**2. Materials and methods**

#### **2.1 Study area**

*Habitats of the World - Biodiversity and Threats*

compete for the many different services that grassland habitats provide [8–10]. Although the value of grasslands, from a socioeconomic perspective and for the environmental services provided, is widely recognized, their degradation process is continuous and global [1, 11–13]. Grassland degradation generally implies a negative reduction in biodiversity, vegetation coverage, plant height, and biomass production [14–16]. Also, the deterioration of ecosystem services and functions was also included in this definition [1, 17]. The degradation process generates a series of ecological problems—loss of biodiversity, carbon sink, and water storage capacity—as well as the intensification of soil degradation and dust storms [3, 6, 18]. Worldwide, up to 50% of grasslands are affected by degradation, mainly due to human activities and climate change [12, 13, 19]. Several studies reveal that land-use changes are responsible for up to 66% of the grassland degradation, whereas the climate dynamics account for approximately 20% [13, 14, 19, 20]. At a European level, climate is the primary degradation agent in some areas of Northern and Northwestern Europe and the southern part of European Russia, but in most areas, including Eastern Europe, degradation is mainly caused by land-use issues [13, 21]. Sudden changes in land-use intensity such as overgrazing or abandonment of traditional farming practices are among the main factors identified to cause the degradation of grassland habitats [13, 22–25]. The alteration of agricultural practices (intensification or abandonment), along with the area of degraded grasslands and the associated environmental problems, shows an upward trend [26–28]. The most important policies aiming to manage and mitigate these issues that have been developed within the European Union (EU) are the Common

Agricultural Policy (CAP) framework and the Environmental Directives (especially Habitats Directive 1992/43/ECC-HD and Birds Directive 2009/147/EC-BD). For grassland habitats, these policies mainly focus on agricultural production (livestock density) and, respectively, on biodiversity conservation. The CAP directives include livestock density determination according to the grassland carrying capacity, while the Habitats Directive (1992/43/ECC) implements the conservation of the habitats of community importance which were selected according to their structure (floristic diversity) and environmental ecological functions. The favorable conservation status of these habitats must be reached or maintained within all the sites which are included in the European Natura 2000 Network (N2000). The network includes a large number of protected areas (27863 sites), being acknowledged as one of the world's most effective legal instruments for biodiversity and nature conservation, with an important function in conserving Europe's natural capital. It is estimated that approximately 16% of the habitats in N2000 areas depend on a perpetuation of extensive farming practices and especially on maintaining the extensive management of grasslands [29]. In the EU-27, approximately 18% of the permanent grasslands are within the protected N2000 Network [30]. However, the effects of grazing livestock density (stocking rate) on the grassland habitats protected within N2000 sites have rarely been considered so far, particularly the context of their actual spatial overlap. Moreover, for some countries hosting very large areas of permanent grasslands in the EU (e.g., Romania), spatially detailed data at the landscape level are not yet available, although the agricultural statistics are reported at the national level by each member state. For instance, the spatial distribution of livestock in Europe was modeled using statistical downscaling of province-level livestock statistics [31], but the possible deviations from the grassland habitats' optimal livestock

This paper aims to evaluate the degradation status of grassland habitats by modeling and mapping the grazing livestock density and the subsequent deviations from the optimal grassland carrying capacity within and outside the N2000 sites from Romania. The permanent grassland habitats from Romania are among the

**30**

density (carrying capacity) are not yet assessed.

Romania (area 238397 km<sup>2</sup> ; capital Bucharest 44°25′57″N, 26°06′14″E) is located in Southeastern Europe, bordering on the Black Sea and the Danube, with the Southeastern Carpathian Mountains in its center (**Figure 1**). The natural landscape includes almost even proportions of mountains (31%), plains (33%), and hills (36%) that expand rather uniformly from the mountains, reaching elevations above 2500 m, to the Danube Delta, a few meters above sea level. The climate is transitional between temperate and continental. The average annual temperature goes from 11°C in the south to 8°C in the north. Annual precipitation decreases eastward and downward, averaging up to 1010 mm in some mountainous areas, 635 mm in the Transylvanian Plateau, 521 mm in Moldavia, and only 381 mm in Dobruja and close to the Black Sea.

