Because negative yield anomalies can accompany both dry and
moist conditions in Brazil, resulting from droughts, floods and
moisture-loving pests or diseases, simple regression-based yield forecasts using moisture-related satellite indices can be confounded in
some regions. A more robust approach to yield estimation would
involve integration with crop modeling systems accounting for likelihood of pest and disease outbreak. The regression analyses presented
here provide insight into when and where different indicators are likely
to add significant value to yield forecasts.