The current loss in biodiversity also includes the loss of widespread and common species. Small declines in common species like the common toad, already have a large impact on the ecosystem. To avert further biodiversity loss, understanding the causes of toad declines is key. As a first step to investigate causes behind toad declines, analysis of decline rates is needed. A sufficiently long time series of counting data is available for more than 200 populations. Here we show how to determine the variation in trends among populations in Flanders. The shorter time series are analyzed using Generalized Linear Model (GLM), while the others used Generalized Additive Models (GAM) making it possible to look at different contrasts: end-begin timepoint and average slope. As we define the trend for multiple populations, we corrected for multiple testing using the false discovery rate (FDR). In most of the locations with a significant trend in the data, we found a decrease in the populations, which are comparable to results in other countries (Italy, Switzerland, UK & Canada). This thesis demonstrates how to perform trend analysis, using GAM as alternative to the commonly used GLM, while still achieving correct statistical inferences.

Read more...