Crop Growth Model Data Assimilation for Harvest Timing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for determining the optimal harvest time for corn grain are inaccurate due to reliance on historical data, which fails to account for individual variances in corn plant growth, leading to errors in predicting grain moisture content and subsequent discounts or losses during storage and shipment.
Innovation Solution
A computer-implemented system that combines observed agricultural crop phenology data with a data model to estimate growth stage threshold values for specific hybrid seeds at specific locations, using historical and observed data to generate a posterior distribution of growth stage durations and calculate mean and variance values for each growth stage, thereby providing accurate crop growth stage thresholds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If historical data and weather models are used to predict R6 stage, then harvest timing can be estimated, but individual variances in corn plant growth are not accounted for, leading to errors in predicting grain moisture content
Solution Approach 1:
The system incorporates actual field observations of corn growth stages as feedback to continuously update and refine the predictive model. By comparing predicted growth stages with actual observations, the model adjusts its parameters to reduce prediction errors and account for individual plant variances, thereby improving grain moisture content prediction accuracy.
Solution Approach 2:
The system performs preliminary data collection and model calibration before the critical harvest decision period. By gathering historical weather data, hybrid seed information, and previous growth stage observations in advance, the model is pre-positioned to provide accurate predictions when harvest timing decisions are needed, while already incorporating known individual variances.
2Measurement precision
If visual observation of black layer formation is used to determine R6 stage, then physiological maturity can be confirmed, but the corn must be dissected making the method impractical for field use
Solution Approach 1:
The system replaces the mechanical/dissection-based visual observation method with a computational model that uses weather data, hybrid seed characteristics, and previous growth stage observations to predict the R6 stage. This substitution eliminates the need to physically dissect corn plants while maintaining accurate determination of physiological maturity through data-driven predictions.
Solution Approach 2:
The predictive model acts as an intermediary between observable external factors (weather, growth stages) and the hidden internal state (black layer formation). Instead of directly observing the black layer through dissection, the system uses the intermediary model to infer R6 stage based on correlated external variables that are easily measurable in the field.
3Loss of time
If harvest timing is determined using approximated R6 date, then harvest can be planned, but individual fluctuations in grain moisture content at R6 lead to errors in predicting optimal harvest time
Solution Approach 1:
The system dynamically adjusts model parameters based on actual field observations and specific hybrid seed characteristics. By changing parameters such as growth rate coefficients and maturity thresholds to match observed individual plant behavior, the model provides both timely harvest planning and accurate optimal harvest time predictions, resolving the conflict between efficiency and precision.
Data Source
AI summary
A method for estimating growth stage threshold values for a specific hybrid seed at a specific geo-location using historical growth stage data and observed growth stage data comprises using a server computer system, storing a historical crop growth model of one or more hybrid seeds measured from one or more fields over a particular period of time. The historical crop growth model includes growth stage threshold estimates for one or more hybrid seeds. The server computer system receives, via a network, one or more digital measurement values specifying one or more observed growth stage values for a particular hybrid seed at a particular field over a particular period of time. The server computer system transforms the growth stage thresholds into growth stage duration values for the historical crop data and the observed crop data. The server computer system then generates a posterior distribution of growth stage duration values for the particular hybrid seed using a multivariate distribution of growth stage duration value data, which is comprised of historical and observed growth stage data, a covariate matrix describing correlations between different growth stages, and an error matrix used to represent variations in the multivariate distribution. The server computer system estimates mean duration values and variance values for the different growth stages for the particular hybrid seed and then calculates estimated crop growth threshold values for the particular hybrid seed. The server computer system then sends the estimated crop growth threshold values to one or more external computer systems for the purposes of updating and programming crop management instructions.