Weather Data Imputation

Team

• Faculty: Lav Khot (WSU), Prasad Tadepalli (OSU), Weng-Keen Wong (OSU), Alan Fern (OSU) 
• Senior Collaborators: Paola Pesantez-Cabrera (WSU, Data Scientist), Bryan Curtis (WSU AWN Data Scientist), Sean Hill (WSU AWN System Analyst) 
• Graduate students: Mohammad Rafid Ul Islam (OSU PhD in AI)
• Undergraduate students: Robert Barragan (HU, CS)

Objectives

• Develop and implement weather station correlation methodology for site-specific usage in case of missing data from a preferred station.
• Develop robust weather data imputation methods and evaluate them on AgWeatherNet. 
• Develop and implement site-specific weather nowcasting/forecasting for 
• improved decision support through ported models.

Hypotheses

• Weather station correlation methodology can help in running AI-based agricultural decision support models on alternate station data in case of missing data
• The neural-network-based imputation methods can outperform classical statistical  approaches
• The higher accuracy, station-specific forecasts will allow growers to improve anticipation of weather events and AI model predicted crop loss scenarios at their specific location and better plan on appropriate mitigation strategies. 

Related Publications