Software — free catalog
GOES-SIF: Near-Real Time Solar-induced Chlorophyll Fluorescence (SIF) Estimation over the United States derived from OCO-3 and GOES-R ABI
A NASA software catalog entry available for reuse or request.
About this technology
Overview
The objective of this research was to train a machine learning model to predict Solar-Induced Chlorophyll Fluorescence (SIF) by fusing data from the Orbiting Carbon Observatory-3 (OCO-3) instrument with multispectral data from the Geostationary Operational Environmental Satellites-R (GOES-R) series of weather satellites. This model was able to predict SIF at 1 km (0.01 degree) spatial resolution with hourly temporal resolution during daytime over the continental United States (CONUS). Validation of the model with reserved data points was performed to ensure that the predictions of SIF showed agreement with observations. A dropout test, where all points within a small region were excluded from the training set, was also performed to ensure that spatial biases were not incorporated into the data through training.