• Reduce text

    Reduce text
  • Restore text size

    Restore text size
  • Increase the text

    Increase the text
  • Print


Grazing behaviour in cattle and horses

When grazing in natural grassland pastures, cows and horses do not graze randomly. They select certain plants or certain grazing sites in the pasture. This leads to heterogeneous sward structures that may enhance pasture biodiversity.

Grazing: diversified environments where the grass consumed always depends on the same factors: management (stocking rate, grazing time), type of pasture (species, stage, season), animal traits (age, weight, production), other feed distributed (forage, concentrates, minerals). Charolais cattle, Laqueuille - Auvergne.. © INRA, AGABRIEL Jacques
By Louise Bergès, translated by Daniel McKinnon
Updated on 10/02/2013
Published on 08/30/2013

Researchers from the Herbivore Research Unit spent two years collecting sward measurements in pastures of the French Massif Central. They have spent considerable time studying the ways animals graze in order to analyse the extent to which selective grazing may create varied habitats in pastures.

Grazing creates patches in pastures

Researchers demonstrated that cattle, as well as horses, favour certain grazing sites within a pasture. They return regularly to graze the same sites over the course of a grazing season and from one season to the next. Animals prefer high-quality regrowth over low-quality reproductive stems. This preference creates patches where the vegetation remains short. Patches are grazed more often,  leading to stable short areas within pastures.

Patch stability is a function of grassland fertility and stocking rates

The study revealed a correlation between sward fertility (1) and the size and stability of grazed patches. Patches in fertile pastures measure several square metres, whereas low fertility pastures have smaller, mottled  patches measuring approximately ten square centimetres. The stability of these smaller patches is strongly tied to the presence of tussocks of grass species that animals avoid, such as Avenula spp or Nardus stricta (matgrass).

In the case of cattle, the development over time of stable short patches depends on the number of animals present on the pasture. When stocking rates (2) are low, grazing areas appear at the start of the season and are maintained until the autumn. When the number of animals per hectare increases, a larger portion of the pasture is grazed and stable vegetation patterns only appear towards the end of the grazing season with persistent tall sward surface around species that cattle avoid such as nettles and thistles.

In the case of horses, patch stability is high, irrespective of stocking rates.

Patches play a role in biodiversity

In turn, cattle and horse behaviour has an effect on vegetation communities. Large patches in fertile pastures lead to contrasted habitats within pastures, and thus to the coexistence of different plant communities. On less fertile pastures, heterogeneity patterns appear on a finer scale. These smaller patches create germination microsites that may serve as a refuge for certain plants which may have difficulty establishing themselves otherwise.

“Cattle and horse grazing creates habitats that benefit biodiversity” says Bertrand Dumont, head of the Animal–Plant Interactions and Forage team. “Knowing if there is an ‘optimal’ level of patch stability for high- or low-fertility pastures would allow us to offer advice on ‘good grazing practices’. A number of practices have already been tested by our team, for example removing animals from some pastures at flowering peak to maintain contrasting sward height and increase flowering intensity. The aim is to develop ‘win–win’ strategies, practices that would allow pasture biodiversity to be preserved without a negative impact to farm production levels. Understanding the beneficial effects of these practices to biodiversity would allow us to calculate the ecosystem services provided by grassland-based systems.”

(1) Fertility: annual plant biomass production (measured in grams of dry matter per metre squared).
(2) The stocking rate is an expression of the number of animals per unit of surface area. It is often measured in kilograms of live-weight per hectare, or in livestock units (LU)  per hectare.  It is an indicator of the intensity of livestock production.

Scientific contact(s):

Associated Division(s):
Animal Physiology and Livestock Systems , Social Sciences, Agriculture and Food, Rural Development and Environment.
Associated Centre(s):
Auvergne Rhône-Alpes

Measurements and modelling

Researchers selected six natural grassland pasture sites in oceanic, Mediterranean, and upland climate zones. Site fertility varied among (and sometimes within) sites; there was a large gradient in annual biomass production, from 130 g of dry matter / m2 to 800 g of dry matter / m2.

Within each pasture, the researchers chose fixed sites where they measured sward height – the study’s core data – over a grazing season, and analysed the local composition of the vegetation cover. Measurements were made twice over the course of the grazing season, once in July at biomass peak, and two and a half months later in the autumn. Measurements were repeated the following year at the same locations.

Measurements were made at two spatial scales: a fine-scale grid with an area of 1 m2 divided in 10 cm2 squares, and a large-scale 500 m2 grid. The two scales allowed differences in patch size and stability to be observed.

Starting with the premise that grass height indicates the likelihood that grass was grazed, the researchers developed a Boolean model to calculate the probability of grid points being grazed or not. Spatial covariogram statistics then measured the stability of grazed areas over time.

The model thus mapped grazed and ungrazed areas at the two spatial scales. Comparisons among the six sites are straightforward as no threshold level was defined.

Visual representation of patches according to pasture fertility. © INRA
Visual representation of patches according to pasture fertility © INRA