Georeferenced Data for Agricultural Management
- Thursday, November 29, 2018 from 3:10pm to 4:00pm
- Wilson Hall, 1-144 - view map
Applied Math Seminar
Dr. Bruce Maxwell (Montana Institute on Ecosystems):
Using Georeferenced Data to Make Better Agricultural Management Decisions
Farms are increasingly at risk due to management practices that fail to account for uncertainties in crop price, rising input costs and climatic variability. We are developing an automated on-farm precision experimentation (OFPE) framework by combining precision agriculture technologies with remote sensing and other GIS based information to deliver profit maximizing, pollution minimizing and resilience maximizing prescriptions for crop production inputs (e.g. fertilizer application strategies as a proof of concept). Our research over the past 20 years has consistently indicated that crop response to inputs like fertilizers are site specific in nature. Every field is different and most years are different. Therefore, in order to make accurate recommendations, the OFPE takes advantage of modern technologies to conduct experiments on the fields where recommendations will be made. Recommendations require optimizing the use of inputs based on maximizing farmer profits and minimizing pollution. We are producing the tools that allow farmers to determine on their own the best management practices on each field given uncertainty in crop response to climate, prices received and costs of production. We are capturing the unprecedented modern data stream available to agricultural producers by creating software that manages the data and automates the process of conducting experiments and analyzing the data field by field. A farmer that has yield monitors, GPS, and variable rate application (VRA) capability can have experimental fertilizer treatments placed in their fields to create crop yield and quality response functions that can be used to assess the economic response to inputs (e.g. fertilizer). The data is noisy and we are challenged with how to appropriately characterize the uncertainty in recommendations that emerge from our analysis of the data.