Quantitative and qualitative yield loss caused by red deer (Cervus elaphus L.) grazing on permanent organic grasslands

ABSTRACT

This paper addresses yield loss due to red deer (Cervus elaphus) grazing on permanent grasslands on three organic farms in SE Slovenia over two consecutive years. Deer graze on permanent grasslands throughout the year, but the resulting crop loss varies during the growing season. Accounting for three mowings, the quantitative yield loss due to deer grazing totalled 48–52%, or 3.9–4.3 t/ha. Deer had the highest consumption capacity in the spring (first mowing: yield loss of 1.7–1.9 t of air-dried hay/ha), and towards the end of the growing season, consumption decreased (third mowing: loss of 0.9 t/ha). A floristic inventory indicated low species diversity and production in permanent grasslands in SE Slovenia. The crude protein content in unprotected plots was generally greater than that in protected plots, but due to a large quantitative yield loss, the yield of crude protein and net energy for lactation on heavily grazed land were much lower than those on protected land in all mowings. The average additional cost per unit of heavily grazed area ranged from 182 to 344 EUR/ha, and the differences are the result of the different intensities of game grazing in individual locations and grassland production capacities.

KEYWORDS:

• Game
• Yield loss
• Chemical composition
• Floristic composition
• Animal husbandry

Introduction

In Slovenia, which is one of the most heavily forested countries in Europe, game animals are an important biotic factor for reducing the productivity of crops. The most economically harmful species are wild boar (Sus scrofa L.) (Laznik and Trdan 2014), which causes damage by trampling across meadows and eating and crushing corn, and red deer (Cervus elaphus L.) (Trdan and Vidrih 2008), which are especially harmful to grasslands because fodder for grazing makes up 50% of its food. Both species of game cause great problems for cattle and small cattle farmers in SE Slovenia, who have to buy large amounts of fodder due to the losses that these animals cause. The rehabilitation of land and its protection from game animals are expensive and, due to the large numbers of game in the area, no longer guarantee efficiency. Farmers can expect adequate crop loss compensation for wild boar damage to the grasslands they report, while compensation for crop losses due to deer grazing is almost impossible to obtain from relevant organisations.

According to previous results of research on the influence of deer grazing on grassland productivity in the Kočevsko region (SE Slovenia), two groups of researchers have worked in an area with heavy grazing by deer for at least 30 years (Trdan et al. 2000). They found a 34% decrease (from 0 to 80%) (Verbič et al. 2013) and a 40 to 59% decrease (Trdan et al. 2003) in dry matter yield. The first group of researchers included only the first mowing, while the second group included all 4 mowings (2002). The second group, who are the initiators of research related to the harmfulness of deer due to grazing on permanent grassland in Slovenia and beyond, conducted the first study (2002) on meadows with a conventional production method. Based on research conducted from 2013-2014, the results of which are presented in this paper, we evaluated the harmfulness of grazing on grasslands in organic production areas for deer (Figure 1).

Figure 1. Due to the large number of deer in the area of SE Slovenia where our research was carried out, red deer also graze on permanent grassland during the day (photo: Dušan Glušac).

In recent years, the harmfulness of deer – quantitative and/or qualitative losses of grassland –yield – on permanent grasslands has also been reported in other European countries and areas, such as Italy (Marchiori et al. 2012; Corgatelli et al. 2019), Spain (Carpio et al. 2021; Casals et al. 2021), Germany (Riesch et al. 2019), Great Britain (MacMillan 2004) and Finland (Virtanen et al. 2002), as well as on other continents, such as Asia (Kamei et al. 2010; Tsukada et al. 2013; Hata et al. 2019).

In the present research, which represents a continuation of and improvement on our research studying the influence of deer grazing on permanent grasslands in SE Slovenia, we used standard procedures to evaluate the direct and indirect losses in the quantity and quality of crops (volume of fodder) due to deer grazing on grasslands. We conducted our research on organic grasslands, which is novel because this type of agricultural land has not yet been the subject of studies on the damage caused by deer grazing. Before conducting our study, we speculated that red deer, even on permanent organic grasslands that are in poor condition, would affect quantitative and qualitative yields to an extent that is comparable to the loss on permanent grasslands that are better supplied with nutrients. Together with economic indicators at the farm level, we aimed to determine the increase in the cost of producing large volumes of meadow forage compared with conditions in which deer do not negatively affect the production of this type of fodder.

Materials and methods

Study area

The research took place in 2013 (year 1) and 2014 (year 2) on permanent grasslands at three locations with organic cattle farms in the Kočevsko region (SE Slovenia): Kačji Potok (45°34′41.06″N, 14°57′57.23″E, 527 m a.s.l.), Novi Lazi (45°34′18.7″N, 14°50′56.12″E, 546 m a.s.l.) and Stari Breg (45°41′7.12″N, 14°55′16.63″E, 527 m a.s.l.). In all locations, the research was carried out on land managed by the Farmland and Forest Fund of the Republic of Slovenia, whose tenants are the Kerneža (Novi Lazi), Senekovič (Kačji Potok) and Zemljič (Stari Breg) families.

The locations were selected because these cattle farmers have been facing a significant loss of permanent grassland (bulky fodder) due to deer grazing for more than 10 years. The land included in the experiment was used for grazing and mowing, and the tenants on average performed one to two mowings per year. The farms conducted no damage prevention management. The electric fences they use are solely for livestock grazing purposes. More data about the farms are available in Table 1.

Table 1. Characteristics of the studied farms in Novi Lazi, Kačji Potok and Stari Breg.

Characteristics of farms	Tenants of permanent grasslands
	Senekovič	Kerneža	Zemljič
Location	Kačji Potok	Novi Lazi	Stari Breg
Livestock type	Suckler cows	Suckler cows	Suckler cows
Number of animals	55	60	35
Agricultural land (ha)	78	93	34
Arable land (ha)	0.5	1	0.75
Grassland (ha)	77.5	92	33.25
Grassland use (%- cut:graze)	45:55	60:40	20:80
Fertilisation	organic	organic	organic

Assessment of red deer grazing

In year 1, we started the experiments on May 10, when we set up 15 iron cages made from construction mesh (diameter 5 mm) at each of the three locations (the locations are surrounded by forest on all sides, and the deer stay there throughout the entire calendar year). The external dimensions of the cages were 1 × 1 × 0.5 m, i.e. even larger than we used in our previous research (Trdan et al. 2003; Trdan and Vidrih 2008). The field experiments were conducted in randomised blocks. At each of the three locations, we set up 5 cages in each of the 3 blocks, which were separated from each other by more than 100 metres. We set the cages in an area 100 to 200 m from the forest edge, and the distance between the cages was 15 to 20 m. To prevent the deer from moving the cages, we attached (anchored) each of the cages on both shorter sides to the ground with an approximately 15 cm long iron wedge. In year 2, we started the same type of field experiment on May 9, when we set up 3 cages in Novi Lazi and 5 cages each in Kačji Potok and Stari Breg. The distance between the cages in all locations ranged from 10 to 20 m.

