Publication News
Two 2025 Papers and One Spatial Prediction Thread
Two papers published in 2025 grew from a shared question: how can spatial information improve both prediction and explanation?
The Thread
The first paper introduces second-dimension outliers for spatial prediction. The second uses geospatial machine learning to identify climate and environmental determinants of wheat production disparities. They are different studies, but they share a concern with spatial context.
Prediction and Explanation
Prediction is not only about lowering error. In geographic problems, prediction also needs to tell us where models behave differently and why those differences matter across regions.
Next Questions
The next challenge is to make spatial models easier to inspect, compare, and explain. That is especially important when results are used in climate, agriculture, remote sensing, and spatial planning.