The Corine Land Cover dataset [33] reports for Romania the following land cover classes: artificial areas (5.34%), arable land and permanent crops (39.37%), pastures mosaics (17.65%), forested land (31.68%), seminatural vegetation (2.78%), open space and barren soils (0.10%), wetlands (1.35%), and water bodies (1.69%). The General Agricultural Census in Romania, performed in 2010, indicates that the permanent grasslands cover 44940 km<sup>2</sup> , including both grazed pastures and hay meadows, that together make up about 33% of total utilized agricultural land [32, 34]. The greatest surface covered by permanent grasslands is in the Carpathian Mountains region and in the Transylvanian Plateau, where every county has between 1000 and 3500 km<sup>2</sup> of grassland.

The geomorphological and climatic diversity of Romania, the geographical position at the intersection of several floristic provinces, and the extensive traditional

**Figure 1.**

*The geographical position (a) and the elevation model (b) of Romania.*

land use all contribute to the vegetation diversity [35, 36], reflected also in the large variety of grassland habitat types [4, 37, 38]. Most herbaceous vegetation types (except ruderal) are comprised in 15 N2000 grassland habitat types [39, 40].

The Romanian grassland habitats are diverse, including dry grasslands, mesophilous grasslands, high-mountain grasslands, and wet grasslands. The detailed description of the floristic structure specific to these vegetation types can be found in phytosociological studies [41] and the Romanian grassland inventory [42]. According to the latter source that mapped an area of 3900 km2 , the best-represented habitat types were mesophilous (39.1%, mostly *Arrhenatheretalia* vegetation order) and dry grasslands (38.2%, mostly *Festucetalia valesiacae*), followed by highmountain grasslands (12.7%, *Nardetalia*, *Caricetalia curvulae* etc.), wet grasslands (5.35%, mostly from *Molinietalia*), and ruderal-degraded grasslands (4.2%).

## **2.2 Data and spatial modeling**

The spatial distribution of the deviations from the optimum livestock density (DEVOLD) was modeled in GIS in order to quantify and map its effect on the

**33**

**Table 1.**

*Deviation from Grazing Optimum in the Grassland Habitats of Romania Within...*

grassland habitat degradation status throughout the grassland habitats from Romania. The data presented in **Table 1** were geoprocessed in ModelBuilder, and a GIS toolbox was developed for analyzing the DEVOLD (grazing carrying capacity). All the GIS processing and spatial analysis were performed using ArcGIS 10.5 [43].

This study encompassed all the permanent grassland polygons (GP)

longer [32, 34, 44]. Also, the data regarding the area (40451.91 km2

Romania [44]. The permanent grassland area is the land used to grow grasses or other herbaceous forage that is not subject to crop rotation for at least 5 years or

the 435 Romanian Sites of Community Importance (ROSCI) was included in the model. The dataset with the numbers and types of livestock from each TAU was downloaded from the National Statistics Institute of Romania [34]. The total number of the different livestock types (animal heads) was recorded during the General Agricultural Census from 2010 for 3177 TAUs from 41 counties. Only the following types of grazing livestock were included in the analysis: cattle (dairy and beef), sheep, goats, horses, donkeys, and mules. Livestock numbers were converted into livestock units (also called animal units) using specific coefficients indicated in the official Romanian guidelines [45]. The livestock unit (LU) is a reference unit which facilitates the aggregation of livestock from various species and ages. One LU is the grazing equivalent of one adult dairy cow producing 3000 kg of milk annually, without additional concentrated foodstuff. According to the transformation coefficients from the national regulations [47–49], the formula and coefficients used for the conversion of the animal numbers and types in number of LUs (for each TAU) is:

 LU number = Cattle Number + (Sheep Number × 0.15) + (Goats Number × 0.15) + Horses Number + (Donkeys & Mules Number × 0.4) (1)

http://geoportal.ancpi.ro/geoportal

Census 2010 [34]

http://www.mmediu.ro/articol/date-gis/434

Digital elevation model over Europe (EU-DEM) https://www.eea.europa.eu/data-and-maps/data/ copernicus-land-monitoring-service-eu-dem#tab-gis-data

National Statistics Institute of Romania—The General Agricultural

The official regulations in Romania regarding the grazing management plans and grassland experts [47–49]

