Master of Arts, Miami University, 2024, Geography

Rapid urbanization and the Urban Heat Island (UHI) effect are key global challenges, particularly in the Global South. Dhaka City, Bangladesh, one of the fastest growing and most densely populated cities in this region, facing increasing heat trends which threaten its public health, social and environmental conditions. Mitigating future UHI requires a geospatial analysis of urbanization and temperature trends at the neighborhood-scale. However, proprietary geospatial data and applications are often prohibitively costly for planners. This study combines cost-effective, publicly-accessible geospatial applications and time series satellite data from 2003 - 2023 to analyze land use and land cover (LULC) and land surface temperature (LST) changes in Dhaka City at regional and neighborhood-scales. For monitoring LULC, Landsat images were analyzed through supervised classification at a regional scale, and spectral mixture analysis (SMA) helped understand complex urban development patterns at the neighborhood-scale. This study analyzed MODIS daily LST images to understand diurnal temperature trends and found a strong positive correlation between urban development intensity and increased day and nighttime temperatures, contributing to neighborhood-specific UHI impacts. The study emphasizes the importance of developing publicly-accessible and inexpensive geospatial methods to support UHI mitigation planning that can benefit other similar cities.

Committee: David Prytherch Dr. (Advisor); Jessica McCarty Dr. (Committee Member); Robbyn Abbitt Ms. (Committee Member); John Maingi Dr. (Committee Member) Subjects: Climate Change; Environmental Studies; Geographic Information Science; Geography; Remote Sensing; Urban Planning

Doctor of Philosophy, The Ohio State University, 2020, Geography

Remote Sensing offers a unique perspective of our Earth and is crucial for managing its resources. Currently, there is no single satellite data product that is suitable for all applications. Satellite data are limited by their spatial, spectral, and temporal resolution. Additionally, satellite images can be affected by sensor noise and cloud cover. One of the solutions to overcome these limitations is by combining existing satellite products to minimize the drawbacks of a dataset. In this dissertation, we improve the spatial and temporal resolution of satellite data products, minimize sensor noise, and remove cloud cover from satellite images by combining data from multiple satellite sensors using deep learning methods. Deep learning has been successful in natural image superresolution, denoising, and translation and these methods perform efficiently given sufficiently large datasets and computational resources. Therefore, publicly available satellite datasets and recent computational advancements provide an ideal opportunity for applying deep learning for our tasks. In our first study, we downscale low resolution optical and thermal spectral bands of MODIS to match higher resolution NIR and Red bands. Information extraction from satellite data often requires the combined use of multiple spectral bands. Usually, the low-resolution bands are downscaled using naive interpolation methods or high-resolution bands are upscaled to create spectral indices. We train a deep learning model for downscaling MODIS spectral to create a spatially consistent MODIS dataset. Our model is compared to a state-of-the-art satellite image downscaling method and a deep learning image superresolution method. Additionally, we investigate the importance of prior natural images towards downscaling satellite images. Next, we increase the effective spatial resolution and denoise MODIS spectral bands with the help of Landsat 8 images. MODIS and Landsat 8 have similar measurement principles and (open full item for complete abstract)

Committee: Desheng Liu Dr (Advisor); Alvaro Montenegro Dr (Committee Member); Srinivasan Parthasarathy Dr (Committee Member); Rongjun Qin Dr (Committee Member) Subjects: Geographic Information Science; Geography; Remote Sensing

Master of Science, The Ohio State University, 2019, Environmental Science

Quantifying and understanding the variability in vegetation gross primary production can help us understand vegetation's response to climate change and predict the climate system. This thesis first compares the performance of newly developed vegetation index—Near-infrared Reflectance of Vegetation (hereafter, NIRV)—with two traditional vegetation indices (hereafter, VIs) based on eddy covariance flux tower. It turns out that NIRV performs poorly in capturing growing season length but may be good at tracking vegetation productivity at the annual scale. Next, parametric and non-parametric regression models are developed for estimating annually integrated GPP (hereafter, AIGPP). Results show that random forest models achieve lower out-of-sample error based on cross-validation. The optimal random forest model is applied to characterize the decadal changes of AIGPP in Alaska, where 6.17% of the area showed significant greening while 2.10% of the area showed significant browning at the 95% confidence level. Trend analysis has also been conducted for vegetation traits—growing season length (hereafter, GSL) and peak value (hereafter, Peak). Peak shows similar trend patterns with AIGPP, while GSL shows limited significant trends in Alaska, which contrasts with previous studies and calls for further study. These findings suggest that physiological changes of vegetation may dominant the AIGPP changes in Alaska during 2003-2014.

