Predicting plant water use is an important component of agricultural, ecological, and climate modeling. Plants regulate their water use by adjusting microscopic pores on their leaves known as stomata. When exactly stomata open and close is notoriously difficult to predict accurately but sophisticated equipment is widely available to measure this phenomenon and many models exist and are in use to predict stomatal behavior. However, the challenges of collecting good data to be able to use these models is not often discussed, and the reliability of published model parameters are questionable.

In this study, we looked at the difficulties that arise when measuring stomatal conductance of leaves. We used new data from different experiments and a detailed 3D computer model of a virtual plant to help us disentangles the sources of these difficulties. We found that leaf-to-leaf variation in stomatal behavior and slow stomatal kinetics significantly challenge model calibration.

To improve these measurements, we provide recommendations of when to sample, how many samples to achieve a desired accuracy, and how to mix different measurement types to improve efficiency and reliability.