We did not change the positions of the cages at any of the three locations to sample the vegetable mass, and these plots (with an area of ⁣⁣1 m²) represented protected areas (‘protected’ or ‘P’ treatment) (Figure 2) or areas with optimal yield. In year 1, approximately two to three weeks before each of the three samplings (mowing), we set up 3 new cages in each of the 3 blocks per site, where the turf was heavily grazed (eaten) by deer. We called this ‘regrowth’ or ‘R’ treatment. With this treatment, we wanted to determine the ability of the grassland, which was intensively grazed by deer for up to 2-3 weeks before mowing, to regrow. The most direct loss of the crop (‘control’ or ‘C’) due to deer grazing on the grassland was determined on sampling days when we set up 3 cages in each of three blocks of the three locations in places that were grazed as intensively as possible.

Figure 2. The iron cage of the ‘protected’ treatment in Stari Breg during the third mowing, October 14, year 1 (photo: Stanislav Trdan).

In year 1, we carried out the first mowing in Novi Lazi on June 11, in Kačji Potok on June 12, and in Stari Breg on June 18. The second mowing was performed at all three locations on August 27, and the third mowing was performed at all three locations on October 14. In year 2, at all three locations, the first mowing was performed on June 6, the second on August 13, and the third on October 3. The dates of other research activities (setting up cages to determine the productivity and regrowth capacity [the latter was performed only in year 1] of the grazed grassland) are listed in Table 1. The dates of installation of the iron cages depended not only on the height and development of the plant species in the grassland but also on the activities of the farmers. Thus, we were able to install cages for the 2nd and 3rd mowings (events 2/3 and 3/3) only after the farmers performed the first or second mowing (Table 2).

Table 2. Timeline of tasks in the research to determine the impact of deer grazing on the productivity of permanent grassland at three locations in the Kočevska region in year 1 and year 2.

Location	Event	Year	Date of setting up the cages	Date of relocation of cages (regrowth)	Mowing date
Kačji Potok	1/3	1	May 10	May 27	June 12
Novi Lazi			May 10	May 27	June 11
Stari Breg			May 10	May 27	June 18
Kačji Potok	2/3		July 4	July 31	August 27
Novi Lazi			June 26	July 31	August 27
Stari Breg			July 4	July 31	August 27
Kačji Potok	3/3		August 27	September 19	October 14
Novi Lazi			August 27	September 19	October 14
Stari Breg			August 27	–	October 14
Kačji Potok	1/3	2	May 9	May 22	June 6
Novi Lazi			May 9	–	June 6
Stari Breg			May 9	–	June 6
Kačji Potok	2/3		July 12	July 1	August 13
Novi Lazi			June 18	–	August 13
Stari Breg			July 6	–	August 13
Kačji Potok	3/3		August 22	September 4	October 3
Novi Lazi			August 22	–	October 3
Stari Breg			August 13	–	October 3

When sampling, we first mowed the grass in the immediate vicinity of the cages with a mower (Honda UMK435 motor), and then we removed the cages and mowed the grass in the previously protected area (1 m²) with a sickle mower (type BCS 615 SL MAX). We transported the grass in labelled canvas bags to the Biotechnical Faculty, Department of Agronomy in Ljubljana, where we dried it in a drying room at 45 °C until constant weight was reached.

Determination of floristic composition

In year 2, we also determined the floristic composition of the grasslands (on June 4, August 12 and September 30) at all locations before each of the three mowings. We separately inventoried ungrazed and deer-grazed grasslands at each location. The size of the census plots was 5 × 5 steps. The names of the plants were taken from the Mala flora Slovenije (Martinčič et al. 2007). We used an atlas to determine the plant species (Rothmaler 1988), and we divided the listed plant species in the grassland into three groups: grasses, legumes and herbs.

The cover and abundance of species according to the Braun–Blanquet scale (+, 1, 2, 3, 4, 5) were converted into mean cover values ⁣⁣according to the table provided by Diersche (1994). From the mean cover values, we calculated the quality coefficients of grass litter and the corresponding fodder values of grass litter according to Sinkovič et al. (2009) and Šoštarić-Pisačić and Kovačević (1974).

Drying, weighing, grinding and chemical analysis of the samples

The samples were dried in a drying oven at 45 °C at the Biotechnical Faculty, Department of Agronomy in Ljubljana, to a constant weight. After drying, we weighed the air-dried matter from the samples from all the plots in the experiment. For further chemical analyses at the Veterinary Faculty, all air-dried samples were ground to a size of 1 mm (with a Brabender mill).

Additionally, this research analysed the chemical composition of the fodder; dry matter, moisture, crude protein (CP), crude fibre (CF), crude fat and crude ash contents were determined using Weende analysis. The metabolic energy (ME), net energy for lactation (NEL) (Universititaet Hohenheim 1997), crude protein yield and NEL yield were also calculated. The analysed and calculated parameters of chemical composition and energy value were classified into selected quality classes (Verbič et al. 2011). Eighty-one samples were included in the chemical analysis.

Economic analysis

To perform an economic analysis of the impact of game grazing on permanent meadows in the Kočevsko region, we used an evaluation method based on additional costs due to yield loss. The calculations assume that despite the loss of fodder for their animals, farmers do not adjust the volume of animal husbandry to the available fodder in the short term but instead maintain a constant stock and replace the lost fodder yield (deer and pasture) by purchasing hay. Therefore, we calculated the difference in the average amount of fodder produced without the influence of deer and that produced with the influence of deer in terms of hay equivalents, which were purchased to maintain the fodder balance. The resulting costs are estimated with the help of data from the Catalogue of calculations for management planning on Slovenian farms (Jerič 2011).

For the purposes of the economic analysis, we collected several types of information on each farm by interviewing the owner of the farm. In year 2, interviews with Mr. Kerneža and Mr. Senekovič were conducted on August 22, and interviews with Mr. Zemljič were performed on September 4. The data collected for further economic analysis included the types and stocks of animals on the farm, the types and stocks of animals sold annually, the categories and areas of ⁣⁣agricultural land use, the methods of use (mowing, grazing) and fertilisation, and the estimated productivity of the grasslands (Table 3). The estimated grassland productivity was represented by machinery use, including the use of different categories and of tractor and tractor attachments with varying power levels for the mechanical harvesting of fodder from grassland.

Table 3. Assessment of the production capacity (annual yield of hay in t/ha) of grasslands on selected farms in the Kočevska region by class.