Committee: Desheng Liu (Advisor); Gil Bohrer (Committee Member); Bryan Mark (Committee Member) Subjects: Environmental Science

PHD, Kent State University, 2018, College of Arts and Sciences / Department of Earth Sciences

Remote sensing instruments are powerful tools for discriminating between color producing agents (CPAs) that comprise the complex optical signal of Lake Erie. The western basin of Lake Erie (WBLE) is affected by annually recurring cyanobacteria and harmful algal blooms (CyanoHABs), which impact water quality, industry and tourism. In addition to cyanobacteria and algae, sediment and colored dissolved organic material contribute to the optical reflectance signal of Lake Erie waters. CyanoHABs are driven by nutrient input from the Maumee River, and grow in size from spring to fall. Due to the potential of cyanobacteria to produce toxins detrimental to humans and wildlife, identification of the constituents in the CyanoHABs is vital. Both airborne and satellite-based remote sensing instruments provide detailed spectral and spatial information at different scales, and are effective tools for assessing CyanoHABs in Lake Erie. Four images acquired in July of the 2015 CyanoHAB by the Moderate Resolution Imaging Spectroradiometer (MODIS), onboard the NASA Aqua satellite, are analyzed using the Kent State Varimax-rotated principal component analysis (VPCA) spectral decomposition method. MODIS provides basin-wide imagery with a 1 km2 pixel resolution. Four primary signals were extracted, identifying the CyanoHAB signal, a clay sediment and chlorophyll degradation product signal, a hematite and chlorophyll a signal, and a cryptophya signal. The results of the VPCA spectral decomposition of the MODIS showed good agreement with both in situ measured chlorophyll a values in the WBLE, and with the VPCA spectral decomposition results of the reflectance spectra of field samples measured in the lab. The HyperSpectral Imager v2 (HSI2) is an airborne instrument operated by a research team at NASA Glenn, and flown during the 2016 CyanoHAB season. Four images collected on 21 June 2016 and three images collected on 13 September 2016 were analyzed. These two sets of images are of near (open full item for complete abstract)

Committee: Joseph Ortiz (Advisor); Anne Jefferson (Committee Member); Elizabeth Herndon (Committee Member); Adem Ali (Committee Member) Subjects: Geology; Hydrology; Remote Sensing

Master of Arts, University of Toledo, 2017, Geography

In the United States of America, several studies have quantified evapotranspiration (ET) in different ecosystems at a local scale. Accurate and spatially explicit information on ET is rare in small cities like Toledo mainly due to the lack of appropriate tools. A remote sensing ET product from the MODerate Resolution Imaging Spectrometer (MOD16) has been developed. However, its accuracy is not known in and around the ecosystems of Toledo. The main objective of this study was to compare the MOD16 ET product using the data from eddy covariance based Flux Tower of The University of Toledo. Eight day and monthly cumulative ET was calculated from the Flux Tower data to coincide with the MOD16 product over the year 2014. The Flux Tower results showed significant number of missing values in the period of June 26 to September 19 and October 29 to November 13. The 8-day MOD16 ET was not compared for the inconsistencies in seasonal variations. Due to data gaps, only monthly ET of Flux measurement from January to June was compared with the MOD16 ET of all selected places. All places were shown positive relation with Flux observation at the R2 level of more than 0.5 which means that more than 50% data were explained the best fit for the interrelationship among them. The appropriateness of the MOD16 and Flux Tower based ET comparison can be attributed to, among other things, (1) the parameterization of the Penman-Monteith model, (2) Flux Tower measurement errors, and (3) Flux Tower footprint vs. MODIS pixel coverage. MOD16 is important for global inference of ET, but to use it for the water management of Toledo, a locally parameterized and improved product should be developed.

