A note on the reporting of stocking rate, stocking density, and “grazing intensity” in pasture and rangeland research

It has been an observation of mine over the last few years that in articles dealing with animal allocation to pasture or rangeland in various countries, there are differences between authors in the units used for some common terms such as stocking rate, stocking density or grazing intensity when reporting experiment treatments and data. Some authors are using units incorrectly, in my opinion. Hence, I thought that it would be useful to write an editorial for Grassland Research, setting out a logical framework and units for authors' consideration when preparing manuscripts on this topic. This is neither an in-depth review nor an attempt to redefine terms and concepts, but simply a call for authors to use units correctly within the currently accepted framework and definitions.

To begin, the authoritative reference is Allen et al. (2011). Here, stocking rate is defined as “the relationship between the number of animals and the total area of the land in one or more units utilized over a specified time,” with a note, “where needed, it may be expressed as animal units or forage intake units per unit of land area over time (animal units over a described time, per total system land area).” Meanwhile, stocking density is defined as “the relationship between the number of animals and the specific unit of land being grazed at any one time; an instantaneous measurement of the animal-to-land area relationship” and grazing pressure is defined as “the relationship between animal live weight and forage mass per unit area of the specific unit of land being grazed at any one time; an instantaneous measurement of the animal-to-forage relationship.” Extrapolating from these definitions, relevant units for stocking rate would be animals (of a specified species and class) per ha, and for grazing pressure, it would be kg animal body weight per kg of forage mass. Grazing pressure and its reciprocal, forage allowance (Allen et al., 2011; Sollenberger et al., 2005), are unitless ratios, with both animal live weight (kg) and forage mass (kg) expressed for the same land area. Considering stocking rate, nine sheep on 2 ha for 6 months of a year with 6 months with plots ungrazed is not the same as nine sheep on 2 ha continuously throughout the year. This distinction could only be represented in the units if time were included in both the numerator and the denominator (e.g., animal. years per ha. year) in which case, time (years) would cancel out. Hence, especially where animals graze a pasture for only a part of a year or other time period, as in extreme environments such as the Qinghai-Tibet Plateau, it should be explicitly stated over what period of time the animals are allocated to the land area and how any fluctuation in animal number, body weight or land area over the reporting period is dealt with.

Accordingly, in the writer's home country, New Zealand, stocking rate has been historically reported on sheep and beef farms in sheep stock units per ha and on dairy farms as cows per ha. For sheep and beef farms, one sheep stock unit was defined more than 50 years ago as a breeding ewe raising a lamb and consuming 550 kg DM per year (Hoogendoorn et al., 2011; Parker, 1998). Rams, hoggets and deer and cattle of various classes were allocated sheep stock unit values proportionate to their expected annual forage consumption. To standardize for variation in animal numbers as lambs and calves were born in spring and sold in autumn, the reported stocking rates were usually the number of animals carried on the farm through the winter months. With an increase in lambing percentage through the years, a ewe typically now rears about 1.3–1.5 lambs on average (rather than a single lamb) and consumes around 620 kg DM per year. With animal body weight accounted for in this way, and with the particular systems management emphases in New Zealand, a grazing pressure metric was not needed and never evolved, even though it can be useful for comparison across environments or management practices (Sollenberger et al., 2005).

On New Zealand dairy farms, some cows die at calving and cows that will not be kept on the farm in the following year are usually culled progressively as the feed supply diminishes in summer with intensifying soil moisture deficit, so the cow number declines through the milking season. The cow number for stocking rate calculations has normally been taken as the number of cows on farm at the peak of the milking season. Previously, New Zealand dairy farms were largely self-contained from a feed supply perspective, so stocking rate was used in extension circles as a comparative unit of farm performance. In recent decades, there has been significant import of feed supplements such as palm kernel or maize silage from off farm, with variation between farms in the quantity of feed imported.

