Disaster Science Operation Center (DSOC) · Dept. of Earth, Environmental & Atmospheric Sciences · Western Kentucky University
WKU EEAS Dr. Singh's Page Google Scholar
WKU · DSOC · Earth, Environmental & Atmospheric Sciences

AI Research
Laboratory

AIR Lab · Western Kentucky University

Building next-generation AI systems for weather, climate, and disaster science. From street-level forecasting to flood digital twins — AIR Lab pushes the frontier of machine learning in Earth science to protect lives and communities.

Explore Research View Publications
4,300+
Research Citations
50+
Publications
11+
Years of Research
4
Research Domains
WKU
Disaster Science Operation Center (DSOC) Dept. of Earth, Environmental & Atmospheric Sciences · Bowling Green, KY
EM1
KEM
Industry & Government Partnerships AI company EM1 · Kentucky Emergency Management · NSF EPSCoR CLIMBS
About AIR Lab

Pioneering AI for
Earth & Disaster Science

The AI Research Laboratory (AIR Lab) is a specialized unit within the Disaster Science Operation Center (DSOC) at Western Kentucky University, embedded in the Department of Earth, Environmental, and Atmospheric Sciences.

We build cutting-edge machine learning systems to improve forecasting of extreme weather events, floods, wildfires, and other natural hazards — helping communities and emergency managers make faster, better-informed decisions. Our AI systems demonstrated this impact most visibly when they outperformed traditional physics-based models during Hurricane Beryl.

We collaborate actively with Kentucky Emergency Management and AI company EM1, and contribute to the NSF EPSCoR CLIMBS storm resilience project and the World Meteorological Organization's research programs.

🌪️

AI-Augmented Forecasting

Deep learning for high-resolution weather and climate prediction at street level

🌊

Digital Twins

Flood and wildfire simulation systems for real-time emergency management

🛰️

Earth Observation

Satellite remote sensing and gap-filling for global environmental monitoring

🧬

Climate-Health Nexus

Modeling links between climate variability and public health outcomes

Part of DSOC at WKU

AIR Lab operates within the Disaster Science Operation Center — a hub connecting academia, government, and industry in the fight against natural disasters across Kentucky and beyond.

  • 🏛️ Western Kentucky University · EEAS Department
  • 🚨 Kentucky Emergency Management (KEM)
  • 🤖 EM1 — AI for Emergency Management
  • 🔬 NSF EPSCoR CLIMBS — Storm Resilience
  • 🌍 World Meteorological Organization (WMO)
  • 📡 IPCC AR6 Earth System Model Contribution
  • 🎓 Fulbright-Kalam Fellow · UT Austin
Research Projects

What We're Building

Our four core research domains tackle grand challenges at the frontier of AI and Earth science, with direct impact on disaster preparedness and climate resilience.

🌪️
01

AI-Augmented Weather Prediction

Developing deep learning systems that push forecast resolution from 25 km grids down to street-by-street scale. During Hurricane Beryl, our AI models outperformed traditional physics-based systems. Our WMO Research Demonstration Project delivered ultra-high-resolution AI forecasting for the 2024 Paris Olympics.

Active Research
🌊
02

Disaster Forecasting & Flood Digital Twins

Building real-time AI-powered digital twins of flood and wildfire events in partnership with EM1 and Kentucky Emergency Management. These platforms allow emergency managers to model disaster scenarios and optimize community response before, during, and after high-impact events.

Active Research
🌡️
03

Climate Modeling & Earth System Science

Advancing large-scale climate science through AI — from downscaling projections with physics-inspired neural networks, to aerosol-cloud interaction modeling that contributed to the IPCC Sixth Assessment Report (AR6). We apply causal discovery, complex networks, and non-linear time series analysis to Earth system challenges.

Active Research
🛰️
04

Remote Sensing & Satellite Data

Applying AI to fill critical gaps in satellite remote sensing datasets, enhancing Earth observation records for climate monitoring. We also develop high-resolution urban gridded datasets and global building height products using Google Earth Engine, Microsoft Planetary Computer, and GPU-accelerated HPC pipelines.

Active Research
🧬
05

Climate-Health Nexus

Investigating links between climate variability and public health. Our recent work includes the EpiClim dataset — a weekly, district-wise all-India climate-health dataset for accelerated GeoHealth research covering multiple epidemics and their environmental drivers across diverse geographic regions.

New · 2025
🏙️
06

Urban Climate Downscaling

Developing novel physics-inspired deep learning for high-resolution downscaling of climate data over urban areas, including global building height datasets for city-scale climate impact assessment, urban heat island analysis, and the design of climate-resilient communities.

Ongoing
Where We're Headed

Our Current Focus

AIR Lab is actively pushing forward on two major fronts that will define the next generation of AI-powered Earth science — combining the depth of climate physics with the speed and scalability of modern machine learning.

🌐
Focus Area 01

High-Resolution Climate & Weather Datasets

Developing state-of-the-art high-resolution climate and weather datasets for the United States and globally. These datasets power downstream AI models, enable climate impact assessment at the community level, and fill critical observational gaps in Earth science research.

📅
Focus Area 02

AI SubSeasonal-to-Seasonal (S2S) Models

Building the next generation of AI-driven SubSeasonal-to-Seasonal (S2S) prediction models — forecasting weather and climate patterns weeks to months ahead. S2S prediction sits in the "predictability gap" between weather forecasts and climate projections, and AI offers transformative potential to unlock this challenging timescale for disaster planning and agricultural decision-making.

