Table 2. Optimal models of environmental variables predicting seahorse presence/absence (i) and density (ii) in the Rio Formoso Estuary, Brazil.

Model	Predictor variables	Estimate	Std. error	z value	p -value	AIC	AICc	Δ
i: Presence/absence ∼habitat + (1—river)						210.1	211	0
	Intercept	−0.398	0.142	−2.791	0.005 (**)
	Habitat: II	−0.171	0.311	−0.551	0.581 (ns)
	Habitat: III	−19.819	5789.562	−0.003	0.997 (ns)
	Habitat: IV	−0.006	0.721	−0.010	0.992 (ns)
	Habitat: V	−1.741	0.721	−2.414	0.015 (*)
ii: Density ∼depth + habitat + (1—river)						447.7	449.0	0
	Intercept	1.586	0.354	4.476	7.61e−06 (***)
	Depth	−0.014	0.006	−2.240	0.025 (*)
	Habitat: II	−0.074	0.363	−0.205	0.837 (ns)
	Habitat: III	−19.973	6991.369	−0.003	0.997(ns)
	Habitat: IV	−0.449	0.729	−0.616	0.538 (ns)
	Habitat: V	−2.217	0.733	−3.023	0.002 (**)

Notes.

Significance levels were as followed: ns p > 0.05; * p ≤ 0.05; ** p ≤ 0.01; *** p ≤ 0.001.