Crop Evapotranspiration

Evapotranspiration bundles two losses that are hard to separate in the field and pointless to separate in a water balance: evaporation from the soil surface, and transpiration through the plant. The FAO-56 method computes the total in two steps that split the problem along a genuinely useful seam — one term for what the atmosphere is demanding, one term for what this particular crop does about it.

Reference evapotranspiration, $ET_0$, is the demand side. It is defined as the water use of a hypothetical, uniform, well-watered grass surface 12 cm tall with a fixed surface resistance, computed from temperature, humidity, wind and solar radiation by the Penman-Monteith equation. The definition is deliberately rigid so that the number describes the weather and nothing else: two fields under the same sky have the same $ET_0$ whatever is growing on them. Weather networks publish it daily, which is why the method is practical at all.

The crop coefficient $K_c$ is the crop side, and it is an empirical ratio measured against that reference — typically 0.3 to 0.5 for bare or barely covered soil early in the season, rising to 0.95 to 1.20 at full canopy, and falling again as the crop senesces. Mid-season maize at $K_c = 1.15$ under a reference ET of 6.5 mm/day is using $1.15 \times 6.5 = 7.475$ mm/day. A coefficient above 1 is not a paradox: a tall crop with a rougher surface than clipped grass extracts more energy from moving air than the reference does.

The assumption buried in the coefficient is that the crop is well watered. Tabulated $K_c$ values describe a crop transpiring freely, so multiplying by them gives potential use, not actual use; a crop already short of water uses less, and applying the full figure to a stressed field overstates the requirement. FAO-56 handles this with a separate water stress coefficient. The other caution is that early-season coefficients are dominated by evaporation from bare soil, which depends on how often the surface is wetted — so a light, frequent irrigation schedule raises the crop coefficient it was computed with, and the calculation quietly feeds back on itself.