Class	Location
	Kačji Potok	Novi Lazi	Stari Breg
5	7.5-7	8-9	7.5-6.5
4	6.5-7	7.5-8	6.5-5.5
3	6-6.5	7-7.5	4.5-5.5
2	5-6	6-7	3.5-4.5
1	4-5	5-6	2.5-3.5

Statistical analysis

The results of the experiment were statistically evaluated using the Statgraphics Centurion XVI programme (Statgraphics Centurion 2009). Differences between treatments (protected, regrowth, control) were evaluated by analysis of variance (ANOVA) and the Newman Keuls test for multiple comparisons (P < 0.05). The results are shown graphically as the average yield of air-dried samples (± SN), calculated in t/ha, in two (three when accounting for the regrowth results) treatments at three locations in the Kočevsko region and as the average yield of air-dried samples found in the same treatments after individual mowings. Based on these data, we calculated the average loss of grassland crops due to deer grazing in the Kočevsko region.

Results

Air-dried matter yield in protected and unprotected plots

The statistical analyses revealed that the average yield of dry matter in year 1 was significantly influenced by treatment (Df = 2, F = 106.83, P < 0.05), the number of cuttings (Df = 2, F = 439.88, P < 0.05), the location of the experiment (Df = 2, F = 7.73, P < 0.05), the interaction between treatment and the number of cuttings (Df = 4, F = 4.26, P < 0.05) and the interaction between the number of cuttings and the location of the experiment (Df = 4, F = 8.69, P < 0.05). Based on general statistics, the average yield of dry matter was significantly greater in the 1st cutting (4.06 ± 0.1 t/ha) than in the 3rd cutting (less than 800 kg/ha). The highest average yield of dry matter was also determined where grass was protected with cages (2.72 ± 0.17 t/ha) in comparison to the control treatment (unprotected area) (1.22 ± 0.17 t/ha) (Table 4).

Table 4. The average yield of dry matter (t/ha) in the 1st, 2nd, and 3rd rows at the three locations and in the three treatments in year 1. The different letters denote values that represent statistically significant differences (Newman–Keuls test) at the 0.05 significance level.

Location	1st cut			2nd cut			3rd cut
	Protected	Regrowth	Control	Protected	Regrowth	Control	Protected	Regrowth	Control
Kačji Potok	4.5 ± 0.4b	2.7 ± 0.3a	3.0 ± 0.6a	1.6 ± 0.3b	1.3 ± 0.4ab	0.7 ± 0.2a	1.6 ± 0.2b	0.7 ± 0.2a	0.4 ± 0.2a
Novi Lazi	5.4 ± 0.3b	3.8 ± 0.2a	3.8 ± 0.3a	3.1 ± 0.3b	0.4 ± 0.1a	1.6 ± 0.0a	0.8 ± 0.1b	0.4 ± 0.2a	0.2 ± 0.0a
Stari Breg	5.2 ± 0.3c	3.4 ± 0.5b	2.6 ± 0.2a	1.0 ± 0.2b	0.4 ± 0.1a	0.3 ± 0.1a	1.1 ± 0.2b	0.1 ± 0.1a	0.2 ± 0.1a

For comparison within the 1st mowing between different treatments within a specific location (Df = 2,32) (for location Novi Lazi, F = 23.83, P < 0.05; for location Kačji Potok, F = 9.00, P < 0.05; for location Stari Breg, F = 13.23, P < 0.005). For comparison within the 2nd mowing between different treatments within a specific location (Df = 2,32) (for location Novi Lazi, F = 47.82, P < 0.05; for location Kačji Potok, F = 3.35, P < 0.05; for location Stari Breg, F = 10.76, P < 0.005). For comparison within the 3rd mowing between different treatments within a specific location (Df = 2,32) (for location Novi Lazi, F = 12.64, P < 0.05; for location Kačji Potok, F = 22.47, P < 0.05; for location Stari Breg, F = 17.34, P < 0.005).

Based on the results of statistical analyses of the data collected in year 2, the average dry matter yield was significantly influenced by treatment (Df = 1, F = 101.16, P < 0.05), the number of cuttings (Df = 2, F = 154.25), the location of the experiment (Df = 2, F = 19.38, P < 0.05), the interaction between treatment and location (Df = 2, F = 11.64, P < 0.05) and the interaction between the number of cuttings and the location of the experiment (Df = 4, F = 12.19, P < 0.05). Additional values are presented in Table 5.

Table 5. The average yield of dry matter (t/ha) at the 1st, 2nd and 3rd mowings at the three locations and in two or three treatments in year 2. The different letters denote values that represent statistically significant differences (Newman–Keuls test) at the 0.05 significance level.

Location	1st cut		2nd cut		3rd cut
	Protected	Control	Protected	Control	Protected	Control
Kačji Potok	5.7 ± 0.4b	3.2 ± 0.4a	2.4 ± 0.2b	0.9 ± 0.2a	1.5 ± 0.2b	0.4 ± 0.1a
Novi Lazi	3.6 ± 0.2a	3.3 ± 0.2a	3.9 ± 0.6b	2.0 ± 0.4a	0.8 ± 0.2b	0.3 ± 0.0a
Stari Breg	3.7 ± 0.4b	1.2 ± 0.2a	1.7 ± 0.2b	0.3 ± 0.1a	1.3 ± 0.1b	0.2 ± 0.1a

For comparison within the 1st mowing between different treatments within a specific location (Df = 2,32) (for location Novi Lazi, F = 44.83, P < 0.05; for location Kačji Potok, F₉= 16.19, P < 0.05; for location Stari Breg, F = 34.34, P < 0.005). For comparison within the 2nd mowing between different treatments within a specific location (Df = 2,32) (for location Novi Lazi, F = 47.28, P < 0.05; for location Kačji Potok, F = 3.35, P < 0.05; for location Stari Breg, F = 10.76, P < 0.005). For comparison within the 3rd mowing between different treatments within a specific location (Df = 2,32) (for location Novi Lazi, F = 12.64, P < 0.05; for location Kačji Potok, F = 33.37, P < 0.05; for location Stari Breg, F = 17.34, P < 0.005).

Based on the average dry matter yield, which was obtained from the results for year 1 and year 2, we calculated the average percentage of yield loss per specific treatment. The protected treatment had no yield loss. The detailed values are presented in Table 6.

Table 6. Average loss of yield of dry matter (in %) in year 1 and year 2.