Southeastern European countries [21, 50–54]

*The input data for spatial modeling of the deviations from the optimal livestock density within and outside the* 

The official regulations in Romania regarding the methodology for the evaluation of the optimal livestock density per hectare [45, 46]

The upper limit/level for the optimal livestock density recommended by various studies for biodiversity conservation in Central and

Land parcel identification system from Romania [44]

) which are included in the Land Parcel Identification System from

) covered by all

*DOI: http://dx.doi.org/10.5772/intechopen.85734*

**Input data Data source**

Permanent grassland polygons

The polygons of N2000 Sites of Community Importance (ROSCI)

Digital elevation model (DEM

Livestock numbers and types

Optimal livestock density for socioeconomic production

Optimal livestock density for biodiversity conservation

*N2000 sites from Romania.*

The polygons of territorial administrative units (TAU)

(GP)

25 m)

(LU)

(OLD\_E)

(OLD\_B)

from TAUs

Coefficients for the transformation of different animal types in livestock units

(33529.42 km2

*Deviation from Grazing Optimum in the Grassland Habitats of Romania Within... DOI: http://dx.doi.org/10.5772/intechopen.85734*

grassland habitat degradation status throughout the grassland habitats from Romania. The data presented in **Table 1** were geoprocessed in ModelBuilder, and a GIS toolbox was developed for analyzing the DEVOLD (grazing carrying capacity). All the GIS processing and spatial analysis were performed using ArcGIS 10.5 [43].

This study encompassed all the permanent grassland polygons (GP) (33529.42 km2 ) which are included in the Land Parcel Identification System from Romania [44]. The permanent grassland area is the land used to grow grasses or other herbaceous forage that is not subject to crop rotation for at least 5 years or longer [32, 34, 44]. Also, the data regarding the area (40451.91 km2 ) covered by all the 435 Romanian Sites of Community Importance (ROSCI) was included in the model. The dataset with the numbers and types of livestock from each TAU was downloaded from the National Statistics Institute of Romania [34]. The total number of the different livestock types (animal heads) was recorded during the General Agricultural Census from 2010 for 3177 TAUs from 41 counties. Only the following types of grazing livestock were included in the analysis: cattle (dairy and beef), sheep, goats, horses, donkeys, and mules. Livestock numbers were converted into livestock units (also called animal units) using specific coefficients indicated in the official Romanian guidelines [45]. The livestock unit (LU) is a reference unit which facilitates the aggregation of livestock from various species and ages. One LU is the grazing equivalent of one adult dairy cow producing 3000 kg of milk annually, without additional concentrated foodstuff. According to the transformation coefficients from the national regulations [47–49], the formula and coefficients used for the conversion of the animal numbers and types in number of LUs (for each TAU) is:

 LU number = Cattle Number + (Sheep Number × 0.15) + (Goats Number × 0.15) + Horses Number + (Donkeys & Mules Number × 0.4) (1)


#### **Table 1.**

*The input data for spatial modeling of the deviations from the optimal livestock density within and outside the N2000 sites from Romania.*

*Habitats of the World - Biodiversity and Threats*

land use all contribute to the vegetation diversity [35, 36], reflected also in the large variety of grassland habitat types [4, 37, 38]. Most herbaceous vegetation types (except ruderal) are comprised in 15 N2000 grassland habitat types [39, 40].

The Romanian grassland habitats are diverse, including dry grasslands, mesoph-

sented habitat types were mesophilous (39.1%, mostly *Arrhenatheretalia* vegetation order) and dry grasslands (38.2%, mostly *Festucetalia valesiacae*), followed by highmountain grasslands (12.7%, *Nardetalia*, *Caricetalia curvulae* etc.), wet grasslands (5.35%, mostly from *Molinietalia*), and ruderal-degraded grasslands (4.2%).

The spatial distribution of the deviations from the optimum livestock density

(DEVOLD) was modeled in GIS in order to quantify and map its effect on the

, the best-repre-

ilous grasslands, high-mountain grasslands, and wet grasslands. The detailed description of the floristic structure specific to these vegetation types can be found in phytosociological studies [41] and the Romanian grassland inventory [42].

