Distribution and Coverage of Environmental Variables

Electronic Supplementary Material 1

Density and Spatial Distribution of Buffy-tufted-ear Marmosets (Callithrix aurita) in a Continuous Atlantic Forest

Darren Norris ∙ Fabiana Rocha-Mendes ∙ Renato Marques ∙ Rodrigo de Almeida Nobre ∙ Mauro Galetti

Distribution and Coverage of Environmental Variables

To determine how well our extensive census encompassed the environmental gradients in the study areas we compared the distribution of the 6 environmental predictors used to determine habitat suitability for Callithrix aurita between the area covered by our overall line-transect survey effort and the sample area (Núcleo Cunha/Indaiá and Núcleo Santa Virgínia).

For our 4 continuous variables [(a) = EVI (enhanced vegetation index), (b) = NDVI (normalized difference vegetation index), (c) = TWI (topographic wetness index), and (d) = DEM (digital elevation model)] we compared the probability distribution of the environmental variables surveyed 1) over 877.7 km of diurnal line-transect census and 2) throughout the sample area using the nonparametric Kolmogorov-Smirnov test. We used Pearson’s χ2 test to compare the frequency distributions of our categorical variables [(e) = land cover and (f) = vegetation class].

Electronic Supplementary Material 2

We derived a refined density estimate (Dr) for each grid cell by rescaling density estimates by the habitat suitability index (HSI) calculated via MAXENT:

where is the lower 95% CI density estimate scaled to the cell size (in our case 4.61 inds/km2 = 0.0041 inds/900 m2) and is the adjusted density obtained from:

where is the cell HSI value produced by MAXENT, HSImin is the lowest HSI value (in our case we chose 0.125 to represent a widespread starting point for a realized distribution), is the range of HSI values from the region of interest (maximum HSI – HSImin) and is the 95% CI range of density estimates scaled to cell size (upper 95% CI – lower 95% CI). We then obtained an overall density by summing over the study area.

The process can be separated into 4 calculation steps:

1)  Derive HSI map and establish minimum (starting) HSI value.

2)  Convert range of 95% CI density estimates to correspond to the grid size.

3)  Calculate density adjustment: .

= 0.0110 – 0.0041 = 0.0069

= 0.8 – 0.2 = 0.6

= 0.0115

4)  Calculate the refined density in the sample area.