Location	Year	1st cut		2nd cut		3rd cut
		Regrowth	Control	Regrowth	Control	Regrowth	Control
Kačji Potok	1	40.00	33.3	18.8	56.3	56.3	75.0
Novi Lazi		29.6	29.6	87.1	48.4	50.0	75.0
Stari Breg		34.6	50.0	60.0	70.0	90.9	81.8
Average		34.7	37.6	55.3	58.2	65.7	77.3
Kačji Potok	2		43.9		62.5		73.3
Novi Lazi			8.3		48.7		62.5
Stari Breg			67.6		82.4		84.6
Average			39.9		64.5		73.5

Floristic composition of the grasslands

According to the typology of the habitat types of Slovenia (Jogan et al. 2004), the grasslands of the studied area belong to the Central European xeromesophilic lowland meadow habitat, occurring on relatively dry soils and sloping sites with a dominant tall fescue (Arrhenatherum elatius [L.] P. Beauv. Ex J. Presl & C. Presl) (Physis code 38.221), with relatively low biomass and up to two mowings annually. Due to the use of a large dataset, the results are presented only in paragraphs.

In 2023, in total, we identified 80 herbaceous plant species in all the census plots. Of these, 18 species were grasses, 2 species were grass-like plants, 5 species were legumes, and the remaining 55 species were herbs. The collected data showed that the largest number of species was observed in the plots at the Novi Lazi location (from 31 to 35), followed by the Stari Breg location (from 26 to 28); 15 to 19 species were observed at the Kačji Potok location.

The three most common species in the grass family were perennial ryegrass, orchard grass, and rough bluegrass. Red clover and white clover were almost always found among the legumes, and yarrow was found among the herbs in all censuses. The state of productivity of the grass areas at the Kačji Potok location was the most favourable of the three sites, followed by the Novi Lazi location and Stari Breg. We did not account for the incessant rooting of wild boars on these lands. The greater number of herb species that do not produce large crops can be attributed mainly to the long history of deer grazing and the unsatisfactory remediation of grass debris after rooting by wild boars.

During the first inventory in year 2, on June 4, the number of species varied from 15 (KP-ZP) to 25 (NL-ZP). Among the grasses, Kentucky bluegrass (Poa pratensis), Yorkshire fog (Holcus lanatus), meadow fescue (Festuca pratensis), and bulbous oat grass (Arrhenatherum elatius) dominated. White clover and red clover were dominant among the legumes, but they accounted for a smaller proportion of the legumes than the grasses. Among the herbs, yarrow was observed most often. According to the calculated quality coefficients, the listed grasslands were classified as being in medium-to-good or poor-to-medium condition. There were no significant differences between the ungrazed and grazed treatments.

During the second inventory, on August 12, the number of recorded species varied from 13 (Stari Bred – G) to 20 (Novi Lazi – G). Among the grasses, Kentucky bluegrass, perennial ryegrass, yellow oatgrass (Trisetum flavescens) and bulbous oatgrass dominate. White clover and red clover were dominant among the legumes, but they accounted for a smaller proportion of the legumes than the grasses.

Among the other legumes, common bird's-foot trefoil (Lotus corniculatus), meadow vetchling (Lathyrus pratensis), tufted vetch (Vicia cracca), and black medick (Medicago lupulina) were also present at the site. Among the herbs, yarrow was most frequently observed during the inventory (12), followed by wild carrot (Daucus carota) and creeping cinquefoil (Potentilla reptans). The calculated quality coefficients placed the grasslands in the classes of medium to good quality, no fodder value to poor, and poor to medium quality.

At the third inventory, on September 30, just before the final and third mowing, the absolute number of species was similar to that at the second inventory. The number of species varied from 10 (Kačji Potok – G) to 20 (Kačji Potok – U). Among the grasses, bulbous oatgrass, Kentucky bluegrass, perennial ryegrass, and yellow oatgrass dominated.

White clover (present at the Kačji Potok and Stari Breg locations) and red clover (present at all locations) were dominant among the legumes, but they still accounted for a smaller proportion of the legumes than the grasses. Of the other legumes that are worth mentioning, meadow vetchling was also present in the sward. Among the herbs, yarrow was identified in all plots (12), followed by ribwort plantain (Plantago lanceolata) and common self-heal (Prunella vulgaris). The calculated quality coefficients placed the grasslands in the highest classes, i.e. good to very good and very good to excellent fodder values.

Chemical composition and energy value of the studied grassland

The chemical composition and nutritional value of the sampled sward are given in Tables 6 and 7.

Table 7. Contents of crude protein (CP), crude fibre (CF), metabolizable energy (ME) and net energy for lactation (NEL) of hay in the three locations during all three mowings in year 2.

Location	Treatment	Moving	CP (g/kg air-dried matter)	CF (g/kg air-dried matter)	ME (MJ/kg air-dried matter)	NEL (MJ/kg air-dried matter)
Kačji potok	Protected	1st	95 ± 12a	284 ± 14a	8.12 ± 0.01b	4.63 ± 0.01b
	Unprotected	1st	99 ± 12a	323 ± 26b	8.01 ± 0.06a	4.56 ± 0.03a
	Protected	2nd	134 ± 12a	260 ± 19a	7.88 ± 0.48a	4.53 ± 0.27a
	Unprotected	2nd	153 ± 22a	304 ± 99a	8.37 ± 0.13a	4.81 ± 0.09a
	Protected	3rd	167 ± 9a	190 ± 16a	8.28 ± 0.05a	4.78 ± 0.03a
	Unprotected	3rd	180 ± 4a	206 ± 13a	8.30 ± 0.07a	4.78 ± 0.04a
Novi Lazi	Protected	1st	99 ± 2a	285 ± 2b	8.09 ± 0.03a	4.62 ± 0.02a
	Unprotected	1st	97 ± 5a	270 ± 8a	8.11 ± 0.04a	4.63 ± 0.02a
	Protected	2nd	120 ± 7a	282 ± 11a	8.31 ± 0.03a	4.78 ± 0.01a
	Unprotected	2nd	156 ± 5b	266 ± 14a	8.46 ± 0.02b	4.86 ± 0.00b
	Protected	3rd	161 ± 6a	235 ± 5b	8.56 ± 0.01b	4.92 ± 0.01b
	Unprotected	3rd	171 ± 10a	198 ± 8a	8.42 ± 0.06a	4.85 ± 0.03a
Stari Breg	Protected	1st	99 ± 8a	275 ± 12b	8.02 ± 0.04a	4.57 ± 0.02a
	Unprotected	1st	104 ± 2a	218 ± 22a	7.99 ± 0.08a	4.57 ± 0.05a
	Protected	2nd	159 ± 6a	236 ± 13a	8.45 ± 0.09a	4.86 ± 0.05a
	Unprotected	2nd	175 ± 8b	251 ± 20a	8.36 ± 0.05a	4.81 ± 0.03a
	Protected	3rd	156 ± 4a	190 ± 12a	8.38 ± 0.05a	4.83 ± 0.03a
	Unprotected	3rd	157 ± 9a	179 ± 6a	8.39 ± 0.05a	4.84 ± 0.03a

For Novi Lazi, for comparison within different cuts within different treatments (protected/nonprotected) (small case letter) (df = 2.6) (within 1st cut, for ME, F = 9.40, P = 0.0790; for NEL, F = 13.37, P = 0.0451; for CF, F = 41.33, P = 0.0471; for CP, F = 12.05, P = 0.0794; within 2nd cut, for ME, F = 0.97, P < 0.05; for NEL, F = 0.90, P = 0.3958; for CF, F = 88.19, P < 0.05; for CP, F = 0.60, P = 0.4832; within 3rd cut, for ME, F = 12.40, P < 0.05; for NEL, F = 45.10, P < 0.05; for CF, F = 33.12, P < 0.05; for CP, F = 45.59, P < 0.05).

