Temporal development and collapse of an Arctic plant-pollinator network

BMC Ecology

Table 2 Model summary

	Parameter	1996	1997	1996&1997	LTR
Arrival of insects Poisson	λ	2.417 (0.317)	2.423 (0.305)	2.420 (0.220)	Yes
	p	0.094	0.15	0.28	0.99
Links for new insects Modified geometric	r ₁	0.414 (0.065)	0.429 (0.062)	0.421 (0.045)	Yes
	r	0.351 (0.049)	0.356 (0.048)	0.354 (0.034)
	p	0.65	0.37	0.55	0.98
Death of insects Sigmoid binomial	α	0.058 (0.012)	0.037 (0.008)	0.046 (0.007)	Yes
	β	0.436 (0.071)	0.469 (0.143)	0.446 (0.063)
	H	4.901 (4.942)	2.797 (3.688)	3.847 (3.984)
	T	19.66 (0.362)	22.31 (0.747)
	p	0.41	0.51	0.75	0.57
Addition of links Geometric	q	0.752 (0.019)	0.836 (0.015)	0.799 (0.012)	No
	p	0.13	0.078	0.031	<10^-3
Removal of links Sigmoid binomial	γ	0.006 (0.002)	0.006 (0.002)	0.008 (0.002)	No
	δ	0.187 (0.021)	0.401 (0.036)	0.271 (0.020)
	h	18.28 (9.497)	40.12 (-)*	48.68 (-)*
	t	19.06 (0.031)	22.55 (-)*
	p	0.15	0.73	<10^-3	<10^-3

Shown are the fitted parameters for the two years; estimated for each year separately and jointly (except T and t) with standard deviations in parenthesis. The p-values in the 3rd, 4th and 5th columns are goodness-of-fit probabilities based on the chi-square test; the p-values in the 6th column are test probabilities for the LRT (Yes/No indicates whether the LRT is accepted or not). Even when the LRT is not accepted the parameters are of similar magnitude in the two years. The distributions are explained in Methods, Analysis and modelling. Standard deviations marked with * could not be estimated reliably.

ISSN: 1472-6785