A major challenge in environmental modeling is to identify structural changes in the ecosystem across time, i.e., changes in the underlying process that generates the data.
In this paper, we analyze the Baltic Sea food web in order to 1) examine potential unobserved processes that could affect the ecosystem and 2) make predictions on some variables of interest. To do so, dynamic Bayesian networks with different setups of hidden variables (HVs) were built and validated applying two techniques: rolling-origin and rolling-window. Moreover, two statistical inference approaches were compared at regime shift detection: fully Bayesian and Maximum Likelihood Estimation.
Our results confirm that, from the predictive accuracy point of view, more data help to improve the predictions whereas the different setups of HVs did not make a critical difference in the predictions. Finally, the different HVs picked up patterns in the data, which revealed changes in different parts of the ecosystem.
Research news | 2020-02-21
Despite rapid urban growth, agriculture in a wetland area in the south of Mexico City soldiers on, more than a millennium after its birth
Research news | 2020-02-19
Malin Falkenmark calls for a shift towards a water based biosphere stewardship. The alternative, she warns, could be catastrophic
Research news | 2020-02-17
Trying to reach the goals under current business-as-usual will come at a heavy price on the planetary boundaries
Research news | 2020-02-14
The new "Our Future On Earth" report provides risks analysis based on survey of 222 global sustainability experts, including centre researchers
Research news | 2020-02-13
New assessment aims to fill critical gaps in understanding the growing role aquatic foods play in the global food system
Research news | 2020-02-09
A new study harmonizes the water planetary boundary with local boundaries for the La Cienega wetlands in Colombia