For the Kačji potok, for comparison within different cuts within different treatments (protected/nonprotected) (small case letter) (df = 1.5) (within the 1st cut, for ME, F = 3.40, P = 0.1390; for NEL, F = 6.37, P = 0.0651; for CF, F = 1.83, P = 0.2471; for CP, F = 0.05, P = 0.8394; within the 2nd cut, for ME, F = 0.97, P = 0.3800; for NEL, F = 0.90, P = 0.3958; for CF, F = 0.19, P = 0.6860; for CP, F = 0.60, P = 0.4832; within the 3rd cut, for ME, F = 3.40, P = 0.1390; for NEL, F = 0.00, P = 1.000; for CF, F = 0.56, P = 0.4949; for CP, F = 1.59, P = 0.2753).

For Stari Breg, for comparison within different cuts between different treatments (protected/nonprotected) (small case letters) (df = 1,5) (within 1st cut, for ME, F = 0.13, P = 0.7397; for NEL, F = 0.01, P = 0.9201; for CF, F = 5.17, P = 0.0853; for CP, F = 0.38, P = 0.5711; within 2nd cut, for ME, F = 0.62, P = 0.4752; for NEL, F = 0.71, P = 0.4476; for CF, F = 0.37, P = 0.5754; for CP, F = 2.81, P = 0.1692; within 3rd cut, for ME, F = 0.01, P = 0.9355; for NEL, F = 0.02, P = 0.8908; for CF, F = 0.42, P = 0.5629; for CP, F = 0.15, P = 0.7266).

According to the statistical analysis for the Novi Lazi location, the ME parameter was influenced by treatment (protected/regrowth/unprotected) (Df = 4, F = 3.70, P < 0.005) and the number of cuttings (Df = 2, F = 11.89, P < 0.05). We also confirmed that the NEL was influenced by treatment (Df = 4, F = 3.78, P < 0.005) and the number of cuttings (Df = 2, F = 11.89, P < 0.05). The following factors also influenced the levels of the CP and CF parameters: CP: treatment (protected/regrowth/unprotected) (Df = 4, F = 4.56, P < 0.005); number of cuttings (Df = 2, F = 42.45, P < 0.05); CF: treatment (protected/regrowth/unprotected) (Df = 4, F = 1.85, P = 0.1465); and number of cuttings (Df = 2, F = 30.90, P < 0.05) (Table 7).

In Novi Lazi, we did not find any significant differences in crude protein content between the protected (P), regrowth (R) and unprotected or control (C) treatments only at the first mowing. We did not detect any significant differences in the contents of ME or NEL between the three treatments only at the first mowing, while at the second mowing, the highest values ⁣⁣of both parameters were found in the R treatment (cage setting: July 12). Significant differences between the treatments were also found at the third mowing, where the lowest values ⁣⁣of ME and MEL were confirmed on the unprotected parts of the land and on the plots where we studied the regrowth of sows with a cage setting of September 4 (Table 7).

According to our statistical analysis of the Kačji Potok location, ME was not influenced by the number of cuts (Df = 2, F = 1.17, P = 0.3400) or treatment (Df = 1, F = 0.54, P = 0.4763). Furthermore, there was no impact of the number of cuttings (Df = 2, F = 1.17, P = 0.5093) or treatment (Df = 1, F = 0.46, P = 0.5093) on NEL. However, we detected an impact of the number of cuts on the CF (Df = 2, F = 3.74, P = 0.0427) and CP (Df = 2, F = 19.77, P < 0.05) parameters. All additional values are presented in Table 7.

We confirmed a significant difference in the contents of ME and NEL between the two treatments only at the first mowing (in favour of the protection treatment); at the other two mowings, we did not find any significant differences between the treatments. Accounting for the quality classes of the harvests, we note that in the protected plots, we did not find significant differences in the NEL values ⁣⁣of the three mowings, while in the unprotected parts of the land, we found a significantly lower NEL content in the hay. Significantly greater ME contents ⁣⁣were found in the 2nd and 3rd mowings in the protected and unprotected areas of land (Table 7).

According to the results of the statistical analysis for the Stari Breg location, the parameters ME (Df = 2, F = 27.55, P < 0.05), NEL (Df = 2, F = 37.79, P < 0.05), CF (Df = 2, F = 7.19, P < 0.05) and CP (Df = 2, F = 58.08, P < 0.05) were all influenced by the number of cuttings (Table 8).

Table 8. Average yield of dry matter of hay (H), crude protein (CP), and net energy for lactation (NEL) at three locations during all three mowings in year 2.

Area	Treatment	Mowing	Yield of hey (t/ha)	Yield of CP (kg/ha)	Yield of NEL (GJ/ha)
Kačji potok	Protected	1st	5.7 ± 0.4b	544 ± 70b	26.39 ± 0.03b
	Unprotected	1st	3.2 ± 0.4a	317 ± 39a	14.59 ± 0.09a
	Protected	2nd	2.4 ± 0.2b	321 ± 28b	10.89 ± 0.65b
	Unprotected	2nd	0.9 ± 0.2a	138 ± 20a	4.32 ± 0.08a
	Protected	3rd	1.5 ± 0.2b	251 ± 13b	7.17 ± 0.04b
	Unprotected	3rd	0.4 ± 0.1a	71 ± 2a	1.92 ± 0.01a
Novi Lazi	Protected	1st	3.6 ± 0.2a	358 ± 6b	16.61 ± 0.08b
	Unprotected	1st	3.3 ± 0.2a	331 ± 11a	15.26 ± 0.07a
	Protected	2nd	3.9 ± 0.6b	467 ± 26b	18.63 ± 0.04b
	Unprotected	2nd	2.0 ± 0.4a	312 ± 10a	9.72 ± 0.02a
	Protected	3rd	0.8 ± 0.2b	129 ± 4b	3.94 ± 0.00b
	Unprotected	3rd	0.3 ± 0.0a	54 ± 0a	1.78 ± 0.33a
Stari Breg	Protected	1st	3.7 ± 0.4b	366 ± 31b	16.92 ± 0.08b
	Unprotected	1st	1.2 ± 0.2a	125 ± 2a	5.51 ± 0.09a
	Protected	2nd	1.7 ± 0.20b	270 ± 11b	8.25 ± 0.081b
	Unprotected	2nd	0.3 ± 0.1a	53 ± 2a	1.44 ± 0.00a
	Protected	3rd	1.3 ± 0.1b	203 ± 5b	6.28 ± 0.03b
	Unprotected	3rd	0.2 ± 0.1a	32 ± 3a	0.97 ± 0.04a

