Skip to main content

Data from: Knowledge co-production with traditional herders on cattle grazing behaviour for better management of species-rich grasslands

Cite this dataset

Molnar, Zsolt et al. (2020). Data from: Knowledge co-production with traditional herders on cattle grazing behaviour for better management of species-rich grasslands [Dataset]. Dryad.


The research gap between rangeland/livestock science and conservation biology/vegetation ecology has led to a lack of evidence needed for grazing-related conservation management. Connecting scientific understanding with traditional ecological knowledge of local livestock keepers could help bridge this research and knowledge gap. 1. We studied the grazing behaviour (plant selection and avoidance) of beef cattle (ca. 33 000 bites) on species-rich lowland pastures in Central Europe and traditional herding practices. We also did >450 outdoor interviews with traditional herders about livestock behaviour, herders’ decisions to modify grazing behaviour, and effects of modified grazing on pasture vegetation. 2. We found that cattle grazing on species-rich pastures displayed at least 10 different behavioural elements as they encountered 117 forage species from highly desired to rejected. The small discrimination error suggests that cattle recognize all listed plants ‘by species’. 3. We also found that herders had broad knowledge of grazing desire and they consciously aimed to modify desire by slowing, stopping or redirecting the herd. Modifications were aimed at increasing grazing intensity in less desired patches and decreasing grazing selectivity in heterogenous swards. 4. Synthesis and applications: These traditional herd management practices have significant conservation benefits, such as avoiding under- and overgrazing, and targeted removal of pasture weeds, litter and enchroaching bushes, tall competitive plants and invasive species. We argue that knowledge co-production with traditional herders who belong to another knowledge system could help connect isolated scientific disciplines especially if ecologists and rangeland scientists work closely with traditional herders, co-designing research projects and working together in data collection, analysis and interpretation. Stronger links between these disciplines could help develop evidence-based, specific conservation management practices while herders could contribute with their practical experiences and with real world testing of new management techniques.04-May-2020


We observed ingestive bite selection and avoidance behaviour of cattle at and just above the feeding station level in May, June, August, October 2014 and May, June, August, September, October 2015, altogether during 22 days, both during morning and afternoon meals (4-5 hours each) on three nearby pastures. Livestock were calm, individuals were selected randomly. We went as close as possible (0.8-2 m) keeping impact on cattle to the minimum while having a good view of plants near each animal’s mouth. When plants were difficult to identify to the species level, we visited the feeding station just after the animal departed. We documented how and how often certain plant species were approached, eaten or avoided by cattle (see behavioural elements in Table 1). To avoid differences among observers, most observations were done by the first author who had a full knowledge of the local flora and was experienced using binoculars (Nikon, 8*40 mm).

We grouped bite-level observations (ca. 33 000 bites) into frequency categories as our goal was to document behaviour towards as many plant species as possible instead of quantifying daily intake or forage preference (cf. Agreil et al., 2005). We did not study grazing preference (i.e. choice given alternatives, comparing intake and available biomass for each species separately) which would have been impossible in such species-rich pastures. Frequency category 3 (see Appendix S3) means more than 6 (usually 20<) observations per meal, 2 means 3-6 observations per meal, and 1 means 1-2 observations per meal (category boundaries were set arbitrarily). We analysed 117 species (from a total of 241) for which we had observations on at least 3 independent days in spring and autumn, respectively. We developed a desirability index using the proportion of intake-positive elements (the first 5 in Table 1) divided by the total frequency of all 10 elements.

Usage notes



National Research, Development and Innovation Office, Award: K 119478

GINOP, Award: 2.3.2-15-2016-00019

GINOP, Award: 2.3.2-15-2016-00019