Evapotranspiration (ET) is one of the largest fluxes in the water balance, yet it's also one of the hardest to measure directly. Getting accurate ET data matters for everything from irrigation planning to watershed modeling. The three main direct methods each work at different scales and come with distinct trade-offs.

Lysimeters

A lysimeter is an isolated block of soil (with or without vegetation) placed in a container so you can track exactly how much water enters and leaves. The core idea: if you know precipitation inputs and can measure drainage out the bottom, the difference (adjusted for changes in soil moisture storage) gives you ET.

Weighing lysimeters sit on precision scales and detect mass changes continuously. A decrease in mass between rain events equals water lost to ET. These can resolve ET at sub-hourly timescales and are considered the gold standard for accuracy.
Drainage lysimeters collect percolated water at the base and estimate ET from the water balance. They're simpler and cheaper but less precise, especially over short time periods.

The main limitation is spatial scale. A lysimeter represents only the small plot it occupies, so extrapolating results across a landscape introduces uncertainty. They're best suited for research stations and for validating indirect methods.

Eddy Covariance

Eddy covariance measures ET at the ecosystem scale (typically hundreds of meters to a few kilometers of fetch) by tracking turbulent fluxes in the atmosphere directly above the surface.

How it works:

High-frequency sensors (usually 10–20 Hz) simultaneously measure vertical wind speed and water vapor concentration.
Upward gusts carrying moist air represent water vapor leaving the surface; downward gusts of drier air represent the return flow.
The covariance between vertical wind velocity fluctuations and water vapor fluctuations, averaged over 30-minute intervals, yields the latent heat flux, which converts directly to an ET rate.

Eddy covariance captures real-time fluxes over larger areas than lysimeters, making it popular in flux tower networks like FLUXNET. The downsides: the instrumentation is expensive, the data require significant post-processing (gap-filling, energy balance closure corrections), and the method assumes flat, homogeneous terrain for best results.

Bowen Ratio Energy Balance

This method estimates ET from the surface energy balance by measuring the Bowen ratio ( $\beta$ ), which is the ratio of sensible heat flux ( $H$ ) to latent heat flux ( $\lambda E$ ):

$\beta = \frac{H}{\lambda E}$

The procedure:

Measure temperature and humidity at two heights above the surface.
Calculate the gradients of temperature ( $\Delta T$ ) and vapor pressure ( $\Delta e$ ) between those heights.
Compute the Bowen ratio from these gradients: $\beta = \gamma \frac{\Delta T}{\Delta e}$ , where $\gamma$ is the psychrometric constant.
Combine with the energy balance equation ( $R_n - G = H + \lambda E$ ) to solve for latent heat flux and thus ET.

The instrumentation is simpler and cheaper than eddy covariance. However, the method assumes that the eddy diffusivities for heat and water vapor are equal, which breaks down under stable atmospheric conditions (e.g., calm nights) and near the Bowen ratio value of $-1$ , where the equations become unstable.

Indirect Estimation Methods

When direct measurements aren't feasible, indirect methods estimate ET from meteorological data and vegetation characteristics. These approaches scale more easily across regions and time periods.

Penman-Monteith Equation

The Penman-Monteith equation is the most physically based indirect method. It combines two principles:

Energy balance: the available energy at the surface drives evaporation.
Aerodynamic transport: wind and humidity gradients carry water vapor away from the surface.

The full form is:

$\lambda E = \frac{\Delta (R_n - G) + \rho_a c_p \frac{(e_s - e_a)}{r_a}}{\Delta + \gamma \left(1 + \frac{r_s}{r_a}\right)}$

Where:

$R_n$ = net radiation, $G$ = soil heat flux
$e_s - e_a$ = vapor pressure deficit (the dryness of the air)
$r_a$ = aerodynamic resistance (how easily vapor moves away from the canopy)
$r_s$ = surface (canopy) resistance (how easily water passes through stomata)
$\Delta$ = slope of the saturation vapor pressure curve, $\gamma$ = psychrometric constant

The surface resistance term ( $r_s$ ) is what makes this equation powerful for vegetated surfaces: it accounts for how plants regulate water loss through stomatal conductance. The challenge is that $r_s$ varies with species, soil moisture, and atmospheric conditions, so accurate parameterization is critical.