In order to arrive at a standardized measure of the animal-to-forage relationship at farm level, a measure called comparative stocking rate has emerged, defined as kg cow body weight per tonne of pasture plus supplement annual feed supply on the farm (Macdonald et al., 2008). For example, a farm with a stocking rate of 3 cows ha⁻¹, cows weighing 500 kg on average and annual pasture yield of 12 tonnes DM with 3 tonnes ha⁻¹ year⁻¹ maize silage DM fed would have a comparative stocking rate of 100 kg cow body weight per tonne feed. Comparative stocking rate has proved to be a useful metric in helping farmers improve feed conversion efficiency. Where comparative stocking rate is too low, feed supply is in surplus and feed utilization will be correspondingly low and where comparative stocking rate is too high, the metabolic allocation of feed energy between cow body maintenance and milk production will swing in favor of body maintenance, in both cases reducing feed conversion efficiency. In New Zealand pasture-based dairy systems, a comparative stocking rate in the vicinity of 80 kg cow body weight per tonne feed has proved to be optimal and surprisingly, when calculated for a pasture-based tropical beef production system in Sabah, the optimal comparative stocking rate value was similar to this (Gobilik, 2017).

A need of farmers, extension staff and researchers in reporting grazing studies (especially studies involving rotational stocking or mob stocking) is for a term that defines how animals are allocated to a grazing event (as distinct from allocation of animals to a land area or quantity of forage). This is an animal-to-land area relationship but it is neither a stocking rate nor an instantaneous measure; the grazing event may occur over a few hours (e.g., half a day) or a few days (e.g., 2 days), so there is a time dimension involved. Moreover, unlike forage allowance or grazing pressure, the available forage during the grazing event is not considered. It appears that none of the terms defined by Allen et al. (2011) exactly fits this entity, which, following usage in New Zealand, we will refer to here as grazing intensity—the potential animal demand for feed over the time course of a grazing event, with units, animal·days per ha (number of animals × duration of grazing event/area grazed). Potential animal demand is used here, because both rotational stocking and mob stocking use competition between animals for available forage (i.e., hunger) to suppress animal selectivity and create a “lawnmower” effect. Therefore, actual feed consumption per animal is not defined by the number of animals, as animal forage intake decreases with an increase in the number of animals in such a grazing event. In New Zealand, grazing intensity as defined above is typically considered in conjunction with forage harvested since the quotient of forage harvested and grazing intensity equals forage intake (kg DM ha⁻¹/animal·days ha⁻¹ = kg DM animal⁻¹ day⁻¹).

By varying rotation length, both stocking density and grazing intensity can be varied on the same farm at the same stocking rate. A high grazing intensity (long rotation) reduces animal intake compared to a lower grazing intensity (shorter rotation) and is used by New Zealand farmers to ration autumn-saved stockpiled feed during winter. For example, on small demonstration farms of 0.8 ha carrying 16 sheep (20 sheep ha⁻¹), grazing 0.10 ha per 2 days yielded a grazing intensity of 320 sheep·days ha⁻¹ and a 16-day rotation with a high animal intake. In contrast, grazing 0.05 ha per 3 days yielded a grazing intensity of 960 sheep·days ha⁻¹ for each grazing event and a 48-day rotation length with a restricted animal intake (Matthew et al., 2017).

The difference between stocking density and grazing intensity as measures to define a grazing event is evident in the experiment of Tracy and Bauer (2019). These authors compare mob, rotational, and continuous stocking by cattle at the same stocking rate of 11.5 animals ha⁻¹ (though they incorrectly use stocking rate units of animal·months ha⁻¹ with time in the numerator only—see further comment below). In this experiment, the mob and rotational grazing treatments have stocking densities of 109 and 14 animals ha⁻¹, respectively. However, because of different grazing durations and areas of 1 day per 0.1 ha in mob stocking and 4 days per 0.8 ha in rotational grazing, the grazing intensity (if it had been calculated) differs by a factor of just 2, being eight animals × 1 day/0.1 ha = 80 animal days ha⁻¹ for mob stocking and 8 animals × 4 days per 0.8 ha = 40 animal days ha⁻¹ for rotational grazing.

Lastly, in terms of stocking rate theory, if it is intended to create a measure of animal feed demand in a grazing event from an ecosystem perspective of flow of energy from sunlight to plants to animals, accounting for the animal component of the ecosystem as kg body weight is not ideal, since animal body maintenance energy is proportional to (body weight)^0.75 (Nicol & Brookes, 2007). For example, if we take animal body maintenance energy as 0.5 × (body weight)^0.75 MJ day⁻¹, then 50 000 kg ha⁻¹ animal body weight comprised of 1250 head of 40 kg lambs has a metabolic energy demand of 9940 MJ day⁻¹, whereas the same animal body weight comprised of 100 head of 500 kg steers has a metabolic energy demand of 5290 MJ day⁻¹. Ideally, this higher energy demand for animals of smaller body weight should be factored into stocking rate calculations, in a way similar to that set out by Allen et al. (2011) for calculating animal units.

