Diagnosing and Predicting the Earth's Health via Ecological Network Analysis

Ecological balance is one of the most attractive topics in biological, environmental, earth sciences, and so on. However, due to the complexity of ecosystems, it is not easy to find a perfect way to conclusively explain all the potential impacts. In this paper, by considering several important elements, we seek to build a dynamic network model to predict the Earth’s health, trying to identify and explain how the human behavior and policies affect the model results. We firstly empirically analyze both the topological properties and time-dependent features of nodes and propose an Earth’s health index based on Shannon Entropy. Secondly, we identify the importance of each element by a machine learning approach. Thirdly, we use a spreading model to predict the Earth’s health. Finally, we integrate the topological property and the proposed health index to identify the influential nodes in the observed ecological network. Experimental results show that the oceans are the key nodes in affecting the Earth’s health, and Big countries are also important nodes in influencing the Earth’s health. In addition, the results suggest a possible solution that returning more living lands might be an effective way to solve the dilemma of ecological balance.