Reference Crop Evapotranspiration ( $ET_0$ ) and Crop Coefficients

For agricultural and water management applications, the FAO-56 method simplifies things with a two-step approach:

Calculate reference ET ( $ET_0$ ) using the standardized FAO-56 Penman-Monteith equation. This represents ET from a hypothetical grass surface 0.12 m tall, well-watered, with fixed surface resistance (70 s/m) and albedo (0.23).
Apply a crop coefficient ( $K_c$ ) to convert $ET_0$ to actual crop ET: $ET_c = K_c \times ET_0$

Crop coefficients vary by crop type and growth stage. For example, $K_c$ for maize might range from 0.3 during early growth to 1.2 at mid-season, then drop to 0.5–0.6 at harvest. Published $K_c$ values are available from FAO for dozens of crops.

This approach is widely used for irrigation scheduling because it requires only standard weather station data. Its limitation is that it assumes well-watered conditions and may not capture site-specific soil or microclimate effects without further adjustment (e.g., a water stress coefficient $K_s$ ).

Selecting the Right Approach

The best method depends on three factors:

Scale and resolution needs:

Direct measurements (lysimeters, eddy covariance) suit small-scale, high-accuracy studies like field experiments or method validation.
Indirect methods and remote sensing work better for regional assessments and longer time periods, such as evaluating climate change impacts on water resources.

Data availability:

Penman-Monteith requires net radiation, temperature, humidity, and wind speed at minimum. If you lack radiation data, simpler empirical methods (Hargreaves, Priestley-Taylor) may be necessary, though they sacrifice physical rigor.
Assess data quality carefully: missing values, poor instrument calibration, or unrepresentative station locations all propagate into ET estimates.

Study objectives:

Research focused on process understanding benefits from direct methods that capture real fluxes.
Operational applications like irrigation management often favor the $ET_0$ / $K_c$ approach for its simplicity and standardization.
Budget and personnel constraints are real: eddy covariance towers cost tens of thousands of dollars and require trained technicians, while the FAO-56 method runs on a spreadsheet.

Analysis of Evapotranspiration Data

Comparing and Validating Methods

No single ET method is perfect, so cross-validation is standard practice. You might compare lysimeter measurements against eddy covariance data at the same site, or benchmark Penman-Monteith estimates against flux tower observations. Common sources of discrepancy include:

Energy balance closure errors in eddy covariance (typically 10–30% of available energy is "missing")
Lysimeter edge effects or unrepresentative soil disturbance
Parameter uncertainty in $r_s$ and $r_a$ for Penman-Monteith

Spatial and Temporal Patterns

ET varies substantially across landscapes and seasons. Forests typically have higher annual ET than grasslands due to deeper roots and greater leaf area. Wetlands can approach or exceed open water evaporation rates. Seasonally, ET tracks solar radiation and vegetation phenology: it peaks during the growing season and drops sharply in winter or during drought when stomata close.

Investigating these patterns helps identify where water is being consumed in a watershed and how land use changes (deforestation, urbanization, crop conversion) alter the water balance.

Environmental Controls

The main drivers of ET are:

Available energy (net radiation): the primary control in humid environments
Vapor pressure deficit: the primary control in arid environments
Water availability: soil moisture limits ET when supply can't meet atmospheric demand
Vegetation characteristics: leaf area index, rooting depth, and stomatal behavior all modulate how much water plants release

Sensitivity analysis of ET estimates to these variables is valuable for scenario planning. For instance, projecting how a 2°C temperature increase might shift ET rates helps water managers anticipate changes in reservoir inflows and irrigation demand.

Connecting ET to Water Management

ET data feed directly into water balance calculations: $P = ET + Q + \Delta S$ , where $P$ is precipitation, $Q$ is runoff, and $\Delta S$ is change in storage. Accurate ET estimates improve hydrologic models, groundwater recharge assessments, and drought monitoring. For agricultural systems, matching irrigation to actual ET demand improves water use efficiency and reduces waste, which becomes increasingly important in water-scarce regions.

2,589 studying →