Committee member) / Li, Wenwen (Committee member) / Arizona State University (Publisher)
Created2020
Description
This dissertation research studies long-term spatio-temporal patterns of surface urban heat island (SUHI) intensity, urban evapotranspiration (ET), and urban outdoor water use (OWU) using Phoenix metropolitan area (PMA), Arizona as the case study. This dissertation is composed of three chapters. The first chapter evaluates the SUHI intensity for PMA using Moderate Resolution Imaging Spectroradiometer (MODIS) land surface temperature (LST) product and a time-series trend analysis to discover areas that experienced significant changes of SUHI intensity between 2000 and 2017. The heating and cooling effects of different urban land use land cover (LULC) types was also examined using classified Landsat satellite images. The second chapter is focused on urban ET and the impacts of urban LULC change on ET. An empirical model of urban ET for PMA was built using flux tower data and MODIS land products using multivariate regression analysis. A time-series trend analysis was then performed to discover areas in PMA that experienced significant changes of ET between 2001 and 2015. The impact of urban LULC change on ET was examined using classified LULC maps. The third chapter models urban OWU in PMA using a surface energy balance model named METRIC (Mapping Evapotranspiration at high spatial Resolution with Internalized Calibration) and time-series Landsat Thematic Mapper 5 imagery for 2010. The relationship between urban LULC types and OWU was examined with the use of very high-resolution land cover classification data generated from the National Agriculture Imagery Program (NAIP) imagery and regression analysis. Socio-demographic variables were selected from census data at the census track level and analyzed against OWU to study their relationship using correlation analysis. This dissertation makes significant contributions and expands the knowledge of long-term urban climate dynamics for PMA and the influence of urban expansion and LULC change on regional climate. Research findings and results can be used to provide constructive suggestions to urban planners, decision-makers, and city managers to formulate new policies and regulations when planning new constructions for the purpose of sustainable development for a desert city.
ContributorsWang, Chuyuan (Author) / Myint, Soe W. (Thesis advisor) / Brazel, Anthony J. (Committee member) / Wang, Zhihua (Committee member) / Hondula, David M. (Committee member) / Arizona State University (Publisher)
Created2018