Leaf wetness.

Most foliar fungal and bacterial diseases, botrytis, downy mildew, the blights, share one requirement: they need free water on the leaf for a minimum number of hours to germinate and infect. So the single most predictive number for disease isn't temperature or humidity on their own. It's leaf wetness duration, the number of continuous hours the leaf surface actually stays wet.

A disease needs three things at once: a susceptible plant, a pathogen present, and an environment that favors it, the disease triangle. For foliar pathogens, the environmental leg is mostly about water on the leaf surface. A spore that lands on a dry leaf does nothing; the same spore on a leaf that stays wet long enough germinates, pushes a thread into the tissue, and the infection is underway. The clock that matters runs in hours of surface wetness, which is why a number nobody watches turns out to be the one that forecasts outbreaks.

The wet window decides infection.

Free water arrives on the leaf in ways the air's humidity reading can miss entirely: dew settling overnight as surfaces cool below the dew point, guttation droplets the plant pushes out on its own, condensation, or overhead irrigation that never fully dried. A canopy can be soaking wet while the air sensor a meter away reads a comfortable humidity. The leaf surface is the thing the pathogen lives on, so the leaf surface is the thing to measure.

Duration, not the moment.

What infects a crop is not that the leaf got wet but how long it stayed wet. A brief wetting that dries within an hour rarely matters; a wet spell that runs eight, ten, twelve hours crosses the germination threshold for one pathogen after another. So leaf wetness is fundamentally a duration measurement: start the clock when the surface goes wet, stop it when it dries, and count the continuous hours. That accumulated wet-hours figure, not a snapshot, is what feeds a risk decision.

Pair it with temperature, and run the model.

Pathogens have a temperature preference layered on the wetness requirement: each has a published curve of "this many wet hours at this temperature equals an infection event." Combine leaf wetness duration with leaf-or-air temperature and you can run those infection models and get a risk signal before lesions appear, the heart of real integrated pest management. Instead of spraying on a calendar, you act when the model says risk is rising: spray preventively, or better, change the environment to shorten the wet window, more airflow and a touch of heat or venting to drop the canopy below saturation and dry the leaf faster. That ties leaf wetness to VPD and airflow, the levers that decide how long water lingers, and the whole picture lives next to the rest of the threat environment.

How to read it.

A leaf-wetness sensor is an interdigitated grid, a comb of interlocking conductive traces on a flat surface, usually painted or coated to mimic a real leaf and mounted at canopy height, angled like a leaf. When water bridges the traces, the electrical resistance drops (or capacitance rises) and the sensor reads "wet"; as it dries, the reading climbs back. Like the matrix moisture sensor, it must be AC-driven so the measuring current doesn't slowly corrode the traces. The DIY version is exactly this: an interdigitated PCB and a single ADC channel with an AC excitation, and it lands in oat-ods through the gatherer like any other analog input, with the duration accumulated from the wet/dry transitions.

The trap: the calendar spray.

The trap is spraying on a fixed schedule, fungicide every Sunday whether the week was risky or not, which wastes product, breeds resistance, and still misses the high-risk nights. The other trap is steering by air humidity alone, when the leaf can be wet from dew or guttation while the air isn't saturated, or dry under high humidity if it's moving air. Measure the surface, count the hours, and let the model, not the calendar, decide when to act.