Recipient Organization
MICHIGAN STATE UNIV
(N/A)
EAST LANSING,MI 48824
Performing Department
Kellogg Biological Station
Non Technical Summary
The balance between producing more food for a growing population and protecting the environment on which our lives depend is becoming increasingly precarious. As a result, sustainable practices that use soil, water, and other resources more efficiently, and are able to adapt to increase climate variability, are critical.Excessive application of Nitrogen (N) fertilizer over a large area releases significant amounts of nitrous oxide, a potent greenhouse gas, into the atmosphere and can cause harmful algal blooms and hypoxia Farmers already recognize that any fertilizer that doesn't get taken up by crops can be lost from their fields, thereby lowering profits. Optimal application rates should vary across space and time to match the variability of crop growth conditions and soil properties. Applying fertilizer at a variable rate across a field is challenging because it requires understanding detailed variability across each field and the relationship of that variability to weather. The aim of this research is to advance the science needed to achieve that goal.The overarching goal of this projectis to develop and enable solutions to improve the economic and environmental efficiency of US agricultural systems given the knowledge of spatial and temporal variability of corn and soybean yields related to climate variability, landscape characteristics and management.This research will combine new data on spatial and temporal yield variability with advanced analytical tools to investigate the hypothesis that site-specific management strategies based on newly developed yield stability maps and resultsfrom integrated geospatial system approachcan reduce the cost of crop production, reduce off-site environmental impacts, and maintain or increase yields under increased climate variability.
Animal Health Component
40%
Research Effort Categories
Basic
0%
Applied
40%
Developmental
60%
Goals / Objectives
The overarching goal of this projectis to develop and enable solutions to improve the economic and environmental efficiency of US agricultural systems given the knowledge of spatial and temporal variability of corn and soybean yields related to climate variability, landscape characteristics and management.This research will combine new data on spatial and temporal yield variability with advanced analytical tools to investigate the hypothesis that site-specific management strategies based on these new data can reduce the cost of crop production, reduce off-site environmental impacts, and maintain or increase yields under increased climate variability.Specific objectives are to:Analyze stability maps, spatial and temporal variability of N uptake and nitrate leaching in corn and soybean crops across US corn belt.Use long-term crop models' ensemble analyses to identify strategies able to better adapt to climate variability and to lead to more robust and efficient nitrogen management practices.Convert yield maps into profitability maps to evaluate alternative management approaches to removeAnalyze the productivity and environmental benefits of spatially variable N management or allocation of within field low-productivity zones to alternative crops such as periennial bioenergy crops, or native vegetation.To fulfill these objectives, I propose an integrated geospatial approach that links crop modeling with remote sensing and AI big-data analytics to improve our understanding of the US corn belt and its resilience to increased climate variability and change.
Project Methods
Building on new evidence showing the additional predictive power of crop model ensembles [Martre et al, 2016], our ensemble modeling approach will utilize widely-used crop system models (DSSAT, EPIC, SALUS, APSIM, N-Wheat, DAYCENT, CropSyst) to provide accurate evaluations of a range of sustainable management practices. These models will evaluate spatial and temporal crop responses of yield, water, and N loss to crop N fertilizer management and alternative crop arrangements and to quantify output uncertainty.Here is a brief description of the some of the models used in this research:SALUS: The Systems Approach for Land Use Sustainability (SALUS) model was designed with the goal of quantifying the impact of management strategies and their interactions with the soil-plant-atmosphere system on yield and C, N, and P dynamics (Basso and Ritchie, 2015).DSSAT: The DSSAT Cropping System Model (CSM) is a widely-used set of crop models that includes the CERES models for corn, wheat, and other cereals, and the CROPGRO models for soybean and other legumes (Jones et al. 2003).EPIC: EPIC can simulate the growth, development, and productivity of over 100 plant species including all major crops, grasses, legumes, and some trees. These crops can be grown alone or as intercrops, in complex rotations, and under a wide range of management operations including tillage, irrigation, fertilization, and liming (Williams 1995).ApproachThe first step in the proposed research is to enhance the understanding of the biophysical aspects of yield stability maps and to evaluate if more years of simulated yield available from a well-tested crop models would change their proportion between HS, LS and U. This analysis will be performed using long-term remote sensing data coupled with spatial application of crop models at field scale. Changes in stability maps will also be quantified in response to the removal of extreme years (wet or dry) to better identify factors responsible for unstable zones.In this project, we aim to produce stability maps of the N losses and N supplied by the soil at subfield scale for the area under study in the US corn belt. We will improve our estimates by also accounting for application of organic amendments based on the methods of(Luscz et al., 2017, 2015)that spatially distribute county-level census-derived animal inventories for both confined and pastured animals. This method estimates total manure quantities by animal type and number, and distributes that manure to the surrounding agricultural area. We will also synthesize and interpolate available wet and dry atmospheric deposition data from sources such as the National Atmospheric Deposition Program.Field experiments will be conducted at a total of 100 fields located in Michigan, Indiana and Iowa who agreed to participate in this project so that we can validate crop model results, and provide economic analyses for the Variable Rate Nitrogen management scenarios, such as where N:is only applied to the HS Zone and not to the LS Zone;is reduced by 50% in the LS Zone and increased by 50% in the HS Zone;is also reduced in HS zones as well as LS zones.is applied in U zones based on long-term model results and in-season weather observed up to the day of N application, representing a tactical approach based on Basso et al. (2011) and Dumont et al. (2015).We will quantify the probabilities of yield, profit, and environmental outcomes posed by the various N rate and timing management scenarios.The large dataset available to PI Basso will allow the model to be tested under diverse pedo-climatic conditions and management practices. We will collect yield monitored data, and in-season optical and thermal imagery (from UAVs, airborne, and high resolution satellites) to test crop N uptake using a variety of remote sensed vegetation indices.Validation of crop models requires extensive field work.Optical images have been used in agricultural systems for decades, but their use can be limited by cloud cover during the growing season. In this project, we aim to develop new AI approaches based on machine learning to replace pixels affected by clouds by linking pixel with similar performances and changes to soils and position in the landscape. Results from the AI gap fill for cloud images will be validated against airborne and UAV images we will collect at selected farmers' fields and research sites.