Yield prediction

We develop yield prediction models based on a number of approaches, whose benefits depends on the use case and the spatial scale we look at. At farm level, process-based crop simulation models, such as DSSAT, are very skillful. We further take advantage of the rapidly increasing amount of available data to develop statistical or data driven models to provide forecasts at regional level. Suitable data sources for yield prediction depend on the spatial scale we look at. Usually, we work with remote sensing and climate data, historical yield, soil quality, fertilizer application, or planted cultivar. We constantly question our approaches and try to come up with ideas from a new perspective to design mixed approaches with combinations of different data sources. In addition, we also consider innovative solutions, such as vertical indoor farming, to study theoretical yield achievability. Our analysis help to estimate the feasibility of these projects.