Publications

Recent Research Output

Selected from 50+ publications with 4,300+ citations. View the complete list on Google Scholar.

View All on Google Scholar →
2025
EpiClim: Weekly District-Wise All-India Multi-Epidemics Climate-Health Dataset for Accelerated GeoHealth Research
Kaur, G., Ghoshal, S., Marbate, R., Malviya, N., Kaur, A., SB, V., Srivastava, A.K. and Singh, M.
arXiv preprint · arXiv:2501.18602
New · 2025
2024
High-Resolution AI Forecasting for the 2024 Paris Olympics — WMO Research Demonstration Project
Singh, M. et al.
World Meteorological Organization Research Demonstration Project
Featured
2024
Deep Learning Augmented Numerical Weather Prediction: Improvements in High-Impact Short-Range Forecasts
Singh, M. et al.
Geophysical Research Letters
Journal
2023
Physics-Inspired Deep Learning for High-Resolution Downscaling of Climate Data
Singh, M. et al.
Geophysical Research Letters
Journal
2022
Role of Volcanic and Anthropogenic Aerosols on the Tropical Ocean-Atmosphere-Land Coupled System and the South Asian Monsoon
Singh, M. — PhD Dissertation
Indian Institute of Technology Bombay · CGPA 9.0/10.0
Thesis
2021+
IITM Earth System Model Contributions to IPCC Sixth Assessment Report (AR6)
Singh, M. and collaborators · Indian Institute of Tropical Meteorology, Ministry of Earth Sciences, India
IPCC AR6 Contributing Research · CMIP6 Simulations
IPCC · AR6
Capabilities & Services

What AIR Lab Offers

AIR Lab offers research partnerships, consulting, and training services to government agencies, emergency management offices, and academic institutions.

🤖

AI Model Development

Custom deep learning and ML models for weather, climate, and environmental applications. We build AI systems that outperform traditional physics-based approaches.

  • CNN & Transformer models for geospatial data
  • Physics-informed deep learning architectures
  • Real-time inference pipelines for operations
  • Uncertainty quantification & interpretability
🚨

Disaster Intelligence & Emergency AI

Embedding AI decision-support tools into emergency management workflows, helping communities prepare and respond to disasters faster.

  • AI-powered early warning systems
  • Real-time flood and wildfire risk mapping
  • Digital twin scenario simulations
  • Integration with KY Emergency Management
📡

Climate Data Science

Advanced data science for large-scale climate and satellite datasets using Google Earth Engine, Microsoft Planetary Computer, and HPC workflows.

  • Satellite remote sensing & gap-filling
  • High-resolution urban gridded datasets
  • Climate model bias correction & downscaling
  • CMIP6 Earth system model analysis
🎓

Education & Training

Preparing the next generation of AI and climate scientists through rigorous graduate mentorship, workshops, and open-science training programs.

  • Graduate research mentorship (MS/PhD)
  • Workshops on deep learning for geoscience
  • Open-science & reproducible research
  • Guest lectures & invited seminars
Faculty & Students

Meet the Team

A dedicated team of researchers at the frontier of AI and Earth science at Western Kentucky University.

Faculty · Principal Investigator
Dr. Manmeet Singh
Principal Investigator
Lab Director · Assistant Professor
Dr. Manmeet Singh
Assistant Professor, Dept. of Earth, Environmental & Atmospheric Sciences
Western Kentucky University · Bowling Green, KY
Dr. Singh is a computational climate scientist with over 11 years of research experience, bringing expertise in AI/ML for Earth and atmospheric sciences. He previously served as a Distinguished Postdoctoral Fellow at the Jackson School of Geosciences, UT Austin, and as a Staff Scientist at the Indian Institute of Tropical Meteorology for 11 years. His research spans AI-augmented weather prediction, climate model development, causal discovery in climate systems, and remote sensing — with contributions to the IPCC AR6 report and the WMO's 2024 Paris Olympics Research Demonstration Project.
AI / Deep Learning Numerical Weather Prediction Climate Modeling Disaster Science Computer Vision Causal AI Remote Sensing IPCC AR6
Personal Website Google Scholar Email WKU Profile
Graduate Researchers
Somnath Luitel
Somnath Luitel
Graduate Researcher
somnath.luitel416@topper.wku.edu
Somnath Luitel is a graduate researcher at AIR Lab, Western Kentucky University, conducting AI and climate research under the supervision of Dr. Manmeet Singh. Originally from Nepal, he earned his Bachelor’s degree in Civil Engineering from Pulchowk Campus, Tribhuvan University. His research focuses on AI applications in weather and climate science.
AI / ML Climate Science WKU EEAS
+

Open Position — Graduate Student

We are seeking motivated MS students passionate about AI, weather, climate, and disaster science to join our growing team at WKU.

Apply Now →
+

Open Position — Postdoc / Visitor

Postdoctoral researchers and visiting scholars with backgrounds in AI, climate modeling, or remote sensing are also welcome to inquire about opportunities.

Get in Touch →
Get in Touch

Connect with
AIR Lab

Whether you're a prospective student, collaborator, government partner, or curious about our research — we'd love to hear from you.

📍
Location
Disaster Science Operation Center (DSOC)
Dept. of Earth, Environmental & Atmospheric Sciences
Western Kentucky University
1906 College Heights Blvd., Bowling Green, KY 42101
✉️
🌐
🎓
Prospective Students
We welcome applications from students with strong backgrounds in mathematics, computer science, atmospheric sciences, or related fields. Email your CV and a brief statement of research interests.
Send a Message
✓ Message received! We'll be in touch soon.