Topics for Thesis Projects

Below find a list of potential thesis projects. Contact.

(1) Multi-Scale Predictive Modeling of Environmental Properties

Ecosystem theory rooted in resilience, ecological indicators and thresholds has influenced restoration efforts in aquatic and terrestrial ecosystem. Transforming these conceptual ideas into spatially- and temporally-explicit models has been hampered by ecosystem complexity, multiple nested levels of interrelated physical, biological and chemical processes, and the lack of sufficient quantitative data. Unresolved questions entail: (i) Which key ecological variables show self-similar spatio-temporal behavior? (ii) Do generalized algorithms exist to describe functional ecoprocess-relationships? (iii) How do these functional relationships behave in space and through time? and (iv) How much complexity must be modeled to effectively describe ecosystem behavior and associated scale dependent behavior?

The specific objectives of this project are to investigate the effect of different grain sizes (1 m to 1000 m) of remote sensing-derived landscape properties (e.g. Normalized Difference Vegetation Index, Leaf Area Index, land cover, net primary vegetation production, biomass, fraction of absorbed photosynthetically active radiation (fAPAR), land surface temperature, etc.) on the prediction of biogeochemical soil properties (e.g. soil phosphorus, nitrogen, carbon, calcium, etc.). Remote sensing images with high spatial resolution (e.g. IKONOS - 1 to 4m), medium resolution (e.g. ASTER - 15 to 30 m) and coarse resolution (e.g. MODIS - 250 - 1000 m) will be used to characterize landscape features. The accuracy of spatially-explicit predictive soil-landscape models will be assessed. Multi-scale behavior of landscape properties will be investigated using landscape indices. Potential study area: Greater Everglades ecosystem or Santa Fe River Watershed.

Skills and research interests: Remote sensing, GIS, geospatial modeling and ecology.

(2) Digital Predictive Soil Mapping - Pedometrics

Digital soil mapping techniques have shown much promise to reduce soil mapping costs and produce high-resolution soil maps covering large areas. Numerous pedometric methods have been developed that use Geographic Information Systems (GIS), global positioning systems (GPS), advanced statistical and geostatistical methods, and remote sensing. These digital soil mapping technologies have greatly improved our capabilities for (semi)automated soil mapping. Despite these advancements knowledge gaps exist that translate into the following questions:

(i) Which data inputs (factors) produce the best predictive soil model within a given region?
(ii) What is the optimal spatial resolution to produce accurate soil attribute maps?
(ii) Which method performs best to predict soil properties and classes within a given region?

The goal is to compare and evaluate different semi-automated digital soil mapping techniques rooted in the SCORPAN modeling concept in a landscape unit in northeastern Florida. The quantitative SCORPAN concept relates landscape factors (S: soils, C: climate, O: organisms/vegetation, R: relief, P: parent material/geology, A: age, and N: geographic space) to soil properties.

The specific objectives of this project are to: (i) Compare statistical and geostatistical/hybrid methods to predict selected soil properties (pH, total carbon, total nitrogen, total phosphorus, and diagnostic horizons) and classes (drainage, Soil Series, and Soil Orders) using a sequential soil-landscape modeling approach; (ii) Assess the usefulness to incorporate remote sensing imagery at different grain sizes (spatial resolutions) into soil prediction models; (iii) Identify the soil prediction model that performs best in terms of prediction quality, costs and labor; and (iv) Develop a standardized protocol to enable transfer of the proposed methodology to other soil survey regions.

Skills and research interests: GIS, geostatistics, and soil/environmental science. 

(3) Visible/Near-Infrared Diffuse Reflectance Spectroscopy (VNIR-DRS)

VNIR-DRS has the potential for cost-effective, rapid and accurate mapping of soil properties across larger landscapes. In this project soil samples will be scanned with a spectroradiometer and spectral scans related to lab soil data. Multiple chemometric modeling techniques will be compared to identify the method that performs best to predict select soil properties (e.g. soil organic carbon, texture, calcium carbonate content, etc.). Interest in quantitative modeling is a prerequisite for this project.

Skills and research interests: Quantitative methods and soil science.

(4) Ecological Indicators

Different types of ecosystems which can be characterized by different environmental factor combinations perform different or similar environmental functions to differing degrees. These functionally diverse ecosystems are linked into an intricate network of material and energy flux. Ecological indicators describe a characteristic of an ecosystem that is related to, or derived from, a measure of biotic or abiotic variable, that can provide quantitative information on ecological structure and function. Indicators can be used to assess the ecological integrity, sustainability and the current condition of an ecosystem. They diagnose the cause of an environmental problem and provide an early warning signal of changes in the environment.