For Novi Lazi, for comparisons between specific cuttings between different treatments (for the 1st cutting, NEL, F_2,8= 585.98, P < 0.05; CP, F_2,8= 11.40, P < 0.05; for the 2nd cutting, NEL, F_4,14= 595.26, P < 0.05; CP, F_4,14= 5.30, P < 0.05; for the 3rd cutting, NEL, F_3,11= 58.36, P < 0.05; CP, F3,11 = 166.81, P < 0.05)

For the Kačji potok, for comparison between specific cuttings and different treatments (small case letters) (for the 1st cutting, NEL, F_1,5= 14943.87, P < 0.05; CP, F_1,5= 8.02, P < 0.05; for the 2nd cutting, NEL, F_1,5= 99.36, P < 0.05; CP, F_1,5= 28.86, P < 0.05; for the 3rd cutting, NEL, F_1,5= 15388.13, P < 0.05; and CP, F_1,5= 173.47, P < 0.05).

For Stari Breg, for comparisons within specific cuttings between different treatments (small case letters) (for the 1st cutting, NEL, F_1,5= 8309.07, P < 0.05; CP, F_1,5= 61.10, P < 0.05; for the 2nd cutting, NEL, F_1,5= 307.36, P < 0.05; CP, F_1,5= 392.87, P < 0.05; for the 3rd cutting, NEL, F_1,5= 10701.42, P < 0.05; CP, F_1,5= 277.11, P < 0.05).

In Stari Breg, we found significant differences in crude protein content between the protected and unprotected treatments only during the second mowing and in favour of the crop in the unprotected area. Considering this forage quality parameter, the yield of all three cuttings in both treatments was excellent. We did not find any significant differences in the contents of ME or NEL between the two treatments in any of the cuttings. Accounting for the quality classes of the harvests, we note that in the protected and unprotected plots, significantly lower values ⁣⁣of NEL and ME were found in the yield of the first mowing, while we did not find any differences in the yields of the 2nd and 3rd mowings in either treatment (Table 7).

According to the results of the statistical analysis, the CP yield in the Novi Lazi location was significantly influenced by treatment (protected/regrowth/unprotected) (Df = 4, F = 7.60, P < 0.05) and the number of cuts (Df = 2, F = 129.68, P < 0.05). Furthermore, the NEL yield was also significantly influenced by treatment (protected/regrowth/unprotected) (Df = 4, F = 9.56, P < 0.05) and the number of cuts (Df = 2, F = 161.93, P < 0.05) (Table 8).

At the first mowing in Novi Lazi, we found a significantly greater dry matter yield in the protected and unprotected (N) plots. The crude protein yield exhibited a similar pattern, as a significantly lower crude protein production rate was observed in the R treatment (296.7 kg/ha). A significantly greater yield of NEL (16.61 GJ/ha) was observed in the protection treatment than in the other two treatments.

Based on the statistical analysis, the CP and NEL parameters at the Kačji Potok location were both influenced by the number of cuttings and treatment (protected/unprotected). The CP yield was affected by the number of cuttings (Df = 2, F = 30.29, P < 0.05) and treatment (Df = 1, F = 44.73, P < 0.05), while the NEL yield was affected by the number of cuttings (Df = 2, F = 1936.47, P < 0.05) and treatment (Df = 2, F = 1257.26, P < 0.05) (Table 8).

At the Kačji Potok location, in the control treatment, the yields of dry matter, crude protein and NELa in each of the three cuttings were significantly lower than those in the protected treatment. The crude protein yield in both treatments decreased with each subsequent mowing, as did the NELa yield. A significantly greater crude protein yield (543 kg/ha) was found in the 1st mowing in the protected treatment than in the other treatments, and a significantly lower yield (71.92 kg/ha) was found in the 3rd mowing in the control treatment (Table 8). Comparing the yields of crude protein and NELa between the Novi Lazi and Kačji Potok locations, at the second location, both parameters reached higher values in the 1st and 3rd mowings but lower values in the 2nd mowing.

According to the statistical analysis, both the CP yield and NEL yield for Stari Breg were influenced by the number of cuttings and treatment (protected/unprotected). The CP yield was affected by the number of cuttings (Df = 2, F = 39.32, P < 0.05) and treatment (Df = 1, F = 19391.48, P < 0.05), and the NEL yield was affected by the number of cuttings (Df = 2, F = 7152.85, P < 0.05) and treatment (Df = 1, F = 19391.48, P < 0.05) (Table 8).

As in Kačji Potok, in Stari Breg, in the control treatment, the yields of dry matter, crude protein and NELa were significantly lower for each of the three cuttings than in the protected treatment. Only in the 1st mowing in the protected treatment did we find a dry matter yield that exceeded 3 t/ha, while in all the other treatments, the dry matter yield was less than 1.8 t/ha. Additionally, the yield of crude protein exceeded 300 kg/ha only in the protected treatment during the 1st mowing, after which the yield of crude protein in both treatments decreased with each of the remaining two mowing treatments. Only during the 1st mowing did the NELa yield exceed 15 GJ/ha, and this was the case only in the protected treatment (Table 8). At the Kačji Potok location, the yields in the protected treatment were significantly greater than those in the control treatment at Stari Breg.

Economic analysis

On the basis of model cost estimates, we found that the cost to the farms for forage from grass due to the damage caused by deer (assuming that the farms raised livestock to the extent that the forage produced would allow) was greater due to both the higher costs of production and the purchase of the necessary forage, which allowed them to maintain the desired stock of animals. The results are presented in Table 9, which shows that the average additional cost per unit area varied between €182/ha and €344/ha. The differences are the result of differences in the intensities of game grazing at individual locations and differences in the production capacities of the grasslands.

Table 9. The effect of game grazing on the cost of forage production on grassland.

Location	Area (ha)	Yield withoutthe influence ofgame (t)		Yield withgameinfluence (t)		Difference (t)		Necessary purchase of hayto level the balance (t)	Additionalproduction cost	Average additionalcost/ha
		Hay	Pasture	Hay	Pasture	Hay	Pasture
Kačji potok	88.09	547	111	257	52	290	59	295	30.275,01 €	343,69 €
Novi Lazi	100.08	391	1768	262	1185	129	583	177	18.197,60 €	181,84 €
Stari Breg	87.61	284	641	71	160	213	481	253	25.961,17 €	296,34 €
Total	275.77	1222	2520	590	1397	632	1123	724	74.433,79 €	273,96 €

Thus, the average damage per unit area was highest at the Kačji Potok location. The average yield loss was not the highest in this location (48.5%), but this area had the greatest potential for the most intensive cultivation, with the highest average yields per hectare; therefore, the average damage was also the highest.