According to the latter source that mapped an area of 3900 km2

*The geographical position (a) and the elevation model (b) of Romania.*

**32**

**Figure 1.**

**2.2 Data and spatial modeling**

The total livestock density (LD) measures the stock of animals, expressed in LU, per hectare of permanent grasslands. LD was calculated considering the total number of LUs from each TAU divided by the total area of permanent grassland of the respective TAU. Since there are no available data regarding the spatial distribution of the LD within each TAU from Romania, the LD (LU/ha) was calculated for the permanent grassland area of each TAU. Also, this approach is supported by the fact that a single grazing management plan is designed for all grassland parcels at the TAU level [45].

The difference between the current LD of a grassland and the optimum livestock density for the respective area and conditions represents the deviation from OLD. The equation for generating the deviation from optimum livestock density (DEVOLD) in each grassland polygon is:

$$\text{DEV}\_{\text{OLD}} = \frac{\text{LD} \times 100}{\text{OLD}} - 100 \tag{2}$$

LD is the livestock density as livestock units/hectare (LU/ha); OLD is the optimum livestock density for the grassland polygon.

The areas where the deviation from the DEVOLD is at most plus or minus 10% are considered not impacted. The fragments of grassland habitats where the DEVOLD is between 10.1 and 50% (plus or minus) are considered partially impacted. Those where DEVOLD is over 50% (plus or minus) are considered to be subject to a major impact, and degraded because of inappropriate livestock densities [48, 49]. Impact and degradation can be caused both by overgrazing and abandonment. In the first case, the grass cover decreases and allows the expansion of ruderal species that are good competitors but have a low forage value, or, worse, the soil is stripped of vegetation, favoring erosion. In the second case, the abandonment of grassland usage (as pasture or hayfield) is also harmful, resulting in shrub invasion which decreases grassland biodiversity and finally in the establishment of forest habitats.

Two scenarios were considered for the grassland habitats included in the N2000 ROSCIs. The first one, which was applied for all the grassland habitats of Romania, employs an optimum livestock density considered suitable for the grassland habitat areas with predominant socioeconomic purpose (OLD\_E). OLD\_E was synthesized from the Guidelines for Elaborating the Grazing Management Plans [47] that takes into account the different ecological and production characteristics of the various grassland habitat types. As a consequence, for each of the three main altitude belts of Romania, a specific OLD (LU/ha) was assigned, as follows: 0.46 (20–200 m a.s.l.), 0.6 (201–800 m a.s.l.), and 0.9 (801–2544 m a.s.l.).

The second scenario analyzes the prospect of using a lower OLD, favoring biodiversity conservation (OLD\_B), for the grasslands situated within N2000 ROSCIs, where lower intensity grazing is recommended [21, 50–54]; in the case of OLD\_B, the value of 0.45 LU/ha [21] was employed, although the large range of elevations and ecological conditions from the territory of Romania might require more specific values for different grassland types and altitude belts.

#### **3. Results and discussions**

The assessment and mapping of the deviation from the grazing carrying capacity were carried out within an area of 33529.42 km2 that corresponds to the permanent grasslands from Romania. Our results indicate that 17.34% (5814.75 km2 ) of these grasslands are situated within the N2000 ROSCIs (**Figure 2**). This indicates an important overlap between domestic livestock husbandry and nature conservation

**35**

**Figure 2.**

*Deviation from Grazing Optimum in the Grassland Habitats of Romania Within...*

same values and status for all the grasslands within a TAU.

within ROSCIs, both supported by the EU within the rural, regional, and environmental development policies. The grazing livestock types (cattle, sheep, goats, horses, donkeys, and mules) from each TAU, presented in **Figure 2**, depend on these grassland habitats, which are the most important resource for livestock production systems [34]. It is estimated that permanent grasslands provide at least 60% of the forage necessary for cattle and 80% for sheep [49]. The livestock density is considered to be one of the most relevant indicators of grassland degradation status, being strictly connected to both the socioeconomic factors and the ecological carrying capacity of grassland habitats. Overstocking permanent grasslands as well as understocking them until abandonment impacts them and, at high intensities, causes their degradation. However, the time span between successive grazing events may also be very important besides LD and grazing intensity [31, 55], but it is very difficult to quantify and map at large scale for each individual grassland polygon. Modeling the livestock data reported at the TAU level for evaluating the LD and DEVOLD is the best alternative for the available data, although it has the limitation of assigning the

