A weakly supervised framework for high-resolution crop yield forecasts
This work addresses the need for high-resolution crop yield forecasts for policymakers and stakeholders, but it is incremental as it applies a weakly supervised method to a known data availability bottleneck in agricultural forecasting.
The paper tackled the problem of crop yield forecasting when predictor inputs and label data are at different spatial resolutions by proposing a weakly supervised deep learning framework that uses high-resolution inputs and low-resolution labels to produce forecasts at both levels, evaluated by disaggregating regional yields in Europe for five countries and two crops, showing performance compared to linear trend models and Gradient-Boosted Decision Trees.
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.