Jumi Gogoi
Jumi Gogoi, PhD
About Me
I am a Postdoctoral Fellow at the University of British Columbia, working in partnership with the BC Wine Grape Council (BCWGC) through a Mitacs Accelerate Fellowship. My research focuses on BC wine grape producers, developing predictive models for grapevine cold hardiness and disease risk tailored to BC's growing conditions and integrating these into BCWGC's Decision Aid System (DAS), a grower-facing decision support platform being expanded to wine grape production across the province.
My work bridges agronomic modeling, machine learning, and applied decision support to develop tools that are both scientifically sound and practically useful for agricultural producers — from wine grape growers navigating climate variability and crop health risks to broader crop production systems requiring data-driven yield forecasting.
I bring a cross-disciplinary background spanning economics (MSc, University of Bath), business analytics (MSc, University of Dallas), and a PhD from UBC's Institute for Resources, Environment and Sustainability, where I applied satellite remote sensing and machine and deep learning to crop yield prediction across the Canadian Prairies from county to subfield scale. I have also contributed as a researcher to Agriculture and Agri-Food Canada's Big Data and Predictive Analytics branch.
I am based in BC and committed to building applied research and decision-support capacity within the provincial wine grape sector.
News
May 2026
Started a Mitacs Accelerate Postdoctoral Fellowship at UBC in partnership with the BC Wine Grape Council, focusing on grapevine cold hardiness modeling and disease risk forecasting for BC wine grape producers.
November 2024
Presented at the Global Land Programme's 5th Open Science Conference in Oaxaca, Mexico. Research demonstrated the utility of high-resolution satellite data and advanced machine learning methods for improving crop yield estimation and forecasting.
July 2023
Presented at The International Environmetrics Society (TIES) Regional Meeting — including a paper on multi-source spatio-temporal datasets for crop yield prediction and led a plenary workshop on analyzing big geospatial datasets using Google Earth Engine.
Research & Tools
My current work is focused on building practical decision support tools for BC wine grape producers. Prior research developed the machine learning and remote sensing foundations that underpin this work.
Current Project: Grapevine Cold Hardiness and Disease Risk Modeling for BC Wine Grape Production
(UBC × BC Wine Grape Council, Mitacs Accelerate, 2026–)
Developing and validating predictive models for grapevine cold hardiness and disease risk under BC-specific growing conditions, with outputs to be integrated into BCWGC's Decision Aid System (DAS). The goal: timely, actionable forecasts that support grower decisions on frost protection, canopy management, and fungicide timing across BC's wine growing regions.
Open-source tools: making viticultural and agricultural data more accessible and actionable for growers, industry, and researchers. Tools will be linked here as they become available.
- Canadian Crop Atlas — A free, open-source tool that brings together satellite data, weather records, and crop health indicators for exploring cropping conditions across Canada. Track biomass, moisture stress, and growing conditions across multiple years, and download data as CSV files — all through your browser, no specialized software needed. Designed for growers, researchers, and industry.
- Vineyard Map (in development) — An open-source interactive map for exploring geospatial and climate datasets across BC wine grape growing regions.
- Crop and Vineyard Dashboards (in development) — Interactive dashboards for exploring datasets and climate indicators relevant to BC wine grape production.
Prior Projects
Using municipality-scale datasets across the Canadian Prairies and Ontario, my prior research investigated how combining multi-source satellite imagery — Landsat, Sentinel-2, and MODIS — with weather and biophysical data improves crop yield estimation accuracy for key crops including barley, canola, wheat, corn, and soybean. At sub-field scale, machine and deep learning approaches were evaluated using precision agriculture yield monitor data to generate crop yield forecasts up to three months before harvest. The methods and modeling frameworks developed through this work now inform the data-driven approach applied to BC viticultural modeling (With Navin Ramankutty, Nathaniel K. Newlands, Nicholas C. Coops).
Publications
Jumi Gogoi, Nathaniel K. Newlands, Zia Mehrabi, Nicholas C. Coops, Navin Ramankutty
Jumi Gogoi