Testing Bottom-up and Top-Down Processes in Communities: A Cointegration Approach

Bottom-up ( resource controlled ) and top-down ( predator controlled ) processes are the two widely used ecological concepts to represent species interactions in ecological communities. Previous studies took a correlation approach to resolve these two types of processes in natural communities. Here, we show three common situations, namely non-stationarity, misalignment of times series and complex community structure, that imperil inferences  from the correlation approach. In this paper, from the perspective of factor analysis, we define a top-down process as one where the influence of species interaction only flows from a higher trophic level to a lower trophic level, and vice versa for a bottom-up process. This definition is crucial in resolving the community structure and overcoming those difficulties mentioned above. The new statistical method employed in this study is called a  cointegration method. Together with the concepts defined above, the cointegration method can be used to analyze non-stationary and possibly mis-aligned multi-dimensional time series data to resolve complex community structure, which is often partially observed. Then, we use the Gulf of Mexico SEAMAP (Southeast Area Monitoring and Assessment Program) survey data as an example to illustrate the testing procedure.