Study On Degree And Carbon Flux Distribution Of Food Web Based On The Maximum Entropy Model

The stability of ecosystems has always been the focus of ecologists,food web is the most important part of the ecosystem,and also is a manifestation of material circulation and energy flow,at the same time,it reflects the complex interspecific relationship between communities.This study will explore the stability of ecosystems from the two aspects of food network topology and energy flow.Estimating reasonable model coefficients by applying the maximum entropy model to the food network material energy distribution,and exploring new methods to study the distribution of species in nature has important guiding significance for protecting the diversity of global ecosystems.The data set was compiled and provided by professor Stuart Borrett from the university of north Carolina Wilmington,and has been extensively studied by predecessors.Based on the principle of maximum entropy,the research established and tested the degree distribution and carbon flux distribution of food web.The results show that:(1)In the 20 food webs,80%of the maximum entropy models based on the degree distribution of the maximum entropy model are not significantly different from the empirical distribution,and the fitting effect is good.In the test of confidence interval width,the confidence interval width of nearly 70%food webs was significantly different from the maximum entropy model.This shows that although the distribution curve and data growth trend are consistent with the empirical distribution,the degree of prediction of each data point in most food webs tends to be high or low.(2)After studying the influence of S and C on the maximum entropy model of degree distribution,it is found that the number of species S influences the goodness of fit fG and the relative width W95 There is a positive correlation,and when the connectivity C value is large or small,the goodness of fit and width of the model are relatively good.Therefore,using S?C for maximum entropy hypothesis,the application to the modeling of food web degree distribution is appropriate,but along with the increase of the species in number,some unknown factors derived from food web itself restrain the distribution of the degree,affect the deviation from the empirical distribution of maximum entropy,and have less effect in the species of small food web.(3)In the carbon flux distribution of food webs established based on the maximum entropy model.45%of the food webs(9)had good fitting effect,and 55%of the food webs(11)had poor fitting effect.The ecosystem with the best fitting effect was the lake ecosystem,while the marine ecosystem had the worst fitting effect.In addition to the extremely good fitting effect and the extremely poor fitting effect,the maximum entropy distribution of carbon flux in most food webs is high in the predicted value of the position with large carbon flux,and the fitting effect is poor in the food webs with very small food webs(number of connections<100)and very large food webs(number of connections>1000).For complex food webs like marine ecosystems-the distribution of carbon flux is more suitable for sub-nutrition or separate the distribution of consumer resources.This was also reflected in previous research on carbon flux-related modeling in food webs.(4)The results of modeling the average carbon flux of species in the food web based on the maximum entropy distribution show that the fitting degree is not ideal,which does not mean that the maximum entropy model cannot fit them with data,but the maximum entropy model constraints of the two distributions still need to be explored more.