National Crop Yield Forecasting Using Satellite Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Accurately forecasting national crop yields during the growing season is challenging due to the difficulty in obtaining reliable data at a national level, as local and regional measurements are geographically distributed and costly to collect and calibrate, especially with existing survey-based methods and crop simulation models.
Innovation Solution
A computer-implemented method using a server system that aggregates agricultural data records, including remotely sensed spectral properties and soil moisture data, to create geo-specific time series, which are then used to calculate state and national crop yields through linear regression and adjustment modules, enabling accurate forecasting during the growing season.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If survey-based data collection by NASS is used, then data can be obtained from farmers, but farmers are unable to provide good estimates until harvest time, making it useless for forecasting during the growing season
Solution Approach 1:
The patent introduces an intermediary system that collects satellite imagery, weather data, and soil moisture data to serve as a mediator between the need for timely forecasts and the unavailability of farmer estimates during the growing season. This intermediary data collection system provides the necessary information to generate forecasts before harvest time.
Solution Approach 2:
The system performs preliminary data collection and processing during the growing season by continuously monitoring satellite imagery, weather conditions, and soil moisture levels. This preliminary action enables forecast generation well before harvest time, resolving the timing contradiction.
2Measurement precision
If crop simulation process models are used to predict regional yields, then forecasts can be generated during the growing season, but the cost of collecting numerous local inputs and calibrating parameters makes it too expensive for national level use
Solution Approach 1:
The patent merges multiple data sources (satellite imagery, weather data, soil moisture data) into a unified forecasting system. By combining these diverse inputs through a standardized processing framework, the system achieves national-level forecasting capability without the excessive complexity of traditional process models that require separate calibration for each region.
Solution Approach 2:
The system creates a universal forecasting approach that works across all agricultural regions simultaneously. Rather than requiring region-specific calibration for each local area, the unified model processes all regions through the same framework, dramatically reducing the complexity and cost of national-level implementation.
3Measurement precision
If local and regional measurements are collected to create national forecasts, then accurate regional data can be obtained, but the geographically distributed nature of data makes it difficult and costly to obtain during the growing season
Solution Approach 1:
The patent replaces the mechanical system of manual data collection from distributed regional sources with an automated remote sensing system. Satellite imagery and automated weather stations substitute for human surveyors traveling to collect data, dramatically improving data collection efficiency during the growing season while maintaining regional measurement precision.
Data Source
AI summary
A method for determining national crop yields during the growing season is provided. In an embodiment, a server computer system receives agricultural data records for a particular year that represent covariate data values related to plants at a specific geo-location at a specific time. The system aggregates the records to create geo-specific time series for a geo-location over a specified time. The system creates aggregated time series from a subset of the geo-specific time series. The system selects a representative feature from the aggregated time series and creates a covariate matrix for each specific geographic area in computer memory. The system determines a specific crop yield for a specific year using linear regression to calculate the specific crop yield from the covariate matrix. The system determines a forecasted crop yield for the specific year using a sum of the specific crop yields for the specific year, as adjusted.