Large‐scale habitat selection by agricultural steppe birds in Spain: identifying species–habitat responses using generalized additive models

Abstract

Summary1. Predictive models of species' distributions are used increasingly in ecological studies investigating features as varied as biodiversity, habitat selection and interspecific competition. In a pilot study, we based a successful model for the great bustard Otis tarda on advanced very high resolution radiometer (AVHRR) satellite data, which offer attractive predictor variables because of the global coverage, high temporal frequency of overpasses and low cost. We wished to assess whether the approach could be applied at very large spatial scales, and whether the coarse resolution of the imagery (1 km 2 ) would limit application to those bird species with large home ranges or to simple recognition of broad habitat types. 2. We modelled the distributions of three agricultural steppe birds over the whole of Spain using a common set of predictor variables, including AVHRR imagery. The species, great bustard, little bustard Tetrax tetrax and calandra lark Melanocorhypha calandra , have similar habitat requirements but differently sized home ranges, and are all species of conservation concern. Good models would reveal differences in distribution between the species and have high predictive power despite the large geographical extent covered. 3. Generalized additive models (GAMs) were built with the presence-absence of the species as the response variable. Individual species' responses to the habitat variables were identified using partial fits and compared with each other. We found that this modelling framework could successfully distinguish the habitats selected by the three species, while the response curves indicated how the habitats differed. Model fits and cross-validations assessed using receiver operating characteristic (ROC) plots showed the models to be successful and robust. 4. We overlaid the predictive maps to identify key areas for agricultural steppe birds in Spain and compared these with the present network of protected sites in two sample regions. In Castilla León the provision of protected sites appears appropriate, but in Castilla La Mancha large areas of apparently suitable habitat have no protection. 5. These results confirm that large-scale models are able to increase our understanding of species' ecology and provide data for conservation planning. AVHRR imagery, in combination with other variables, has sufficient resolution to model a range of bird species, and GAMs have the flexibility to model subtle species-habitat responses.

Density and habitat preferences of male little bustard across contrasting agro-pastoral landscapes in Sardinia (Italy)

“…Contrary to other studies (Suarez-Seoane et al 2002;Silva et al 2004;Osborne and Suarez-Seoane 2007;Silva et al 2010), we found no effect of proximity to anthropogenic infrastructures on the occurrence of little bustards at the landscape scale. However, at the smaller scale, displaying males located at points further from roads, suggesting that when not constrained by the availability of pastures and new fallows, roads were avoided.…”

“…These land-uses, associated with low-intensity agro-pastoral activities, are rapidly disappearing from most of the Palaearctic but still occur locally in southern Mediterranean regions and are well known as a suitable habitat for the little bustard in Iberia (Martinez 1994;Salamolard and Moreau 1999;Suarez-Seoane et al 2002) and southern France (Wolff et al 2001). A strong preference for high heterogeneity of land-use substrates was also evident in the non-spatial model but not in that incorporating spatial autocorrelation.…”

“…We also compared density estimates for displaying males among three contrasting landscapes of western Sardinia that differ in land-use composition and structure, recent land-use trends and disturbance levels. From available knowledge of habitat preference of the species, we predicted that males' occurrence would be determined by both large-scale landscape features and small-scale habitat characteristics, with densities higher in the landscape with a balance of pasture and cropland cover and low disturbance levels (Lane et al 2001;Suarez-Seoane et al 2002), and predicted lower densities in the more homogeneous landscapes exposed to greater levels of disturbance.…”

Review article: Thirty years of analysing and modelling avian habitat relationships using satellite imagery data: a review

“…The Advanced Very High Resolution Radiometer (AVHRR) from the National Oceanic and Atmospheric Administration (NOAA) was applied in 22 studies; six times it was used together with other satellite sensors. These coarseresolution imageries were often used in large-scale studies covering entire or large parts of countries (Wallin et al 1992, Suarez-Seoane et al 2002, Venier et al 2004, continents (O'Connor et al 1996, Johnson et al 1998, Roshier et al 2001, HAcevedo and Currie 2003 or the ocean (Haney and McGillivary 1985a, Jouventin et al 1994, Huettmann and Diamond 2001. With the help of NOAA AVHRR imagery often the actual or integrated normalized difference vegetation index (NDVI) was calculated to produce maps of primary production or to detect green biomass (Wallin et al 1992, Andries et al 1994, Hepinstall and Sader 1997, Verlinden and Masogo 1997, Osborne et al 2001, Saveraid et al 2001, in order to detect habitat change (Sader et al 1991, Debinski et al 2000, to map wetland distribution (Roshier et al 2001) or to explain the distribution of centres of bird endemism (Johnson et al 1998).…”

Linking habitat quality with genetic diversity: a lesson from great bustards in Spain

“…Recent developments in geographic information systems (GIS) and habitat modelling provide an ideal methodological framework in which high-resolution analyses of habitat attributes can be performed at the landscape level. Models allow gaining insights into the environmental variables (such as biotic, physical, climatic or human-induced habitat conditions) acting on species survival by comparing current species distribution with the optimal values for each variable (Guisan and Zimmermann 2000;Osborne et al 2001;Suárez-Seoane et al 2002). The output of the models is a habitat suitability index (HSI) which ranges from 0 (low quality) to 1 (high quality).…”