Site index & Yield class
Is it time to re-think how we measure forest productivity?
In UK forestry, yield class remains one of the most fundamental measures of site productivity, it is a simple yet powerful way of describing how fast a stand of trees is growing in terms of volume, but are historic top-height/age curves still fit for purpose?
Yield class explained
Yield class is maximum mean annual increment (MMAI), expressed as an integer, and rounded to the nearest even number. In other words, a stand of trees whose mean annual increment (MAI) will peak at 18.5m³ ha⁻¹ yr⁻¹ is growing at yield class 18.
Yield class is used in production forecasting (UK yield tables are mapped to yield class divisions and site treatments). It is perhaps most useful, however, for benchmarking site productivity: for example, a stand of Sitka spruce growing at yield class 24 is somewhat more productive than a stand of Sitka growing at yield class 18.
Yield class cautions
We need to be careful when comparing yield classes between species and regimes, because the age at which MMAI is reached varies greatly. For example, a stand of yield class 22 Sitka spruce will reach its MMAI aged just 39, whereas a stand of yield class 22 Norway spruce will not reach its MMAI until age 61. These stands are theoretically equivalent in terms of volume increment over an infinite cycle of biological rotations, but are not necessarily equivalent if your forecast horizon is shorter. It is an oversimplification to think of them as growing at the same rate.
We must also keep in mind that top height age curves are only available for 16 tree species (with other species mapped to those 16) and for those 16 species, the range of MMAI classes may not be exhaustive. For example, stands of grand fir with an MMAI of <12m³ ha⁻¹ yr⁻¹ or >30m³ ha⁻¹ yr⁻¹ are beyond the extend of the curves. For some species, such as improved Sitka spruce, these 'off the charts' yield class values may not be that extraordinary.
Another common mistake is using yield class alone to estimate volume per ha by multiplying yield class by stand age. This approach works at exactly one point in the rotation – the age of MMAI. To apply this method to a 20-year-old YC20 Sitka stand, will result in an error of 80%.
Yield class vs Site index
Yield class is usually derived using static top height-age curves, which were developed in the 1980s. For a number of reasons, including the approach that was used to derive them, and the nature of the data that was available at the time, these are no longer considered reliable by some. However, they remain widely, if not universally used. They are well bedded into workflows, and many practitioners are simply unaware of alternatives.
In 2021, Manso et al (2020) challenged the top height model status quo by publishing a paper in Forestry exhibiting brand new dynamic top height models for UK tree species. These dynamic models are based on more extensive data, covering a wider range of age-classes, and are far more statistically robust than the historic top height age curves.
These models do not predict MMAI directly, instead, they predict top height at a reference age, a metric better known site index (SI). This is reported along with the reference age, for example, SI50 would be the height of dominant trees when the stand reaches 50 years of age.
Site index is a metric which is used internationally as a benchmark of site productivity. It is more interpretable than yield class, and it isn’t constrained by the extent of the current yield class curves (i.e., <= 36 for Sitka spruce). Further, it could be used in combination with existing top-height dependent yield simulators to eliminate some of the bias in existing production forecasting workflows.
For example, a stand with a top height of 12.00m at year 20, would be treated as yield class 20 based on top height age curves. The dynamic top height model, however, suggests that at year 40 (close to age of MMAI) this stand would have a top height of 26.63m. This suggests a yield class of 24 at the end of the rotation, an uplift of two GYC divisions. Using the MMAI age and volume for YC24 might be more appropriate, but we do need to accept a large error either way; updated yield models are urgently needed.
A final note of caution is that neither the historic top-height/age curves that we use to derive general yield class, nor the newer dynamic models, account for growth differences associated with genetically improved Sitka spruce, or other species with active breeding programmes. There is still plenty more modelling work to do, as more mature improved stands become available for measurement.
Summary
Together, yield class and site index provide complementary perspectives on forest productivity, one grounded in historic yield models, the other in more modern, and statistically rigorous approaches. As dynamic modelling becomes the norm, site index may soon take the lead as the preferred benchmark for UK forest site productivity, but for now, why aren’t more of us making use of both?
Access the site index calculator here
Ben Crisford MICFor
5th November 2025