Visibility Graph Network Analysis Of Natural Gas Price Fluctuations:The Case Of North American Market

As energy supply and climate problem is increasingly serious, natural gas, as a type of clean and efficient alternative energy, plays an important role on improving energy security and reducing carbon emissions, its price fluctuations will have a significant impact on the global economy. Therefore, the research on the characteristics of natural gas price fluctuations is of great importance. Currently, a unified global natural gas market is not formed, global natural gas trade concentrates on three regional markets: the North American market, the European market and the Asia-Pacific market. After one hundred years’ development, the North American market gradually formed a mature and open market system by improving the legal supervision and releasing the control of price. The fluctuations of North American natural price also can represent the fluctuate direction of international natural gas price.In this paper, we study on the North American natural gas spot market specifically, investigate the center nodes of gas price and its influential period on subsequent price. We converting the natural gas spot price time series into a visibility graph network, investigating several indicators: degree and degree distribution, average path length, community structure; the internal mechanism underlying price fluctuation have been explored from the perspective of complex network theory. It provide a new direction for macro analysis of time series. The study found that, the natural gas spot price complex network have small-world and scale-free property simultaneously. After random rearrangement of original gas prices time series, the network degree distribution character becomes exponential distribution which is different from the original ones, that means the original gas price time series is a long-range negative correlation fractal time series. Nodes which have large degree correspond to significant geopolitical or economic events and communities corresponding to time periods in visibility graph networks. The cycle of time series and the impact scope of the central nodes can be found by community structure partition.