*Livestock distribution in the territorial administrative units from Romania (a). The permanent grassland* 

*habitats and the limits of the N2000 Sites of Community Importance from Romania (b).*

*DOI: http://dx.doi.org/10.5772/intechopen.85734*

#### *Deviation from Grazing Optimum in the Grassland Habitats of Romania Within... DOI: http://dx.doi.org/10.5772/intechopen.85734*

within ROSCIs, both supported by the EU within the rural, regional, and environmental development policies. The grazing livestock types (cattle, sheep, goats, horses, donkeys, and mules) from each TAU, presented in **Figure 2**, depend on these grassland habitats, which are the most important resource for livestock production systems [34]. It is estimated that permanent grasslands provide at least 60% of the forage necessary for cattle and 80% for sheep [49]. The livestock density is considered to be one of the most relevant indicators of grassland degradation status, being strictly connected to both the socioeconomic factors and the ecological carrying capacity of grassland habitats. Overstocking permanent grasslands as well as understocking them until abandonment impacts them and, at high intensities, causes their degradation.

However, the time span between successive grazing events may also be very important besides LD and grazing intensity [31, 55], but it is very difficult to quantify and map at large scale for each individual grassland polygon. Modeling the livestock data reported at the TAU level for evaluating the LD and DEVOLD is the best alternative for the available data, although it has the limitation of assigning the same values and status for all the grasslands within a TAU.

**Figure 2.**

*Livestock distribution in the territorial administrative units from Romania (a). The permanent grassland habitats and the limits of the N2000 Sites of Community Importance from Romania (b).*

*Habitats of the World - Biodiversity and Threats*

(DEVOLD) in each grassland polygon is:

DEVOLD = \_\_\_\_\_\_\_\_

optimum livestock density for the grassland polygon.

the TAU level [45].

The total livestock density (LD) measures the stock of animals, expressed in LU, per hectare of permanent grasslands. LD was calculated considering the total number of LUs from each TAU divided by the total area of permanent grassland of the respective TAU. Since there are no available data regarding the spatial distribution of the LD within each TAU from Romania, the LD (LU/ha) was calculated for the permanent grassland area of each TAU. Also, this approach is supported by the fact that a single grazing management plan is designed for all grassland parcels at

The difference between the current LD of a grassland and the optimum livestock density for the respective area and conditions represents the deviation from OLD. The equation for generating the deviation from optimum livestock density

LD is the livestock density as livestock units/hectare (LU/ha); OLD is the

grassland biodiversity and finally in the establishment of forest habitats.

a.s.l.), 0.6 (201–800 m a.s.l.), and 0.9 (801–2544 m a.s.l.).

values for different grassland types and altitude belts.

were carried out within an area of 33529.42 km2

**3. Results and discussions**

LD × 100

The areas where the deviation from the DEVOLD is at most plus or minus 10% are considered not impacted. The fragments of grassland habitats where the DEVOLD is between 10.1 and 50% (plus or minus) are considered partially impacted. Those where DEVOLD is over 50% (plus or minus) are considered to be subject to a major impact, and degraded because of inappropriate livestock densities [48, 49]. Impact and degradation can be caused both by overgrazing and abandonment. In the first case, the grass cover decreases and allows the expansion of ruderal species that are good competitors but have a low forage value, or, worse, the soil is stripped of vegetation, favoring erosion. In the second case, the abandonment of grassland usage (as pasture or hayfield) is also harmful, resulting in shrub invasion which decreases

Two scenarios were considered for the grassland habitats included in the N2000 ROSCIs. The first one, which was applied for all the grassland habitats of Romania, employs an optimum livestock density considered suitable for the grassland habitat areas with predominant socioeconomic purpose (OLD\_E). OLD\_E was synthesized from the Guidelines for Elaborating the Grazing Management Plans [47] that takes into account the different ecological and production characteristics of the various grassland habitat types. As a consequence, for each of the three main altitude belts of Romania, a specific OLD (LU/ha) was assigned, as follows: 0.46 (20–200 m

