The database compiled for the Braun-Blanquet project is a compilation of various national and regional vegetation databases. The maintenance of these databases is in principle in the hands of the custodians. However, before uploading the databases into Braun-Blanquet database a quality check is performed by Alterra and Masaryk University. If possible, detected errors are corrected and reported back to the data provider. For the modelling of the habitat suitability map the Maxent software is used ( http://www.cs.princeton.edu/~schapire/maxent/). The grid values in the map represent the probability (ranging from 0-1) that the cell is suitable for the habitat.

The grid file represents the habitat suitability of the EUNIS type. For the modelling the widely used software Maxent for maximum entropy modelling of species’ geographic distributions was used. Maxent is a general-purpose machine-learning method with a simple and precise mathematical formulation, and has a number of aspects that make it well-suited for species distribution modelling when only presence (occurrence) data but not absence data are available (Philips et al. 2006). Because EUNIS habitats have a particular species composition, they are assumed to respond to specific ecological requirements, allowing us to generate correlative estimates of geographic distributions. Modelling habitats that have been floristically defined is a well-known procedure for ecological modelling at local scales, and a promising technique to be applied also at the continental level.

The Maxent method considers presence data (known observations of a given entity) and the so-called background data. Background data comprise a set of points used to describe the environmental variation of the study area according to the available environmental layers. It is assumed that these layers represent well the most important ecological gradients on a European scale. As layers the following environmental parameters have been used: Potential Evapotranspiration, Topsoil pH, Solar radiation, Temperature Seasonality (standard deviation *100), Mean Temperature of Wettest Quarter, Annual Precipitation, Precipitation Seasonality (Coefficient of Variation), Precipitation of Warmest Quarter & Distance to water (rivers, lakes, sea).

Statistical output of the model:

#Training samples: 2368

Regularized training gain: 0.9878

Unregularized training gain: 1.0481

Iterations: 500

Training AUC: 0.8628

#Test samples: 263

Test gain: 1.0651

Test AUC: 0.863

AUC Standard Deviation: 0.0064

#Background points: 12363

bio_12_etrs2_ras contribution: 0.1128

bio_15_etrs2_ras contribution: 1.4857

bio_18_etrs2_ras contribution: 21.5749

bio_4_etrs2_ras contribution: 58.255

bio_8_etrs2_ras contribution: 0.2729

dist2water1km contribution: 0.0045

pet_he_yr contribution: 14.6326

soil_ph contribution: 0.7197

solar_1km contribution: 2.9419

bio_12_etrs2_ras permutation importance: 0.9595

bio_15_etrs2_ras permutation importance: 0.9637

bio_18_etrs2_ras permutation importance: 16.9937

bio_4_etrs2_ras permutation importance: 59.872

bio_8_etrs2_ras permutation importance: 0.8316

dist2water1km permutation importance: 0.0386

pet_he_yr permutation importance: 16.1328

soil_ph permutation importance: 0.6691

solar_1km permutation importance: 3.539

Entropy: 8.4372

Prevalence (average of logistic output over background sites): 0.1835

Fixed cumulative value 1 cumulative threshold: 1

Fixed cumulative value 1 logistic threshold: 0.0631

Fixed cumulative value 1 area: 0.4362

Fixed cumulative value 1 training omission: 0.0042

Fixed cumulative value 1 test omission: 0

Fixed cumulative value 1 binomial probability: 0.00E+00

Fixed cumulative value 5 cumulative threshold: 5

Fixed cumulative value 5 logistic threshold: 0.2271

Fixed cumulative value 5 area: 0.35

Fixed cumulative value 5 training omission: 0.0236

Fixed cumulative value 5 test omission: 0.0114

Fixed cumulative value 5 binomial probability: 0.00E+00

Fixed cumulative value 10 cumulative threshold: 10

Fixed cumulative value 10 logistic threshold: 0.3212

Fixed cumulative value 10 area: 0.3014

Fixed cumulative value 10 training omission: 0.0714

Fixed cumulative value 10 test omission: 0.0646

Fixed cumulative value 10 binomial probability: 0.00E+00

Minimum training presence cumulative threshold: 0.0968

Minimum training presence logistic threshold: 0.0029

Minimum training presence area: 0.6497

Minimum training presence training omission: 0

Minimum training presence test omission: 0

Minimum training presence binomial probability: 5.34E-33

10 percentile training presence cumulative threshold: 12.5828

10 percentile training presence logistic threshold: 0.3507

10 percentile training presence area: 0.2823

10 percentile training presence training omission: 0.0997

10 percentile training presence test omission: 0.0989

10 percentile training presence binomial probability: 0.00E+00

Equal training sensitivity and specificity cumulative threshold: 24.4193

Equal training sensitivity and specificity logistic threshold: 0.4588

Equal training sensitivity and specificity area: 0.2185

Equal training sensitivity and specificity training omission: 0.2183

Equal training sensitivity and specificity test omission: 0.2395

Equal training sensitivity and specificity binomial probability: 0.00E+00

Maximum training sensitivity plus specificity cumulative threshold: 8.0176

Maximum training sensitivity plus specificity logistic threshold: 0.2921

Maximum training sensitivity plus specificity area: 0.3181

Maximum training sensitivity plus specificity training omission: 0.0494

Maximum training sensitivity plus specificity test omission: 0.0342

Maximum training sensitivity plus specificity binomial probability: 0.00E+00

Equal test sensitivity and specificity cumulative threshold: 23.1332

Equal test sensitivity and specificity logistic threshold: 0.4513

Equal test sensitivity and specificity area: 0.2243

Equal test sensitivity and specificity training omission: 0.2073

Equal test sensitivity and specificity test omission: 0.2243

Equal test sensitivity and specificity binomial probability: 0.00E+00

Maximum test sensitivity plus specificity cumulative threshold: 7.9676

Maximum test sensitivity plus specificity logistic threshold: 0.2914

Maximum test sensitivity plus specificity area: 0.3186

Maximum test sensitivity plus specificity training omission: 0.0494

Maximum test sensitivity plus specificity test omission: 0.0304

Maximum test sensitivity plus specificity binomial probability: 0.00E+00

Balance training omission, predicted area and threshold value cumulative threshold: 1.2849

Balance training omission, predicted area and threshold value logistic threshold: 0.0851

Balance training omission, predicted area and threshold value area: 0.423

Balance training omission, predicted area and threshold value training omission: 0.0051

Balance training omission, predicted area and threshold value test omission: 0

Balance training omission, predicted area and threshold value binomial probability: 0.00E+00

Equate entropy of thresholded and original distributions cumulative threshold: 3.3697

Equate entropy of thresholded and original distributions logistic threshold: 0.1877

Equate entropy of thresholded and original distributions area: 0.3733

Equate entropy of thresholded and original distributions training omission: 0.0139

Equate entropy of thresholded and original distributions test omission: 0.0038

Equate entropy of thresholded and original distributions binomial probability: 0.00E+00