How Reliable Is the FAO-56 Method? The Original Authors Answer, 17 Years Later
In 2015 the original FAO-56 authors published a review 17 years on: what held up against later research, what didn't, and why the gap between standard and actual crop coefficients explains part of the criticism of the method.
Who better than the original authors to take stock
In the first article of this series we covered what the FAO Irrigation and Drainage Paper No. 56 (1998) says: why the Penman-Monteith equation became the world standard for calculating evapotranspiration. Nearly three decades on, does the method still hold up? The answer came from the four original FAO-56 authors themselves, Luis Pereira, Richard Allen, Martin Smith and Dirk Raes, who in 2015, 17 years after the original publication, wrote a review dedicated to exactly that: what worked, what didn't, and what's still needed.
It is rare for the people who created a standard to come back and assess it against the research accumulated since. Below is what they wrote, explained for non-researchers.
The paper at a glance
| Section | Content |
|---|---|
| 🎓 Paper | Pereira, L.S., Allen, R.G., Smith, M., Raes, D. (2015). Crop evapotranspiration estimation with FAO56: Past and future. Agricultural Water Management, 147, 4-20 |
| ❓ Question | 17 years after FAO-56's publication, how valid are its guidelines in light of the research, data, and modelling capabilities that have become available since? |
| 🔬 What they did | Systematic review of FAO-56's four main pillars (reference ET, single crop coefficient, dual crop coefficient, non-standard conditions) checked against research published since 1998 in very different climates |
| 📈 What they found | The Penman-Monteith reference ET method held up in comparisons across very different climates; the dual crop coefficient improved the estimate of soil evaporation and is now used in many applications; only the 'standard' Kc, measured on well-watered crops, is transferable, while the 'adjusted/actual' Kc is not, and according to the authors some of the criticism of Kc transferability comes from mixing the two up |
| 🌱 What it means for your garden | A Kc value found online for 'lawn' or 'hedge' only applies if it represents optimal conditions (well-watered, unstressed). If the source measured it on stressed plants or in a different climate from yours, applying it as-is produces wrong estimates |
| 🗺️ How it connects to SprinklerMap | SprinklerMap uses fixed peak-summer water-need values by vegetation and exposure (5 mm/day for a full-sun lawn, ×0.8 in partial shade, ×0.6 in shade) to size the system; day-to-day scheduling still has to follow local weather, because, as the authors point out, coefficients depend on real conditions |
| 📚 References | See the References section at the end of the article |
The question: does a 1998 method still hold up today?
When a scientific method becomes a global standard, it risks no longer being tested: everyone uses it because 'it's the standard', without anyone checking that it still deserves to be. Pereira and colleagues address this risk head-on: FAO-56 had been validated mainly against data available through the late 1990s. Since then, decades of additional lysimeter networks, tens of thousands of weather stations, and satellite remote sensing for estimating evapotranspiration have arrived, tools that did not exist in their current form in 1998.
So the question isn't rhetorical: with all this new data, has the Penman-Monteith method standardised in FAO-56 held up to scrutiny, or has later research found cracks that couldn't be seen back then?
What they did: a systematic check on four fronts
The authors organised the review around FAO-56's four main contributions, checking each against later literature. First: reference ET (ET0) calculated with Penman-Monteith, checked against comparative studies and regional analyses in very different climates: from semi-arid Spain (Berengena and Gavilán, 2005) to southern Italy (Steduto et al., 2003) and China, where a sensitivity analysis of 41 years of daily data from 150 national weather stations found the result most sensitive to relative humidity (Gong et al., 2006).
Second: the single crop coefficient (Kc), checked against new studies on crops not present in the original edition, from quinoa to teff, from tree crops like peach and citrus to natural vegetation in semi-arid regions. Third: the dual crop coefficient approach (Kc = Kcb + Ke, which separates plant transpiration from direct soil evaporation), checked against applications on annual crops, orchards and vineyards. Fourth: procedures for non-standard conditions such as water stress, salinity, mulching and intercropping.
What they found: what held up, and the distinction that matters most
The most solid finding concerns the heart of the method: the Penman-Monteith equation for reference ET was largely confirmed by comparative studies against local ET measurements, and regional analyses confirm it applies across a wide range of environments. The starting idea was that physics is the same everywhere, so a physically grounded equation should work in any climate, and later comparisons largely bore that out.
The dual crop coefficient (Kcb + Ke), which separates plant transpiration from soil evaporation, has been adopted in many applications: Idaho and Nevada, for example, used it to revise statewide water requirements at more than 100 weather stations each. For orchards and vineyards, where canopy size, ground cover and shading make Kc vary widely, the authors point to the method of Allen and Pereira (2009), based on the fraction of ground covered or shaded by the crop and its height.
The part of the paper with the most direct practical payoff is the distinction between 'standard' and 'adjusted' (or 'actual') Kc. Standard Kc represents a crop under optimal conditions (well-watered, unstressed) and is the one transferable from one climate to another with the adjustments the method provides. 'Adjusted' Kc, on the other hand, represents the real plant, often under water or salinity stress or managed sub-optimally, and by definition is not transferable: it depends on site-specific conditions that don't repeat elsewhere. The authors write that some of the criticism about 'poor transferability' of Kc comes from not appreciating the difference between standard and actual Kc, and comparing values measured on crops under different conditions.