Committee: Kevin Czajkowski (Committee Chair); Patrick Lawrence (Committee Member); Bhuiyan Alam (Committee Member) Subjects: Geography

Doctor of Philosophy (PhD), Ohio University, 2014, Chemical Engineering (Engineering and Technology)

Emissions from local industries, particularly coal-fired power plants, have been shown to enhance the ambient pollutant budget in the Ohio River Valley (ORV) region. One pollutant that is of interest is PM2.5 due to its established link to respiratory illnesses, cardiopulmonary diseases and mortality. State and local agencies monitor the impact of the local point sources on the ambient concentrations at specific sites; however, the monitors do not provide satisfactory spatial coverage. An important metric for describing ambient particulate pollution is aerosol optical depth (AOD). It is a dimensionless geo-physical product measured remotely using satellites or ground-based light detection ranging instruments. This study focused on assessing the effectiveness of using satellite aerosol optical depth (AOD) as an indicator for PM2.5 in the ORV and two cities in Ohio. Three models, multi-linear regression (MLR), principal component analysis (PCA) – MLR and neural network, were trained using 40% of the total dataset. The outcome was later tested to minimize error and further validated with another 40% of the dataset not included in the model development phase. Furthermore, to limit the effect of seasonality, four models representing each season were created for each city using meteorological variables known to influence PM2.5 and AOD concentration. GIS spatial analysis tool was employed to visualize and make spatial and temporal comparisons for the ORV region. Comparable spatial distributions were observed. Regression analysis showed that the highest and lowest correlations were in the summer and winter, respectively. Seasonal decomposition methods were used to evaluate trends at local Ohio monitoring stations to identify areas most suitable for improved air quality management. Over the six years of study, Cuyahoga County maintained PM2.5 concentrations above the national standard and in Hamilton County (Cincinnati) PM2.5 levels ranked above the national level fo (open full item for complete abstract)

Committee: Kevin Crist Ph.D. (Advisor) Subjects: Atmosphere; Engineering

Master of Science (MS), Wright State University, 2008, Earth and Environmental Sciences

This study used sap flow measurements and satellite imagery to estimatewater use by cottonwood (Populus fremontii S. Wats. ssp) trees in an irrigated restoration plot at Cibola National Wildlife Refuge on the Lower Colorado River. Several thousand hectares of irrigated plots of this type are planned to improve riparian habitat on the river, hence it is important to know how much water the trees require. In this study, the ET rates for 20 Freemont cottonwood trees, from an 8 ha plot, were monitored over a 30-day period. ET rates were estimated by measuring sap flow through branches of the trees. Biometric scaling was used to project ET at branch to ET at tree and plot level through the ratio of basal trunk area with the cross-sectional area of the branches. The mean biometric ratio exhibited a 1:1 relationship. Sap flow ET results showed that the cottonwood tree consumed 6-11 mm day-1 of water. My main contribution in this project was working with vegetation indices from MODIS and Landsat 5 TM (TM) time-series imagery and air temperature data. I developed projected ET rates over annual cycles, based on an empirical method calibrated against moisture flux tower data in previous studies. ET estimates from satellite data were similar to concurrent measurements of ET by sap flow methods. Annual estimates of ET from satellite data were approximately 1,200 mm yr-1, with an error or uncertainty of 20-30% inherent in both the ground and remote sensing methods.