The specific objectives of this project are to identify ecological indicators in the Greater Everglades ecosystem. A comprehensive available database of biogeochemical properties will form the basis of the project. A suite of quantitative statistical and geostatistical methods (e.g. Canonical Correlation Analyses, Generalized Linear Models, Classification and Regression Trees, multi-variate kriging) will be used to identify indicator variables.

Skills and research interests: GIS, statistics, geostatistics, and environmental science. 

(5) Development of Reusable Learning Objects (RLOs) [to earn teaching/service credits]

Digital libraries of Reusable Learning Objects (RLO) offer new avenues to support education efforts that benefit multiple instructors, courses and graduate programs. Because RLOs are shared educational resources they are cost-effective and can be used to support distance education and on-campus instruction. Reusable Learning Objects are a new type of online instruction that provide a digital educational resource that can be reused, scaled and shared from a central online repository in the support of instruction and learning. Each RLO supports a single learning objective, which are streamlined into a Digital Library of RLOs. These include, but are not limited to text entries, web sites, bibliographies, charts, figures, maps, models, photographs, illustrations, mini-case studies, assessments, tutorials, simulations, animations, audio and video clips, movies, and interactive tools. They vary in size, scope and level of granularity ranging from small chunks of instruction to a series of combined resources to provide a more complex learning experience.

Specific objectives are to: (i) Develop one or more standardized RLOs on one of the following topics:

• Biogeochemistry
• Ecohydrology
• Ecosystem services
• Microbial ecology
• Nutrient management
• Soil and water chemistry
• Soil-landscape analyses
• Environmental pedology
• Soil mapping and surveys
• Soil physics
• Soil chemistry
• Sustainability of water resources
• Water-borne pathogens
• Wetlands and water quality

(ii) and submit the RLO(s) into the environmental RLO digital repository hosted in the Soil and Water Science Department, UF. The RLOs will be peer-reviewed and after approval included in the digital library that is used by instructors and students in support of teaching.

(6) Ecosystem Services

Ecosystem services refer to the benefits that ecosystems provide to the globe and humanity and cover fine (microscopic) to coarse (global) scales. Ecosystem services characterize the functions that are useful to humans and contribute to ecosystem stability, resilience, sustainability and integrity, such as nutrient cycling and productivity. These services are diverse ranging from physical (e.g. carbon sequestration, green corridors that act as buffers, best management practices that reduce nutrient leaching, etc.), cultural, social-economic and aesthetic. The assessment and quantification of ecosystem services are important because of the multi-functional land uses that compete for the same land and water resources and to minimizes adverse effects on environmental systems, preserve social equity and maximize economic output.  

Ecosystem services have been categorized into the following functional groups: provisioning services, regulating services, cultural services and supporting services. The process of placing economic value on non-market goods remains problematic and is still evolving. A crucial component to the study of ecosystem services is the valuation of those services. The value of many ecosystem services, particularly provisioning services such as agricultural and forest products, is market-oriented. However, a market value is not as easily placed on services that can be considered 'outside the market'.  

Common valuation methods of ecosystem services include: travel cost method, hedonic pricing method, and contingent valuation method, each differing in underlying assumptions and possessing unique limitations and uncertainties. The travel cost method places value on ecosystem services via observed consumption in related markets, and proxy consumption costs are substituted for the market price of the ecosystem goods or services in question. The hedonic pricing method assigns value by estimating a statistical relationship between the attributes of the system and the price of goods for which a market actually exists. Lastly, the contingent valuation method is based on social scientific survey techniques of sample populations and asking people for their "willingness to pay" estimates.

The objectives of this project are to:

(i) Assemble a spatially-explicit dataset to characterize landscape properties across the study area (including: climate, soils, land use, land cover, land use management, topography, hydrology, geology, demography, infrastructure, recreational and cultural points of interest, etc.). GIS will be used to manage the spatially-explicit datasets.

(ii) Compare two different valuation methods to assess ecosystem services within the study area.

(iii) Develop a spatially-explicit model that describes the relationships between landscape properties and environmental response variables ("cause-effect relationships").

Potential study areas:

(i) Hydrologic unit within the Lake Okeechobee Basin or the Suwannee Basin; (ii) Hydrologic unit undergoing rapid urban growth.

Skills and research interests: GIS, geospatial modeling and environmental science.