Observing the Non-Optically Active: An Observability Budget for Inferring Stormwater BMP Nitrogen-Removal Performance from Earth Observation

Saurav Kumar 

Some regulatory load-accounting programs credit distributed stormwater best management practices (BMPs) using standardized, type-based nitrogen-removal efficiencies, yet field records show highly variable removal with at most a weak age signal. Earth observation (EO) can map aspects of a practice's physical condition, such as open-water extent and vegetation, and may provide indirect storage-related proxies, but it cannot directly observe the pollutant removal that regulation credits. This paper asks whether assimilating observed condition through a process model may recover otherwise unobservable performance, and it advances a single principle. On a proper, held-out score, assimilating remotely observed condition appears to recover treatment performance according to how much of that performance is routed through observable condition, and little beyond it. The principle is examined with a calibrated observing-system simulation experiment in which a known synthetic truth is observed with noise and reconstructed. On disjoint held-out seeds, the observability-budget skill, defined as the CRPS skill gain, the fraction of the gap to a perfect-performance reference that assimilation closes, rises with the observable share of hidden-rate-driven performance. It is only about 16% when the rate modifier is fully hidden. It reaches about 65 to 76% where the rate modifier is fully explained by observed condition, although those fully observable cells are overconfident under the known-forcing idealization and are read as an upper bound. Storage-routed aging performance is separately recovered from about 24% to 68%, because storage is observed. Because these are properties of the simulated system under stated assumptions, they are intended to bound rather than measure the operational budget, positioning the observability budget as a design quantity that targeted monitoring investment could enlarge.