A Metric for Quantifying Spatial Heterogeneity in Gridded Atmospheric Fields

Abstract Spatial heterogeneity influences many nonlinear atmospheric processes but it is often described qualitatively or with simple statistical measures that do not capture spatial organization. We introduce a mixing‐oriented metric for gridded fields based on the expected deviation of subdomain means from the global average. This metric captures a multiscale aspect of spatial organization and can be normalized for comparisons across fields or scenarios. We apply it to idealized spatial patterns, real emission inventories, and a set of large‐eddy simulations of aerosol coagulation with identical initial total particle numbers but varying spatial configurations. For these idealized coagulation simulations, larger initial metric values are associated with faster coagulation and lower final number concentrations, consistent with theoretical expectations for clustered particle fields. The metric provides a quantitative framework for comparing departures from spatial uniformity in model fields and for interpreting how spatial organization may influence nonlinear processes. This approach can support more systematic assessments of spatial heterogeneity in aerosol modeling and other applications involving gridded atmospheric data.