Figure captions
Figure 1. Experimental design of the artificial substrates placed within in the streams. A total of five substrates (AS) were placed in four rows transversal to the stream. The frequency of the disturbance (control, F1, F2 or F3) was assigned randomly to each row of each stream. An entire column was randomly selected to be sampled at each sampling procedure. The same design was used for the two times the experiment was conducted, with different intensities of disturbance.
Figure 2: Immediate effect of the disturbance on the permanence probability of the functional groups. Effect is the parameter estimated in the model that affects the probability of occurrence.
Figure 3: Probability density function (PDF) of permanence (\(\psi-\ \text{blue}\)) and colonization parameters (\(\theta-\text{red}\)) for functional groups of aquatic insects, during 39 days in streams of the Brazilian Cerrado. The letters A - I correspond to the functional groups of 1 - 9 (A = Depressed body Coleoptera; B = Cylindrical Coleoptera; C = Chironomidae; D = Ceratopogonidae; E = Empididae and Simuliidae; F = Baetidae; G = Depressed body insects; H = Larvae with anal claws; and I = Tricoptera shelter builders).
Figure 4. Effects of experimental treatments (disturbance intensity and frequency) on the probability of permanence, with the mean values and credible intervals (95%). A = Depressed-bodied Coleoptera; B = Cylindrical Coleoptera; C = Chironomidae; D = Ceratopogonidae; E = Empididae and Simuliidae; F = Baetidae; G = Depressed-bodied insects; H = Larvae with anal claws; and I = Trichoptera shelter builders. C=control; F1, F2, and F3=frequency of disturbance (interval in days of 6, 3, and random). The gray band corresponds to the credibility interval of the control.