Summer School 2024 Projects

Prediction of Pharmaceutical Removal from Wastewater with Machine Learning

Jude Okolie ·
Assistant Professor · University of Oklahoma · College of Engineering
CV · Personal Website · LinkedIn Profile

Water Quality · Wastewater Treatment · Sustainable Technology · Environmental Health · Pollution Mitigation

The proposed project aims to address concerns related to the removal of pharmaceuticals pollutants from wastewater, a critical environmental challenge due to the increasing prevalence of these pollutants. Pharmaceuticals as emerging pollutants pose significant risks to both human health and the aquatic ecosystem. Their microscopic nature allows them to often go undetected, thereby complicating efforts to mitigate their impact.

Biochar produced waste materials could be used as a promising adsorbent for the remediation of pharmaceutical pollutants. The specific questions to be answered include: How effective is biochar, compared to other adsorbents in removing various pharmaceutical micropollutants from wastewater? What are the optimal conditions under which biochar can maximize the adsorption of these pollutants? Additionally, the research aims to understand the relationship between the properties of biochar and its efficiency in removing different types of pharmaceuticals. Furthermore, publicly available software will be developed for predicting biochar effectiveness in removing pharmaceuticals from wastewater.

The proposed problem is intrinsically linked to the theme of water security, a central focus of the Summer School 2024. Water security involves ensuring sustainable access to safe and clean water, a necessity that is compromised by the presence of pharmaceutical pollutants. The research into biochar as a cost-effective and eco-friendly adsorbent aligns with the goals of water security by providing a viable solution for clean water and waste valorization. It addresses the urgent need for innovative and sustainable technologies to mitigate water pollution, thereby safeguarding both environmental and public health. The proposed project contributes to the broader goal of maintaining and improving water quality, a key aspect of water security.

Themes: Sustainable Development, Water Security
Sub-Themes: Geo-Enabling Reproducible and Open Science

Skills and Background: Basic or intermediate coding skills in Python, R, or MATLAB.  Participants with any academic background could work on the project as introductory materials on the nature of the problems will be provided along with the dataset.

AI-Driven Analysis for Enhancing Biodiversity through Urban Green Spaces

Elie Alhajjar ·
Senior Information Scientist · RAND Corporation
CV · Personal Website · LinkedIn Profile

Urbanization · Biodiversity Conservation · Urban Green Space · Ecological Impact · Urban Planning

Urbanization has profound effects on local ecosystems, often leading to reduced biodiversity and the degradation of natural habitats. However, urban green spaces (UGS) like parks, gardens, and green corridors can play a pivotal role in preserving biodiversity within city landscapes. These areas provide sanctuary for a variety of species while contributing to the health and well-being of urban populations. The challenge lies in optimizing the design and management of these spaces to maximize their ecological benefits. Traditional methods of monitoring and analyzing biodiversity and the effectiveness of UGS are labor-intensive and may not capture the full picture of their ecological impact.

This project seeks to leverage Artificial Intelligence (AI) to analyze various data types related to urban green spaces, biodiversity, and urban planning to enhance biodiversity conservation strategies in urban environments. The specific questions we aim to answer include:

  • How can AI and machine learning models predict the impact of existing urban green spaces on local biodiversity?
  • What AI-driven recommendations can be made for the design of future urban green spaces to maximize their biodiversity potential?

This project directly addresses the "Biodiversity" and "Sustainable Development" application areas of the Summer School 2024, with a thematic focus on "Leveraging AI for Environmental Sustainability." By applying AI techniques to urban ecological data, the project aims to develop actionable insights for the creation and management of urban green spaces that support biodiversity. This approach not only contributes to the preservation of species but also aligns with sustainable development goals by promoting green urban planning practices that benefit both the environment and human populations. By tackling a real-world problem with clear societal and environmental implications, this project offers participants the opportunity to contribute to meaningful sustainability outcomes. It also opens up avenues for interdisciplinary learning, as participants will need to consider ecological, urban planning, and social dimensions in their analyses.

Themes: Biodiversity, Sustainable Development
Sub-Themes: Geovisualization, GeoAI and Spatial Data Science via HPC

Data Sets:  The datasets for this project will enable a comprehensive analysis of the influence of urban green spaces on biodiversity, facilitating the development of AI-driven recommendations for urban planning and community engagement strategies.  Data will include

  • Satellite imagery and urban maps from NASA Earth Observing System Data and Information System (EOSDIS) and Google Earth Engine
  • Biodiversity records including species occurrence data within urban areas from the Global Biodiversity Information Facility (GBIF)

Skills: Python (Scikit-learn, TensorFlow or PyTorch, Pandas, and NumPy libraries)

Changing Water Footprints in Arid Environmental Settings of the Southwestern United States

Santosh Palmate ·
Assistant Professor · Texas A&M University · AgriLife Research Extension
CV · Research Profile · LinkedIn Profile

Arid Climates · Water Scarcity · Desert Ecosystems · Drought · International Water Conflicts

The arid climate region of the southwestern United States includes diverse landscapes characterized by low precipitation, high temperatures, and unique desert ecosystems. This US-Mexico border region is home to a population of over 2 million people and is also constantly threatened by chronic hazard drought impacting regional water resources availability. International water conflicts and disputes in the border region are caused by the high demand for limited water. Border community consists of a majority of the underrepresented Hispanic population. Agriculture in the border region is the source of income for many farmers and substantially contributes to the nation’s economy. Efforts have been made to balance water demand and supply, including water conservation measures and developing water infrastructures like reclamation plants and the world’s largest inland desalination plant to supplement the freshwater supply. Hence, changes in water footprints need to be investigated to sustainably manage water demands and supplies. Also, the border community faces climate risks and potential hazards, constraining community growth. Therefore, this project aims to improve water management in the arid regions of the US.