Committee: Doyle Watts PhD (Committee Chair); Allen Burton PhD (Committee Member); Pamela Nagler PhD (Committee Member); Subramania Sritharan, PhD (Committee Member) Subjects: Agriculture; Biogeochemistry; Botany; Earth; Environmental Science; Geology; Geophysics

MS, University of Cincinnati, 2005, Engineering : Environmental Engineering

Many water resources and agricultural applications require the knowledge of evapotranspiration (ET) over a range of spatial and temporal scales. Satellite remote sensing provides an unprecedented spatial coverage of land surface and atmospheric data that are logistically and economically impossible to obtain through ground based observation networks. This study is focused on ascertaining the feasibility and robustness of using newly launched sensors for the continuous monitoring capability of ET for hydrology, climatology, agronomy and ecology studies. The estimation of ET is based on a modified Priestley-Taylor equation developed by Jiang & Islam (2001), where all parameters can be derived independently using primarily remote sensing data. Important thermodynamic and physical information is revealed by the combination of remotely sensed normalized difference vegetation index (NDVI) and land surface temperature (To). The scatter plot of NDVI and To parameters form a characteristic triangular pattern whose boundaries are interpreted as limiting surface fluxes. We explore the robustness of Jiang-Islam methodology by doing an extensive inter-comparison of spatially distributed parameters derived from MODIS and AVHRR sensors onboard EOS Terra, NOAA14 and NOAA16 satellites respectively. Spatially distributed net radiation maps were retrieved using Bisht et al., (2005) methodology as an estimate of available energy to get both the spatial and temporal distribution maps of ET for clear sky days. We demonstrate the utility of newly launched sensors by validating the estimated ET results to ground flux stations of Southern Great Plains (SGP) with standard error of about 22% and correlation of 0.82.

Committee: Dr. Shafiqul Islam (Advisor) Subjects: Remote Sensing

MS, University of Cincinnati, 2004, Engineering : Environmental Engineering

Net radiation is a key quantity for the estimation of surface energy budget and is used for various applications including climate monitoring, weather prediction and agricultural meteorology. Remote sensing provides an unparallel spatial and temporal coverage of land, thus several studies have attempted to estimate net radiation (or its components) by combining remote sensing observations with ancillary surface and atmospheric data. A simple scheme is proposed to estimate instantaneous net radiation over large heterogeneous areas for clear sky days using only remote sensing observations. This is one of the first studies which abandons the need of ancillary ground information by using various land and atmospheric data products available from MODIS-Terra. It explicitly recognizes the need for spatially varied parameters and provides a distributed net radiation map over large heterogeneous domain with fine spatial resolution. Since instantaneous net radiation estimates have limited scope compared to daily average values or diurnal cycle, thus a sinusoidal model is proposed to estimate Diurnal cycle of Net Radiation.

Committee: Dr. Shafiqul Islam (Advisor) Subjects:

Doctor of Philosophy, The Ohio State University, 2012, Statistics

This dissertation is comprised of an introductory chapter and three stand-alone chapters. The three main chapters are tied together by a common theme: empirical hierarchical spatial-statistical modeling of non-Gaussian datasets. Such non-Gaussian datasets arise in a variety of disciplines, for example, in health studies, econometrics, ecological studies, and remote sensing of the Earth by satellites, and they are often very-large-to-massive. When analyzing ``big data,'' traditional spatial statistical methods are computationally intensive and sometimes not feasible, even in supercomputing environments. In addition, these datasets are often observed over extensive spatial domains, which make the assumption of spatial stationarity unrealistic. In this dissertation, we address these issues by using dimension-reduction techniques based on the Spatial Random Effects (SRE) model. We consider a hierarchical spatial statistical model consisting of a conditional exponential-family model for the observed data (which we call the data model), and an underlying (hidden) geostatistical process for some transformation of the (conditional) mean of the data model. Within the hierarchical model, dimension reduction is achieved by modeling the geostatistical process as a linear combination of a fixed number of basis functions, which results in substantial computational speed-ups. These models do not rely on specifying a spatial weights matrix, and no assumptions of homogeneity, stationarity, or isotropy are made. Another focus of the research presented in this dissertation is to properly account for spatial heterogeneity that often exists in these datasets. For example, with county-level health data, the population at risk is different for different counties and is typically a source of heterogeneity. This type of heterogeneity, whenever it exists, needs to be incorporated into the hierarchical model. We address this through the use of an offset term and by properly weighting the SRE (open full item for complete abstract)

Committee: Noel Cressie PhD (Advisor); Radu Herbei PhD (Committee Member); Desheng Liu PhD (Committee Member) Subjects: Statistics