SEMICE

Seguimiento de los
micromamíferos comunes
de España

CA | ES | EN

Data analyses

Population trends

In order to assess how small mammal populations fluctuate over time, a rate of change is calculated using the TRIM (TRends and Indices for Monitoring data) method. Population trends are estimated for each of the six most common species in Catalonia and Andorra (they appear regularly at more than 10 SEMICE stations): wood mouse (Apodemus sylvaticus), common shrew (Crocidura russula), bank vole (Clethrionomys glareolus), yellow-necked mouse (Apodemus flavicollis), Algerian mouse (Mus spretus) and garden dormouse (Eliomys quercinus). On the other hand, only plots with a minimum of two years of data are taken into account.

The calculation of the trend is based on the number of different individuals captured at each SEMICE monitoring station during each campaign. As this trend can vary depending on the area and the scale used to calculate it, the graphs show three levels for each species: a general trend (considering all plots in Catalonia and Andorra), an alpine and sub-alpine one (Pyrenean and Pre-Pyrenean stations) and one for the Mediterranean region (Litoral and Prelitoral mountain ranges and other more southern mountain ranges).

TRIM calculates indices for each period based on a Generalized Linear Model (GLM), assuming that the counts follow a Poisson distribution. It works with two covariates: site (it enables abundance to differ between sites) and time (from which it calculates the average increase for each level: it treats the increase or decrease for a particular period as the same across all localities and multiply it by the previous index). When a monitoring station has not been sampled in several campaigns, TRIM also calculates their expected values from the changes observed in the sampled plots.

The calculated indices do not reflect the actual abundance of the species, but indicate the percentage change between two periods, with an associated standard error. In addition, they are corrected to minimize the effect of two aspects: overdispersion, when the variance of the results is higher than we would expect by chance, and autocorrelation, which means that the number of captured individuals tends to depend of those captured during the previous campaign.

The first campaign from which the trend of a species is calculated receives the value 100. The rest of the indices are then obtained based on this reference value, thus facilitating visualization and comparison among different species.

The population trend of the (Apodemus sylvaticus) in Catalonia and Andorra varies according to the level at which we analyse it. Despite showing some stability at a general level (considering all the stations in the SEMICE network), when we divide the data we see that in the Mediterranean region there is enough evidence to classify its trend as negative. On the other hand, it remains stable in the Pyrenean stations. By clicking the categories in the legend, you can activate the different graphs.