The second scenario analyzes the prospect of using a lower OLD, favoring biodiversity conservation (OLD\_B), for the grasslands situated within N2000 ROSCIs, where lower intensity grazing is recommended [21, 50–54]; in the case of OLD\_B, the value of 0.45 LU/ha [21] was employed, although the large range of elevations and ecological conditions from the territory of Romania might require more specific

The assessment and mapping of the deviation from the grazing carrying capacity

grasslands from Romania. Our results indicate that 17.34% (5814.75 km2

grasslands are situated within the N2000 ROSCIs (**Figure 2**). This indicates an important overlap between domestic livestock husbandry and nature conservation

that corresponds to the permanent

) of these

OLD <sup>−</sup> <sup>100</sup> (2)

**34**

The geoprocessing steps that were performed and integrated into ModelBuilder in order to identify the status of each grassland polygon regarding the DEVOLD are presented in **Figure 3**. In the first stage, all the input data were processed at the national level. The current LUs and subsequently the LD (LUs/ha), OLD, and DEVOLD were derived for each grassland polygon based on the OLD\_E values recommended for each of the three altitude belts from Romania (**Figure 3a**). In the second stage, the resulting grasslands–DEVOLD\_E dataset was intersected with the limits of ROSCIs in order to analyze the status of the grassland habitats included within these protected areas (**Figure 3b**). Subsequently, in the third stage, the OLD\_B value was input into the model as an alternative to OLD\_E for the grassland habitats included in ROSCIs (**Figure 3c**). The developed GIS toolbox with the OLD model is flexible, allowing to easily test a different OLD or to be adapted for any similar case study. As mentioned above, the results obtained from the model are an approximation that considers the LD as having a uniform distribution throughout all the grassland habitats from each TAU. Although the situation within individual grassland parcels might be different, on average it is accurate at the TAU level, particularly taking into account the spatial and temporal dynamics of grazing, the high probability

#### **Figure 3.**

*The models generated for the analysis of the (a) grassland habitats at the national level, using deviations from the optimal livestock for socioeconomic production (OLD\_E); (b) grassland habitats from the N2000 SCIs, using deviations from the optimal livestock for socioeconomic production (OLD\_E); and (c) grassland habitats from the N2000 SCIs, using deviations from the optimal livestock for biodiversity conservation (OLD\_B).*

**37**

**Figure 4.**

*the national level.*

*Deviation from Grazing Optimum in the Grassland Habitats of Romania Within...*

**density for sustainable economic production (OLD\_E)**

tend to increase with altitude following the available plant biomass. For the analyzed grassland polygons of Romania (33529.42 km2

of livestock grazing within the TAU of their owners, and the grazing management

The assembly of the input data within the OLD model and the used spatial analysis tools are presented below for the grassland habitats situated within and

**3.1 Case scenario 1: status of the grassland habitats with the optimal livestock** 

), the deviation

This scenario considers the values of the optimal livestock density (carrying capacity) recommended by the Romanian grassland experts for sustainable economic production of biomass [47–49]. Most grazing management studies and textbooks recommend different optimal LDs (stocking rates), but they generally

of the existing livestock density from the OLD\_E (**Figure 4**) results in 52.45% of the grassland area being degraded (major impact, current LD with more than 50% over or under OLD\_E). The LD was much higher than the carrying capacity (overgrazing impact) for 44.05% of the area, with 8.40% of the area being impacted by abandonment, the LD being far under the OLD. Of the 39.25% grassland habitat area that is partially impacted (10.1–50% over or under the CC), 23.94% has an LD under the optimal value, while 15.31% is moderately overgrazed. At the national level, only 8.28%

*The spatial distribution of impact and degradation at the national level caused by the deviations from grazing optimum for the socioeconomic production scenario; the deviation classes, status, their percentage, and area at* 

*DOI: http://dx.doi.org/10.5772/intechopen.85734*

plans which are designed at the TAU level.

outside protected areas (**Figure 3**).

#### *Deviation from Grazing Optimum in the Grassland Habitats of Romania Within... DOI: http://dx.doi.org/10.5772/intechopen.85734*

of livestock grazing within the TAU of their owners, and the grazing management plans which are designed at the TAU level.

The assembly of the input data within the OLD model and the used spatial analysis tools are presented below for the grassland habitats situated within and outside protected areas (**Figure 3**).