Discussion

We performed quantitative and qualitative analyses of the productivity of permanent grasslands at three locations in the Kočevsko region (SE Slovenia), an economic evaluation of grassland production in areas where deer represent an important biotic factor in yield failure and other accompanying research activities in the study area from 2013–2014. After accounting for the results of all three mowings at all three locations on the protected plots, the results of our analysis demonstrated that there was a significantly greater total yield of air-dried hay in Novi Lazi, with 9.3 in year 1 and 8.3 t/ha in year 2, followed by Kačji Potok, with 7.7 and 9.6 t/ha, and Stari Breg, with 7.2 and 6.7 t/ha. Disregarding the results of the 2nd mowing (due to drought), in year 1, the average dryland yields at the indicated locations were 6.2, 6.1 and 4.5 t/ha. These results demonstrate that in Kočevsko in year 1 and year 2, there was a loss of yield in permanent grasslands due to deer grazing, which has thus far been the subject of only a few studies on European and global scales (Virtanen et al. 2002; MacMillan 2004; Marchiori et al. 2012; Tsukada et al. 2013; Casals et al. 2021).

Considering the results of all three mowings, in Novi Lazi, we found 40 (in year 1) and 33% (in year 2) yield failures (or losses of total weight of 3.7 and 2.7 t of air-dried hay/ha); in Kačji Potok, 47 and 53% (3.6 and 5.1 t/ha); and in Stari Breg, 56 and 75% (4.0 and 5.0 t/ha). Considering only the first and third mowings, 35% (2.2 t/ha), 44% (3.6 t/ha) and 71% (4.5 t/ha) of the plants experienced crop loss from air-dried hay in year 1, which proves that the yield losses in the control (unprotected) plots, expressed as a percentage of the yield in the protected plots, increased from the first to the last mowing. This finding was also confirmed in the research of Marchiori et al. (2012), who reported a 15-20% yield loss in the pre-Alpine region of NE Italy in the first cut and a 25-40% loss of yield in the second and last cuts. Interestingly, in a related study in northern Italy, yield loss was confirmed for only the second mowing, with a loss of 14% due to red deer grazing (Corgatelli et al. 2019). This proves that the loss of grassland crops in the distribution area of ⁣⁣red deer depends on many factors. In general, high densities of deer alter forest vegetation, orchards and agricultural field, which can cause severe financial loss (Richer et al. 2005).

Considering the results of all three mowings at all three locations, we note that the average optimal total yield (in the protected treatment) of air-dried grassland in the Kočevsko region was 8.1 (in year 1) or 8.2 t/ha (in year 2), and the total yield loss due to deer grazing (unprotected treatment) amounted to 48% (3.9 t/ha) or 52% (4.3 t/ha). Considering the results of the first and third mowings, the average optimal total yield was 6.2 t/ha, and the total yield loss was 44% in year 1 (2.7 t/ha). Thus, we can claim that deer grazing on permanent grasslands has a significant effect on yield, as already reported by Marchiori et al. (2012).

We found that deer graze on the permanent grassland in the Kočevsko region throughout the entire calendar year, but the resulting quantitative loss of yield varies during the growing season; deer have the highest consumption capacity in spring (at the first mowing, we found yield losses of 1.9 (in year 1) and 1.7 t (in year 2) of air-dried hay/ha), and in summer (at the second mowing in year 2, we found a yield loss of 1.6 t of air-dried hay/ha); however, towards the end of the growing season, consumption decreases (at the third mowing in both years, we found a loss of 0.9 t/ha). Due to the more intensive growth of plants in permanent meadows in the spring, the yield of the first mowing was 38 or 40%, that of the second mowing was 55 or 59%, and that of the third mowing in both years was as much as 75%.

The floristic inventories of the permanent grassland at all three locations immediately before each of the three mowings did not significantly differ among the grassland groups (grasses, legumes, and herbs). From 5 to 7 species of grasses, 2 to 5 species of legumes and 7 to 14 species of herbs were identified in the rush hour. The percentages of cover and abundance of these taxa were highly variable, but these results indicate the low species diversity and productivity of permanent grasslands on organic cattle farms in the Kočevsko region. During the inventories, characteristic patches were observed on the ground (in relatively small areas) where the sward was thriving in a dwarf (lawn) form. This was the result of several years of selective early spring grazing by deer. By browsing or selective grazing, red deer cause the development and growth of the sward to trend in a negative direction (the spread of tall stems, low bushes and woody species) in such a way that the plants in the sward adapt to altered growth habits. Findings, that grazing by red deer influences on plant succession by Schutz et al. (2003). However, since farmers also mow such land later, this development pattern does not lead to secondary succession or overgrowth. Interestingly, Riesch et al. (2020) noted that grazing by wild red deer can benefit vegetation structure and diversity and could therefore enrich the set of tools available for the conservation of seminatural open vegetation types. Additionally, Virtanen et al. (2002) and Schutz et al. (2003) stated that red deer grazing sustains the plant species diversity of productive grasslands. On the other hand, Gass and Binkley (2021) reported that intensive grazing is correlated with drier, more compacted soils and nutrient loss, suggesting increasing N limitation and reduced plant biodiversity.

The differences in optimal grassland productivity between locations with otherwise comparable soil pH values ⁣⁣and nutrient supplies (phosphorus, potassium, and organic matter) are attributed to the intensive spring fertilisation of the meadows in Novi Lazi with livestock manure. It is known that intensively managed grasslands are located in the lowland areas and are characterised by high input of fertilisers (Bilotta et al. 2003). The greatest crop loss was observed in Stari Breg because of the already changing floristic composition of the meadows (with larger proportions of low grasses, white clover and creeping herbs), in which the deer grazed with a distinct preference. We must not forget the fact that deer grazing that occurs too frequently destroys plants when they are still in the vegetative stage of development because deer graze in the same areas very often and tear the leaves off the same plants repeatedly; thus, plant root systems cannot fully develop, causing limited growth in summer when there is a lack of precipitation. Schutz et al. (2003) reported that red deer prefer phosphorus-rich, formerly irrigated areas for grazing, while Riesch et al. (2022) reported that the annual nutrient export of nitrogen (N) and phosphorus (P) by red deer grazing was greater than the nutrient import through red deer excreta, resulting in an average net nutrient removal of 14 and 30 kg N/ha and 1.1 and 3.3 kg P/ha in heathlands and grasslands, respectively.