Now, to examine a selection of examples in the literature:

Example 1: Zhu et al. (2023) calculated sustainable stocking rate thresholds on the Qinghai–Tibet Plateau, based on animal numbers converted into stock unit (SU) equivalents in sheep and areas at the county level and reported the results in units of SU ha⁻¹ year⁻¹. Here, the authors standardized their animal units (sheep, yaks, etc) as sheep equivalents to take account of species differences in body weight and derived the annualized stocking rate as half the average stocking rate on separate summer and winter grazing areas. Both of these steps have a sound basis in logic. However, inclusion of “year⁻¹” in the units is incorrect as explained above. By analogy with acceleration (units m/s s⁻¹), including year⁻¹ would signify that stocking rate changes with the passage of time, which it clearly does not. Also, stocking rate would be unchanged regardless of the time base used to calculate it: six sheep·days per ha day, six sheep·months per ha month and six sheep·years per ha year all cancel to six sheep ha⁻¹.

Example 2: In an experiment in Maqu County on the Qinghai–Tibet Plateau, Wang et al. (2018) compared herbage parameters and animal weight gain of groups of eight lambs stocked either 1.0 or 0.5 ha paddocks for periods of 6 months in the warm season and on different paddocks for 6 months in the cold season. The authors describe their placement of 8 sheep on 0.5 or 1.0 ha for 6 months of the year as “stocking rates” of 16 or 8 sheep ha⁻¹, respectively, whereas if time in months is considered in both the numerator and the denominator, the stocking rates are half those cited. Thus, the reporting approach of authors in examples 1 and 2 differs. In the first example, the authors use an “annualized” stocking rate, whereas in the second example, the stocking rate during the part-year grazing period is presented. While the true situation can be discerned by reading the detail in the methods section of the respective papers, these two examples highlight that the current standard terminology is context-sensitive in its application, and this is a point for possible future attention.

Interestingly, from a whole-experiment perspective, the allocation of eight sheep to 0.5 ha for 6 months of the year and citation of the animal numbers during the experiment as 16 sheep per ha is closely similar to “stocking density” as defined by Allen et al. (2011). However, as the animals in this experiment were rotationally grazed on a third of each paddock every 10 days in summer and a half of each paddock every 15 days in winter, the instantaneous stocking densities for this treatment in this experiment in summer and winter are 48 and 32 sheep ha⁻¹, respectively.

Example 3: Welten et al. (2014) investigated the impact of the administration of dicyandiamide to dairy cows via drinking water on nitrogen losses from grazed pasture. In this experiment, pregnant Friesian dairy cows were rotationally grazed on 627 m² plots, with 24 h on each plot before moving to a new plot. Animals grazed for 12-day periods using 20 cows in June and 12 cows in August. The grazing intensity was reported as 319 and 191 cows/ha/day for June and August, respectively. While the numerical values for grazing intensity are correct (20 cows × 1 day/0.0627 ha; 12 cows × 1 day/0.0627 ha), the units are incorrect and should be cow·days ha⁻¹. In this case, by coincidence, because the grazing event is exactly 1-day duration, cow·days per ha = cows per ha per day. This misunderstanding over the correct units for grazing intensity is widespread among authors and can be in part traced back to a chapter in the text book “New Zealand Pasture and Crop Science” (Matthews et al., 1999), where the units for grazing intensity are also incorrectly set out as animals per ha per day. A numerical example may resolve the confusion. Consider the case of 400 sheep grazed on 2 ha, for either half a day or 2 days. There can be no dispute that the 2-day grazing event has four times the potential herbage removal of the half-day grazing event, and thus four times the grazing intensity, where grazing intensity is as defined above. If the units are calculated as animal·days ha⁻¹, we obtain 400 × ½/2 = 100 for the half-day grazing event and 400 × 2/2 = 400 sheep·days ha⁻¹ for the 2-day grazing event, with the relativity as expected. On the other hand, if we calculate animals per ha per day, we obtain values of 400 and 100 sheep·days ha⁻¹ for the half- and 2-day grazing events, respectively—clearly incorrect.