Answers to these two questions support decisions based on environmental data:
-
Why do we observe the values we have?
-
How can we identify natural variability and anthropogenic effects?
The first question can be answered using an appropriate regression model that relates mean or quantile values of the continuous response variable to the range of values of one or more explanatory variables. Explanatory variables frequently are categorical; that is, names such as soil type, compass direction, and dominant vegetation rather than continuous variables such as temperature and slope. Mixed models analyze continuous and categorical variables in the same model and yield insights into why we obtain the values we have. Using several regression models allows us to select the one best fitting the available data by using the Akaike information criterion (AIC) to measure their relative quality in explaining the data.