Introducing controlled grazing by domestic animals on grasslands provides the opportunity to control the duration of idle time after each grazing rotation, and the selectivity of grazing animals is also lower. At the Stari Breg location, productivity may have also been lower because the deer only fed on the grassland and spent the night in the forest; therefore, they left a large part of their excrement (including minerals, nitrogen and undigested organic matter from consumed vegetables) in the forest. This leads to the transfer of nutrients from agricultural land to forestland, which leads to the impoverishment of grasslands in terms of soil fertility. Additionally, Catorci et al. (2016) stated that deer grazing in a grassland where controlled livestock grazing has been introduced has many negative impacts, while it is said to have positive impacts on abandoned (uncultivated) grassland.

Chemical analysis of the sampled vegetation showed that the crude protein content in the unprotected treatment was always greater than that in the regrowth and protected treatments, as published also by Jargue-Bascuñana et al. (2022). This is the result of deer grazing, which, by successive scraping and removing vegetation, rejuvenates the grass and consequently forces the grass to form new leaves, which are the main carriers of this quality indicator. This fact has previously been confirmed in research on domestic grazing animals (Clark et al. 2000). The opposite pattern was observed for crude fibre, which had the highest content in the hay in the protection treatment, and this pattern was generally observed at all three locations. The nutritional value of the hay produced at all locations was poor, even at the 1st mowing, as it did not exceed 5 MJ/kg of dry matter; this finding is mainly attributed to the poor floristic composition of the grassland. Thus, deer grazing also significantly affects the chemical composition of grasslands (Marchiori et al. 2012). Additionally, Corgatelli et al. (2019) concluded that the chemical composition of meadow forages differed only slightly between grazed and ungrazed plots; however, the crude protein and neutral detergent fibre contents were significantly lower in grazed plots.

From the point of view of the later growth of the grassland and the provision of fodder for grazing in the summer period, ‘overgrazing’ by any (domestic or free-living) animal under optimal growing conditions is more desirable than is the herd grazing system; however, such a system also requires a temporally (on an annual scale) appropriate beginning and, of course, conclusion (withdrawal of animals from such lands) of grazing. However, this can be achieved only by fencing deer-exposed agricultural land (Goddard et al. 2001; Baltzinger et al. 2018) or by moving these animals between areas.

In the economic analysis, which is the first of its kind in the field of damage caused by deer on permanent grassland, the model estimates showed that the cost of forage purchased by the farms was greater due to the damage caused by deer, assuming that they raised livestock to the extent that the produced forage would allow. The additional cost was due to both the higher costs of production and the purchase of the necessary fodder, which enabled farms to maintain the desired animal stock. We note that the average additional cost per unit of area varies between €182/ha (Novi Lazi) and €344/ha (Kačji Potok). The differences are the result of differences in the intensities of game grazing at individual locations and differences in the production capacities of the grasslands. Thus, the average damage per unit area was highest at the Kačji Potok location; the average yield loss was not the highest at this site, but this area allows for the most intensive cultivation with the highest average yields per hectare; therefore, the average damage was also the highest. The economic analysis indicated that to assess the appropriate compensation for damage caused by deer grazing on permanent grassland, it is necessary to consider not only the actual crop loss but also the production potential of the areas in question.

In 2000 (Trdan et al. 2000), we were the first to draw attention to the problem of deer grazing on permanent grasslands in Slovenia and beyond, and although we did our best to bring this problem to the attention of competent institutions, nothing has changed in this field to date. Specifically, farmers are still not entitled to compensation for damage caused by deer grazing, lessening their economic gains and leading to the abandonment of farming. Based on many years of research and knowledge of the advantages and disadvantages of farming in the Kočevsko region, we conclude that it is important that the permanent meadows in this area, on which deer graze beginning in early spring, are also mowed later. When these lands are left unmown, they will become overgrown very quickly (Ribeiro et al. 2020). In the future, we would like to study the possibility of reducing the loss of permanent grassland crops due to deer grazing in the Kočevsko region.

Availability of data

The datasets analysed during the current study are available from the corresponding author upon reasonable request.

Author contributions

ST and MV conceptualised the study, carried out field data collection, wrote the first draft of the manuscript and wrote the final version of the manuscript, ŽL and TS contributed to field data collection, BJS, GTK and AU assisted with data analysis, and TB assisted with data analysis and interpretation of the results.

Acknowledgements

We thank the Zemljič family from Stari Breg, the Senekovič family from Kačji Potok and the Kerneža family from Novi Lazi for giving us access to the land they use for the preparation of bulky fodder for livestock and for the production of iron cages, which we used in the research. Jaka Rupnik and Boštjan Medved Karničar are acknowledged for their technical assistance.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research was carried out within the framework of the Programs for the Conservation and Development of Agriculture and the Countryside of the Municipality of Kočevje in 2013 and 2014.

Notes on contributors

Stanislav Trdan

Stanislav Trdan is a full professor–researcher in the Department of Agronomy at the Biotechnical Faculty, University of Ljubljana (Slovenia), where he has worked since 1997. His research interests are focused on environmentally acceptable plant protection, especially in the field of agricultural entomology and zoology. He has published more than 120 refereed articles.

Žiga Laznik

Žiga Laznik has been a researcher in the Department of Agronomy at the Biotechnical Faculty, University of Ljubljana (Slovenia), since 2007. His main research area is biological control, with special emphasis on interactions between plant pests, their plant hosts and biological control agents.

Tomaž Sinkovič

Tomaž Sinkovič is a retired researcher. Until 2021, he worked in the Department of Agronomy at the Biotechnical Faculty, University of Ljubljana (Slovenia). His main research area was botany.

Breda Jakovac Strajn

Breda Jakovac Strajn is an associate professor–researcher at the Veterinary Faculty, University of Ljubljana (Slovenia). Her main research area is animal feed quality and hygiene.

Gabrijela Tavčar Kalcher

Gabrijela Tavčar Kalcher is a researcher at the Institute of Food Safety, Feed and Environment, at the Veterinary Faculty, University of Ljubljana (Slovenia). Her main research interests include animal nutritional feed and analytical chemistry.

Andrej Udovč

Andrej Udovč is a full professor–researcher in the Department of Agronomy at the Biotechnical Faculty, University of Ljubljana (Slovenia). His main research area is agricultural economics.

Tanja Bohinc

Tanja Bohinc has been a researcher in the Department of Agronomy at the Biotechnical Faculty, University of Ljubljana (Slovenia), since 2012. Her research is oriented towards environmentally acceptable plant protection and interactions among plant pests/diseases and the chemical structure of their hosts.

Matej Vidrih

Matej Vidrih is an assistant professor–researcher in the Department of Agronomy at the Biotechnical Faculty, University of Ljubljana (Slovenia), where he has worked since 1999. His research is oriented towards grassland management and field crop production.