Using network analysis to time India’s monsoon
As social networks evolve, the analysis involved with studying the complexity of these networks are evolving as well. Specifically, social network analysis are starting to be used to predict trends in climate systems. Researchers at the Potsdam Institute are using network analysis to predict the timing of the Indian monsoon onset and offset. The heavy summer rains are of vital importance for millions of farmers feeding the subcontinent’s population. Slight deviations in the onset and offset of these rains can lead to drought or flooding, leading to millions of dollars in damages and lack of food in an already impoverished region of the world. Global warming is expected to make the onset and offset of these rains erratic, making it increasingly important to find quicker and more accurate ways of predicting them.
Jürgen Kurths, head of PIK’s research domain Transdisciplinary Concepts & Methods, compares climate systems to complex social networks, with geographical regions communicating in similar ways to how people communicate on the internet. Like Facebook postings or tweets that get shared multiple times, temperature and humidity get transported from one place to another by atmospheric flows, such as winds (the spread of information over a distance in time). Variables like temperature and humidity can be considered nodes, while wind can be considered a connection (edge). Insight into where winds are headed at a specific time could yield better prediction methods for the temperature and humidity at the final destination.
The network analysis is showing good results so far. However, the major innovation of the project is combining the network analysis with the statistical analysis of weather patterns. Integrating these two systems together would probably yield the most accurate results for the future.