The hyperspectral satellite sensing capacity is expected to increase substantially in the near future with the planned deployment of hyperspectral systems by both space agencies and commercial companies. These enhanced observational resources will offer new and improved ways to monitor the dynamics and characteristics of terrestrial ecosystems. This study investigates the utility of time series of hyperspectral imagery, acquired by Hyperion onboard EO-1, for quantifying variations in canopy chlorophyll (Chl<sub>c</sub>), plant productivity, and yield over an intensive farming area in the desert of Saudi Arabia. Chl<sub>c</sub> is estimated on the basis of predictive multi-variate empirical models established via a machine learning approach using a training dataset of in-situ measured target variables and explanatory hyperspectral indices. Resulting time series of Chl<sub>c</sub> are translated into Gross Primary Productivity (GPP) and Yield based on semi-empirical relationships, and evaluated against ground-based observations. Results indicate significant benefit in utilizing the full suite of hyperspectral indices over multi-spectral indices constructible from Landsat-8 and Sentinel-2.