Why These Tools Work

Thermal cameras are widely used in plant research to infer stomatal conductance — how open a leaf's stomata are. The logic is intuitive: open stomata allow transpiration, which cools the leaf. A warmer leaf suggests closed stomata; a cooler leaf suggests open ones.

But this signal is not always strong. The relationship between leaf temperature and stomatal conductance depends heavily on ambient conditions: air temperature, humidity, and wind speed all influence how much a change in conductance actually shifts the leaf's temperature. Under some conditions, the thermal signal is robust. Under others, it is nearly undetectable.

The Sensitivity Index Squantifies this — it answers the question: “If stomatal conductance changes right now, how much will leaf temperature change in response?”

The framework is derived from the leaf energy balance equation (Penman-Monteith). Rather than modelling the full energy balance, it reduces to two key parameters that capture how ambient conditions shape the leaf temperature response to stomatal changes.

c is a dimensionless scaling coefficient. It represents how sensitive the leaf-to-air temperature difference is to a change in stomatal conductance. A larger |c| means the leaf temperature shifts more per unit change in conductance — a stronger, more detectable signal.

T_wet is a wet-bulb proxy temperature — the temperature a fully transpiring leaf would reach under the current conditions. It anchors the lower end of the leaf temperature range.

S is the sensitivity index: the magnitude of the distance between the wet-bulb temperature and the baseline leaf temperature, scaled by the conductance-temperature coefficient. Larger S means thermal imagery will more clearly distinguish differences in stomatal aperture.

The threshold S > 10 is a practical recommendation from the source paper. It does not represent a hard physical boundary — it is a heuristic based on the relationship between S and the ability to discriminate stomatal conductance values with realistic sample sizes and measurement noise typical of field thermal cameras.

Below this threshold, the temperature differences between leaves with different conductance values become small enough that measurement noise, canopy heterogeneity, and image artefacts tend to overwhelm the signal. Above it, a meaningful physiological contrast is likely to be detectable.

The platform uses three bands to communicate certainty:

• Recommended (S > 10) — Conditions are suitable. Thermal imaging is likely to return physiologically meaningful data.
• Borderline (7 ≤ S ≤ 10) — Usable with care. Results may be noisier; increase sample size if possible.
• Low sensitivity (S < 7) — Not ideal. The stomatal signal is likely to be obscured by environmental noise.

This release uses the sunlit leaf version of the simplified equations. The sunlit model assumes the target leaves are directly illuminated, which is the condition where the stomatal thermal signal is strongest and most consistent.

A shaded-leaf version and a canopy-level model exist in the literature. These will be available in a future version of this mission as optional model modes. For most field protocols targeting sunlit canopy leaves, the sunlit model is the appropriate starting point.

The baseline leaf temperature T_l = 25°C is a fixed simplification in this release. Future versions will support user-specified or image-derived T_l values.