What it means for your garden
In practice, if you find a Kc value online for 'lawn' or a specific hedge, that number only applies if it describes a plant under optimal conditions, and not a plant that, in the source's own photo, was clearly drought-stressed or grown in a climate with very different humidity and wind from yours. Using an 'adjusted' Kc measured elsewhere as if it were a transferable standard leads to wrong water-need estimates: the review points out that adjusted Kc values vary widely from site to site, which is why they can't be transferred.
The second practical consequence concerns the quality of input weather data: the review devotes an entire section to the need for weather data quality control (QA/QC), because a physically accurate equation fed poor weather data still produces poor results: garbage in, garbage out applies to Penman-Monteith too. A smart controller pulling data from a distant or poorly sited weather station inherits this same limitation.
How this connects to SprinklerMap
As in the first article: SprinklerMap doesn't calculate daily ET0 with Penman-Monteith and doesn't adjust crop coefficients with local weather data. It uses fixed peak-summer water-need values for each vegetation type and exposure (5 mm/day for a full-sun lawn, multiplied by 0.8 in partial shade and 0.6 in shade) as a design estimate to size litres and run times. The review explains why those values aren't universal: the real need depends on climate, plant condition and data quality, so day-to-day scheduling should be corrected with local weather and observation. For the practical run-time calculation side, see How long to run sprinklers: minutes per cycle by soil and head type, with the math behind them.
The short version
17 years after publication, the FAO-56 authors themselves put their own method under review, and the conclusion is that it held up: Penman-Monteith held up in comparisons across very different climates, the dual crop coefficient has found wide use, and the weak points identified are more about application (data quality, confusing standard with actual Kc) than the method itself. For anyone watering a garden, the most useful practical takeaway is exactly this: a crop coefficient found online only applies if it represents optimal conditions, not the stressed plant in the photo next to it.
References
Pereira, L.S., Allen, R.G., Smith, M., Raes, D. (2015). Crop evapotranspiration estimation with FAO56: Past and future. Agricultural Water Management, 147, 4-20. doi.org/10.1016/j.agwat.2014.07.031
Allen, R.G., Pereira, L.S. (2009). Estimating crop coefficients from fraction of ground cover and height. Irrigation Science, 28(1), 17-34. doi.org/10.1007/s00271-009-0182-z
First article in this series: Why Evapotranspiration Is the Foundation of Modern Irrigation: /en/blog/evapotranspiration-modern-irrigation-fao56-explained
Frequently asked questions
Did the authors change their minds compared to 1998?
No. The people who created the standard went back to assess it and confirmed the original framework. Pereira, Allen, Smith and Raes took the four pillars of FAO-56, meaning reference ET, single crop coefficient, dual crop coefficient and non-standard conditions, and compared them against the research published since 1998. Comparative studies and regional analyses confirm that Penman-Monteith reference ET works across a wide range of environments. The part that needs most care is the crop coefficients, where the review insists on the distinction between standard values and values adjusted for real conditions.
What did they flag as real limitations?
Three above all, all practical rather than theoretical. The first is the quality of the input weather data: the equation is reliable, but if the humidity or wind figures are poor or come from an unrepresentative station, the error enters there and the method cannot correct it. The second is the limited availability of crop coefficients for non-traditional crops: the tables cover well what has been studied most, much less the rest. The third concerns fruit trees and vineyards, where the canopy structure and partially shaded soil make application harder than on a uniform crop. For a domestic garden the first limitation counts most: the calculation is worth exactly as much as the weather data you feed it.
Does the standard/adjusted Kc distinction apply only to farming, or to home gardens too?
It applies to the garden as well, through the same mechanism. The distinction is between a standard Kc, measured under optimal conditions on well-watered, unstressed plants, and an actual Kc, measured on plants in less than ideal conditions. Only the first is transferable, because it is designed to be comparable across places and climates; the second describes one specific situation and cannot be moved elsewhere. In practice a well-watered lawn has a different Kc from a lawn stressed by drought, and if you take a value found online without knowing the conditions it was measured under, you risk applying to your garden a number that described another plant in another climate. That is why the authors insist on this label.
Are there even more recent reviews?
Yes: Pereira, with a largely different group of co-authors (Paredes, Hunsaker, López-Urrea and Mohammadi Shad), published another update in Agricultural Water Management in 2021 devoted to crop coefficients for field crops, a possible subject for a future article in this series. The direction of that work is consistent with the 2015 review: the reference equation is left alone, and what gets refined is the coefficients, which is where real-world variability is greatest and where picking the wrong value has the largest effect on the final result. For anyone designing garden irrigation the practical message does not change: the stable part of the calculation is reference ET, while with Kc values it always pays to know where the number came from.
Technical note: This article explains, in plain language, the findings of a published scientific study, necessarily simplifying the original methodology to make it readable. For the full statistical analysis, the limitations the authors themselves note, and the experimental details, always refer to the original paper linked in the references at the end of the article: a study's results shouldn't be generalised automatically to your specific case.