The North Carolina Institute for Climate Studies (NCICS) is hosting more than a dozen undergraduate interns this summer. These internships offer students a chance to do hands-on work across a wide range of topics, including teleconnections in the climate system, sea surface temperature data, an economic cost–benefit analysis of NOAA satellite observing platforms, and a number of projects exploring cloud-computing and machine-learning techniques to improve access to critical environmental data. Jonathan Brannock, Jenny Dissen, Garrett Graham, Douglas Rao, Carl Schreck, and Denis Willett are among the NCICS staff serving as mentors for this group.
NCICS staff members also frequently serve as mentors for other interns hosted by NOAA’s National Centers for Environmental Information (NCEI) and the NASA DEVELOP program. This summer, Douglas Rao was one of the mentors for a DEVELOP team working in Asheville, North Carolina, on a project titled “Asheville Urban Development II: Mapping Urban Heat to Support Cooling Initiatives and Climate Resilience Planning in the Greater Asheville Area.” The team partnered with the City of Asheville Sustainability Department and and the nonprofit Asheville GreenWorks to identify urban heat islands using data on land surface temperatures, albedo, evapotranspiration, and social vulnerability.
Below we highlight the projects of some of our NCICS interns.
See NCEI’s web story on the Summer 2024 NASA DEVELOP program, and learn more about other NCEI summer interns here.
NCICS 2024 Summer Interns
Soubhagya Akkena
Degree program: Master’s in Computer Science at NC State Mentor: Carl Schreck
Project: My internship project delves into the technical aspects of climatology and teleconnections to understand the impacts of climate change on streamflow patterns and explore the interactions among various climate teleconnections, such as ENSO, MJO, and NAO, using extensive time-series data.
The project involves identifying teleconnection events, adjusting climatological data, and applying bootstrapping techniques to determine statistical significance. This scalable approach, facilitated by cloud technologies, enhances the modeling and prediction of streamflow, contributing valuable insights to the atmospheric drivers section of NCEI’s monthly report on the state of the climate.
My next project employs advanced data mining techniques and statistical analysis to investigate how increasing precipitation trends affect streamflow.
Apurv Choudhari
Degree program: Master’s in Computer Science at NC State Mentor: Denis Willett
Project: In my internship project, I am setting up a hybrid cloud environment for climate research. Traditional on-premises HPC (high-performance computing) clusters come with significant challenges, including scalability, maintenance, and flexibility in setting up new environments. The new hybrid cloud environment addresses these issues by enabling rapid provisioning and deprovisioning of HPC clusters with necessary libraries within minutes, reducing setup and maintenance time and costs.
I am also integrating Open OnDemand software, an open-source solution that provides researchers with a web-based portal to access both on-premises and cloud HPC clusters. This user-friendly interface simplifies the process of leveraging HPC resources, allowing researchers to focus more on their work and less on the underlying infrastructure.
Megh Dedhia
Degree program: Master’s in Computer Science at NC State Mentors: Denis Willett and Jonathan Brannock
Project: I am working on the potential use of Elasticsearch for data discovery on open and public datasets managed by NOAA. While ingesting data from multiple satellites has been convenient, searching through it to find any piece of data has been tedious.
To allow searching through the data more quickly, I am working on a solution that involves the use of Elasticsearch to cache recently ingested files with their names to allow searching them and pinpointing their exact locations across multiple datasets.
Akshada Malpure
Degree program: Master’s in Computer Science at NC State Mentor: Denis Willett
Project: During my internship at NCICS, I have been involved in a multifaceted project focusing on Amazon Web Services (AWS) metrics monitoring and analysis. Key tasks included exporting AWS Steampipe tables to CSV, utilizing Python subprocesses to call export CLI, and deploying a Lambda function using Docker.
This Lambda function orchestrates AWS and Python scripts to process and analyze AWS metrics data stored in S3, enhancing the efficiency and accuracy of our monitoring system.
Antoni Marvin
Degree program: Master’s in Data Analytics at NC State Mentor: Jenny Dissen
Project: I worked on providing a cost–benefit analysis to support the continuation of NOAA’s microwave and infrared sounders. Both instruments provide data and information regarding certain aspects of the atmosphere (temperature, hurricane forecast accuracy, etc.). This data helps users make informed decisions, resulting in both indirect and direct benefits that can be interpreted in terms of dollar amounts.
We worked to determine whether the benefits of these observing systems are greater than the costs associated with the maintenance of these technologies.
Zachary Moss
Degree program: Bachelor’s in Meteorology at NC State Mentor: Douglas Rao
Project: For my internship project, I am developing a computational notebook that will allow users to access, visualize, and analyze NCEI’s data. So far, I have written programming snippets that will give users access to NOAA Climate Data Records (CDRs) through an open-source registry.
I have also written a coding example that retrieves sea surface temperature data from the Woods Hole Oceanographic Institute CDR and generates static and interactive mapping plots.
Shubh Nisar
Degree program: Master’s in Computer Science at NC State Mentors: Garrett Graham and Denis Willett
Project: I am working on developing a dashboard platform to study sea surface temperatures. I am leveraging Next.js and Tailwind CSS to build a modern web platform. GitLab is used to handle version control and code management. Several Amazon Web Services (AWS) are used to perform certain functions. Continuous integration and deployment is performed using AWS Amplify to deploy the production environment.
AWS Cloud9 is used to create a sample Docker container to parse AWS S3 buckets comprising climate parquet files. The Docker container is then published to AWS ECR to deploy the AWS Lambda to provide a serverless solution leading to a microservice.
Homak Patel
Degree program: Master’s in Computer Science at NC State Mentor: Denis Willett
Project: I am working on anomaly detection in NOAA satellite streaming feeds. My responsibilities include enhancing and developing a machine-learning pipeline with real-time NOAA data, preprocessing and extracting data through engineered ETL pipelines, and leading MLOps initiatives to deploy models with robust continuous integration and deployment practices.
Additionally, I develop advanced deep-learning models for multi-time-series forecasting, utilizing open-source frameworks and pretrained language model libraries to improve accuracy and efficiency.
Priya Shah
Degree program: Master’s in Computer Science at NC State Mentor: Denis Willett
Project: I am currently working on integrating large language models such as Mistral AI and LLaMA with a retrieval-augmented generation (RAG) pipeline using AWS Bedrock, Kendra, and DSPy.
This project aims to deploy advanced chatbots tailored for diverse research groups and organizations catering to federal initiatives. The integration focuses on enhancing the efficiency and effectiveness of research communication and information retrieval.