In addition, for each relevé we calculated the mean Ellenberg indicator values of temperature and moisture weighted on species cover.
Moreover, we assigned to each species the Ellenberg indicator value for temperature and moisture, and calculated the means in each relevé.
pa: the analyzed indicator value.
To characterise the microhabitat conditions, the mean light- and humidity indicator values [ 38] of the three most abundant herbaceous plant species at each sampling site were recorded (variables LIGHTIND and HUMIND).
Different key performance indicator values have been calculated by means of custom prediction models based on system characteristics.
The aggregation rules applied were: Hierarchical mean (see Tables S3 for the structure of aggregation), where at each step of aggregation, the transformed indicator values were averaged.
Indicator values range from 1 to 100.
The balancing factor reflects the importance of the maximal deviation of the indicators, where the maximal deviation means the maximum difference between an indicated value and the best value for the indicator (Dahab and Lee [1991]).
If any of the baseline BPs cannot be estimated, we will substitute the overall mean baseline value, set an indicator variable to denote an imputed value and include the indicator variable in the model.
We used three values to estimate the statistical distribution for each set Lijkt: the mean observed value of the indicator, and the associated lower and upper quartiles.
