Microscale nonrandomness revisited: An overlooked predictor of ecological change

Microscale nonrandomness metrics (Morisita Iδ from frequency distributions and Moran’s I/SADIE from spatial patterns) consistently improve prediction of short-term ecological change across four contrasting datasets.

• Why it matters: Early warning signals are often tested for systems nearing transitions under gradual external change, but their usefulness for short-term dynamics without directional forcing is unclear.
• What they did: The framework was applied to time series of phytoplankton–parasitic chytrid interactions, leaf beetles and aphids on willow in crops, and forest vegetation–soil snapshot data, using microscale nonrandomness from frequency and spatial patterns.
• The result: Models that included microscale nonrandomness showed greater cross-validated predictive performance than models using only macroscale variables or spatial averages, with frequency-based metrics sometimes matching or exceeding spatially explicit ones.

AI summary of the abstract · may contain errors · not medical advice · read the paper.