Leaf optical property modelling is a tool that has been used for many decades to describe the interactions of light with plant leaves. Using these tools, measurements of light reflected and transmitted from plant leaves can be used to estimate the biochemical and biophysical properties remotely. Light in the ultraviolet, visible, and near infrared regions of the spectrum is affected by different biochemical components within the leaf sample and the biophysical properties have the potential to affect the whole wavelength spectrum. Furthermore, the angle of illumination and observation affect both the spectral shape and intensity of the reflected and transmitted light spectra.

Some of the most widely used leaf optical property models use spectral data that do not describe the angular components of reflectance and transmittance. While this simplifies the model and results in a more predictable spectral result from defined biochemical parameters, it is difficult to apply this model to field data. The data used in these models requires hemispherical measurements of reflected and transmitted light whereas data collected in the field are typically limited to single angle with variable illumination angle and polarization. These single angles are not always representative of the hemispherical spectra expected by the model and are affected to a greater degree by the biophysical attributes of the leaf.

In this work, an extension to the PROSPECT-D leaf optical property model was developed to combine the effective hemispherical modelling developed by previous researchers with the spatial component that is related to the angular measurements obtained in a field setting. This was done by connecting physically measured attributes of plant leaves to the spatial distribution of the reflected spectra.

Through the course of this work, it was hypothesized that the modelling capabilities could be further improved by calibrating the model with a more diverse and agriculturally-based calibration dataset. However, it was found that the approach to this problem may be better aligned with calibrating the model with species specific datasets for targeted applications. In short, a one-size-fits-all model may not be the best solution to improve biochemical and biophysical parameter acquisition using leaf optical property modelling.