Themes: Effects of a Changing Climate, Sustainable Development, Water Security
Sub-Themes: Geo-Enabling Reproducible and Open Science, Geovisualization, GeoAI and Spatial Data Science via HPC

Data Sets: Elevation, Land Use / Land Cover, NDWI, NDVI, Population, Climate, Soil Textures

GeoAI Applications to Predict Field Scale Actual Evapotranspiration

Sushant Mehan ·
Assistant Professor · South Dakota State University · Department of Agricultural and Biosystems Engineering
CV · Laboratory Website · Faculty and Research Profile · LinkedIn Profile

Manoj Lamichhane ·
Graduate Student · South Dakota State University · Department of Agricultural and Biosystems Engineering
CV · LinkedIn Profile

Water Scarcity · Water Management · Agriculture · Crop Yield Optimization · Arid Climates

Evapotranspiration plays a crucial role in agricultural water management, as approximately 90% of water from crop fields is lost through evapotranspiration. The amount of water crops need, and the timing of their application are especially critical in arid and semi-arid regions, which consistently face water scarcity issues. Providing crops with the optimal quantity of water at the right time conserves water and enhances crop yield. Understanding the amount and timing of agricultural water usage is essential for estimating evapotranspiration from crop fields. Although numerous empirical equations exist for estimating reference evapotranspiration, they only represent specific points and fail to account for ground heterogeneity. Several satellite-based products offer reference evapotranspiration data, but these products often have biases when compared with ground observation, making them unsuitable for precise agricultural water management. Recently, the OpenETa platform became available, providing daily evapotranspiration data at a 30 m spatial resolution across the entire United States. However, our previous assessment indicates that OpenETa struggles to capture the daily dynamics of actual evapotranspiration (ETa) and exhibits significant biases compared to ETa measurements obtained from an eddy covariance tower in a crop. In this context, our goal is to develop a data-driven model capable of predicting daily reference evapotranspiration at a finer spatial resolution for agricultural fields in semi-arid regions. This model, once developed, could significantly enhance water resource management in these regions, ensuring the optimal use of water for crop production. To achieve this goal, we will use some classical machine learning models (Random Forest, Support Vector Machine, and Gradient Boosting) and Deep Neural Network (DNN).

Themes: Water Security
Sub-Themes: Geoethics, Geo-Enabling Reproducible and Open Science, Geovisualization, GeoAI and Spatial Data Science via HPC

Data Sets: We will employ innovative techniques, utilizing spectral indices derived from satellite imagery to develop machine-learning models. We will use the Moderate Resolution Imaging Spectroradiometer (MODIS) imagery, which has a spatial resolution ranging from 500 m to 1 km meters and a temporal resolution of 1 to 16 days. We will use for data retrieval. Spectral indices such as NDVI, EVI, LST day, and LST night will be extracted from the MODIS imagery. These spectral indices will serve as input features for our data-driven models. The target variable, reference evapotranspiration, will be calculated using the Penman-Montieth method. All these data span from 2000 to 2023.

Skills: Jupyter Notebook, Python (Sci-kit learn, Keras, Pandas, Numpy, and Matplotlib libraries), Javascript for data download, Google Earth Engine for data visualization.

Broadening Adoption of Cyberinfrastructure and Research Workforce Development for Disaster Management

Zhe Zhang ·
Assistant Professor · Texas A&M University · Department of Geography
CV · Faculty and Research Profile · LinkedIn Profile

Disaster Management · Curriculum Development · Continuing Education · Workforce Development · Cyberinfrastructure

This project aims to establish an International CyberTraining for Disaster Management (CTDM) network in which disaster management research communities can deepen their cyberinfrastructure (CI) and geospatial skills by participating in CTDM training activities. To achieve this goal, we have established a CI-enabled spatial disaster science network among institutions, federal agencies, hazards centers, industry, and educational organizations to leverage the expertise of our communities in developing the next-generation workforce training material across multiple scales. We plan to integrate the training material and modules into institutions’ graduate curricula and continuing education at Texas A&M, University of Illinois Urbana-Champaign, and Morgan State University. In this project, we will also introduce cutting-edge geovisualization tools (e.g., Digital Twin and Augmented Reality) to train participants to visualize disaster data and interpret spatial disaster patterns.

This project will build and expand upon the NSF-funded projects MIR-FASTER (OAC-2019129), ACES (OAC-2112356), DesignSafe (CMMI-2022469), CONVERGE (CMMI-1841338), and CyberGIS Software (OAC-1047916). We aim to directly train over 800 students and educators, potentially reaching 1200 or more through our education networks. In the 2024 summer school, the team plans to organize a CyberTraining session to tackle challenges related to sustainability and disaster management.

Themes:Effects of a Changing Climate, Sustainable Development
Sub-Themes: GeoAI and Spatial Data Science via HPC

Data Sets:

  • Social media data (downloaded using Twitter API)
  • Socioeconomic data (
  • Wal-Mart data (
  • NetCDF data (

Skills: Python, CyberGISX

Share this: