Large carnivores have made a remarkable comeback in Europe during the last half century, and recovery is still ongoing in large parts of the continent (Andrén 2018; Boitani 2018; Boitani et al. 2022; Chapron et al. 2014; Huber 2018; Ranc 2018; von Arx 2020). While this expansion can be celebrated as a huge conservation success, it also creates considerable challenges for coexistence in the multi-use landscapes of Europe (Linnell 2013).

Having a common understanding of the distribution, size and trends of large carnivore populations in Europe is one prerequisite for a knowledge-based dialogue in the often heated and highly politicized discussions about future scenarios of large carnivore conservation and management in Europe. Because of the scale at which large carnivores utilize the landscape it is essential to conduct periodic continental scale assessments of their status transcending sub-national and national borders. Assessments at this scale require harmonising diverse datasets that arise from different jurisdictions using different monitoring approaches.

These shapefiles provides the best available overview of brown bear (Ursus arctos), Eurasian lynx (Lynx lynx), wolf (Canis lupus), golden jackal (Canis aureus), and wolverine (Gulo gulo) distributions a continental scale.

2.1. Distribution mapping methods

The mapping approach generally follows the methods described in (Chapron et al. 2014) and (Kaczensky et al. 2013). It updates the published Species Online Layers 2012-2016 for brown bear, Eurasian lynx, wolf, golden jackal, and wolverine (Kaczensky et al. 2021; Ranc et al. 2022) for the period 2017-2022/23. Large carnivore presence was mapped at a 10 x 10 km (ETRS89-LAEA Europe) grid scale. This grid is widely used for Habitat Directive reporting to the European Union (EU) and can be downloaded at: http://www.eea.europa.eu/data-and-maps/data/eea-reference-grids-2. The map encompasses the continental EU countries plus Switzerland and Norway, and the EU candidate / potential candidate countries in the Balkan region, in addition to Ukraine and Turkey. For the two latter countries, only parts were included; for Ukraine only the Carpathian region (for this report Ukraine was artificially cut off and the straight line in the east does not represent the national border), and the European part of Turkey. For the 2012-2016 mapping, several countries were not or not fully (not for all species) included (Hungary, Montenegro, Turkey), so that no comparisons can be made of the updated carnivore distributions with those from the last mapping for these countries. Mapping large carnivores for this report had a two-fold goal: 

• Visualizing areas of large carnivore presence 
• Visualizing the variation in the underlying data quality

2.1.1. Presence status

We aimed to distinguish between two presence levels:

• Permanent = suggesting an established population which is reproducing, but also including cells with continuous presence in the absence of documented reproduction.
• Sporadic = suggesting only occasional presences of dispersers or lone individuals. 
• Where this distinction was not possible, but presence was confirmed, we used Undefined = presence confirmed, but not known if it is permanent or sporadic.

“Permanent” is equivalent to the status of “Present regularly” (PRE) as used in Article 17 reporting to the Habitats Directive, while “Sporadic” corresponds to the status of “Occasional” (OCC) in the same system. It was not possible to systematically separate out the “Newly arrived” (ARR) category, although it may well apply to the many new jackal records throughout western and northern Europe. Finding a common harmonized definition that fits all monitoring circumstances is difficult and the distinction required expert assessment. Here are the most common scenarios that we have continued with from the previous 2012-2016 mapping cycles:

1) For countries where the known annual species distribution was monitored annually, the distinction between permanent and sporadic was primarily made based on how consistently the species was detected in a cell over the 5–7-year monitoring period:

• Permanent = presence confirmed in ≥ 3 years in the last 5 - 7 years OR reproduction confirmed at least once within the last 3 years
• Sporadic (highly fluctuating presence) = presence confirmed in <3 years in the last 5 years OR in <50% of the time

2) For countries where the probability of species presence is modelled based on presence signs in combination with habitat parameters and distance rules, the distinction between permanent and sporadic can be made based on the modelled “probability of presence” value of a cell. As models used for different populations will vary in their approach, the cut-off values for permanent, sporadic, and absent were defined by the national/population level species experts.

3) For countries where the total range is covered by rotating annual surveys of parts of the total area over a 5–7-year period (i.e. different sections are surveyed in different years such that the whole area is surveyed at least once during the cycle), other criteria need to be used such as: comparison to presence in a cell (or adjacent cells) during the previous survey cycle, or confirmed reproduction, or the presence of females, to delineate permanent presence from sporadic presence. However, where monitoring is too fragmented and infrequent so that no reasonable distinction between permanent and sporadic can be made, the category “undefined” was used.

In general, telemetry data of long-distance dispersers out of the known range and once off documentation of individuals outside the known range were categorised as sporadic presence.

2.1.2. Large carnivore signs used to map presence

Large carnivore signs as a basis for mapping

We used the following presence categories, which were derived from the SCALP criteria for lynx in the Alps (Molinari-Jobin et al. 2012), but supplemented with two additional data quality information categories:

1. Confirmed presence signs

• Category 1 (C1): “Hard facts”, verified and unchallenged large carnivore presence signs (e.g. dead animals, DNA, verified camera trap images);
• Category 2 (C2): “Confirmed signs”, large carnivore presence signs controlled and confirmed by a large carnivore expert (e.g. trained member of the network), which requires documentation of large carnivore signs (e.g. tracks in the snow).
• We also had to include the category C1*, referring to C1 records which also include observations by trained or experienced personnel; it was not always possible to know if these observations also included non-documented records such as direct observations (which by the SCALP definitions do not quality as confirmed records). This is particularly true for the documentation of bear family groups or where monitoring was heavily based on hunters, foresters, and protected area wardens (e.g., Croatia, Slovakia, Ukraine).

