A weakly supervised framework for high-resolution crop yield forecasts

TL;DR

This paper introduces a deep learning framework that leverages high-resolution input data and low-resolution labels to improve crop yield forecasting at finer spatial scales, addressing data resolution mismatches.

Contribution

It presents a novel weakly supervised deep learning approach for high-resolution crop yield prediction using disparate spatial resolution data, validated across multiple countries and crops.

Findings

Outperforms linear trend and GBDT models in accuracy

Enables high-resolution yield forecasts without high-res labels

Effective across different countries and crop types

Abstract

Predictor inputs and label data for crop yield forecasting are not always available at the same spatial resolution. We propose a deep learning framework that uses high resolution inputs and low resolution labels to produce crop yield forecasts for both spatial levels. The forecasting model is calibrated by weak supervision from low resolution crop area and yield statistics. We evaluated the framework by disaggregating regional yields in Europe from parent statistical regions to sub-regions for five countries (Germany, Spain, France, Hungary, Italy) and two crops (soft wheat and potatoes). Performance of weakly supervised models was compared with linear trend models and Gradient-Boosted Decision Trees (GBDT). Higher resolution crop yield forecasts are useful to policymakers and other stakeholders. Weakly supervised deep learning methods provide a way to produce such forecasts even in the absence of high resolution yield data.