Satellite-derived and regional weather data as scalable alternatives to on-site stations for forecasting corn tar spot

Tar spot, caused by Phyllachora maydis, has become one of the most damaging foliar diseases of corn in North America. Weather-based forecasting of tar spot severity supports fungicide timing and risk assessment, but it currently relies on on-site weather stations that are costly, spatially sparse, and dependent on careful calibration. This study evaluated, for the first time in tar spot research, whether NASA Prediction Of Worldwide Energy Resources (POWER) satellite data and a regional Mesonet network could substitute for on-site measurements in forecasting disease severity. Daily weather data were obtained from three sources spanning different spatial scales — on-site ATMOS 41 sensors, the Purdue Mesonet, and NASA POWER — and paired with georeferenced severity assessments collected across nine Indiana site-years (2021–2024). Agreement among sources was quantified, and severity was forecast using multiple linear regression (MLR), Bayesian estimation, and autoregressive integrated moving average (ARIMA) and seasonal ARIMA (SARIMA) models, evaluated by leave-one-site-year-out cross-validation. NASA POWER air temperature agreed strongly with on-site sensors (r = 0.81) and relative humidity moderately (r = 0.61), whereas precipitation agreed poorly across all source comparisons (r ≤ 0.25). Across frameworks, SARIMA best captured the temporal structure of epidemic progression , producing smoother severity trajectories, while all three frameworks achieved comparably low prediction errors. Freely available gridded weather data are a viable alternative to on-site 1 sensors for tar spot forecasting, supporting precision agriculture and crop biosecurity.