2. Extrapolated confirmed presence signs

• Category “buffered”: Confirmed presence signs with a buffer around them, ideally based on well-documented/ published methods, and usually done to represent home ranges that are typically larger than 10 x 10 km in many parts of Europe (especially in the north).
• Category “modelled”: confirmed presence signs and modelling based on habitat suitability and/or proximity criteria ideally based on well-documented/published methods and explicit cut-off values.

For areas of poor monitoring coverage or infrequent monitoring, we also included:

3. Unconfirmed presence signs

• Category 3 (C3): Unconfirmed reports of category 2 large carnivore presence signs and all presence signs such as sightings and calls which, if not additionally documented, cannot be verified.
• Category “Soft”: Extrapolation of large carnivore presence based on interviews questionnaires, and media coverage from 2017-2022/23,
• Category “Past presence”: Documented presence from the past (but no older than from 2010) and no indication that the situation has changed.

From signs to grid cell

Ideally, GPS locations of large carnivore signs were intersected with the 10 x 10 km ETRS89-LAEA Europe grid. However, for some countries, data was collected at the spatial scale of hunting grounds (e.g., bear observations by hunters in eastern and south-eastern Europe). In this case the hunting grounds with large carnivore presence were intersected with the 10 x 10 km grid. Large carnivore presence was assumed for all cells intersecting the hunting ground (normally using a minimum intersection area in the range of >10%).

2.1.3. Data quality of large carnivore presence cells

We aimed to present data quality information at the grid cell level, but for some datasets it was not feasible to do so, and data quality was provided at the scale of the entire, or parts of, the layer, and not the individual grid cells. For some datasets this resulted in a mix of cells based on confirmed and buffered, modelled or unconfirmed cells without spatially explicit information at the grid cell level. Consequently, the following final data quality categories were used:

• Confirmed presence: based on C1 & C2 signs, cases where C1 included unspecified “observations” were marked with a star (C1*)
• Extrapolated presence: cells which don’t have LC signs but are intersected by buffers or have a high probability of large carnivore presence based on documented modelling approaches.
• Unconfirmed presence: cells with only C3 signs, or cells with only data from prior to 2017, where presence is still assumed to persist.

Where data was only available at the shape file level, we also used the following categories:

• Confirmed and extrapolated presence: mixed layer with buffered C1 & C2 signs and/or documented modelling approach, or when data was only available at the level of hunting grounds.
• Confirmed and unconfirmed presence: a mixed layer with C1-C3 signs; for these datasets it can be assumed that the majority are C1 and C2, but that documentation is not (readily) available - these data sources include hunter observations and some damage inspection data.

2.1.4. Populations

Grid cells were assigned to populations based on the LCIE’s population approach (Linnell et al. 2008). Population delineation follows a combination of ecosystem boundaries, topography, different management regimes, distributional discontinuities and administrative units, selected to create practical and functional management units. Population borders were partly drawn out of convenience, roughly following topographic regions, natural or artificial barriers such as large rivers, and in some cases national borders to facilitate reporting. Ecological conditions and monitoring methods tend to be similar within populations. Populations are primarily based on where animals are detected, not where they have originated from (except for reintroduced populations). Hence even if the genetic origin of an individual was known (e.g., based on genetic analysis), the cell it shows up in will not represent the animal’s origin, but rather the population of the location at which it was detected. Cells in-between existing populations and geographic regions and single cells outside of existing populations were given the Status “Unassigned” - even if their genetic origin was known. An exception to this concerns lynx which derive from many reintroductions. Here we have used the origins as an additional factor in population separation. In the last cycle of large carnivore mapping for 2021-2016, we delineated 11 populations for lynx, 10 for bears, 9 for wolves, 4 for golden jackals, and 2 for wolverines (Kaczensky et al. 2021, Ranc et al. 2022, Kaczensky et al. 2024).

2.1.5. Border cells

Many populations occur along and across national borders and cells are shared by neighbouring countries. Only a few populations reported in a coordinated way for a crossborder region (e.g., for bears in the Pyrenees, lynx in the Alps, and all species in Norway & Sweden). Where large carnivore presence was detected in the same cell by neighboring countries, the cell with the “better information status” wins, following the rule: - Permanent > Sporadic > Undefined - Confirmed presence > Confirmed and extrapolated presence > Confirmed and unconfirmed presence > Extrapolated presence > Unconfirmed presence

2.1.6. Time period

The majority of the large carnivore presence layers cover the time period 2017-2022/23. But because monitoring conditions and logistics vary between countries, regions, and species, we documented the specific monitoring periods. Periods referring to 2022/23 usually mean that the species range or population estimates are reported at the end of winter in early 2023. However, the use of 2022 for the calendar year versus 2022/23 for the biological/monitoring year was not consistently used and therefore 2022/23 can mean: only the first part of 2023 is included, all of 2023 is included, or all or parts of 2023 and the first part of 2024 is included. For all countries that provided data deviating from the 2017 start and 2022 or 2023 end date, we explicitly label the monitoring period on one map together with the range.