Can Eltonian processes explain species distributions at large scale? A case study with Great Bustard (Otis tarda). Issue 2 (11th September 2014)

• Record Type: Journal Article
• Authors: Alvarez‐Martínez, Jose Manuel; Suárez‐Seoane, Susana; Palacín, Carlos; Sanz, Julia; Alonso, Juan C.; Beaumont, Linda
• Is Part Of: Diversity & distributions. Volume 21:Issue 2(2015:Feb.)
• Page Start: 123
• Page End: 138
• Publication Date: 2014-09-11
• Subjects: Biodiversity -- Periodicals; Biodiversity conservation -- Periodicals; 577
• DOI: 10.1111/ddi.12256
• Languages: English
• ISSNs: 1366-9516

Abstract

<abstract abstract-type="main" id="ddi12256-abs-0001"> <title>Abstract</title> <sec id="ddi12256-sec-0001" sec-type="section"> <title>Aim</title> <p>Species distribution modelling (SDM) allows identification of suitable areas where the species might occur. The Grinnellian approach relates species occurrence to abiotic limiting factors, such as topography and climate, measured at present or averaged over time at large spatial scales, whilst Eltonian approaches relate processes as reproductive rates or species abundance to factors as resource availability or functional traits described at a finer spatio‐temporal resolution. In this study, we aimed to compare the performance and reliability achieved by Eltonian and Grinnellian models developed at geographical scale when addressing two questions that may affect species distribution: breeding site selection and historical local extinctions.</p> </sec> <sec id="ddi12256-sec-0002" sec-type="section"> <title>Location</title> <p>Spain.</p> </sec> <sec id="ddi12256-sec-0003" sec-type="section"> <title>Methods</title> <p>We used a temporal series (1988–2010) of breeding success data of Great Bustards (<italic>Otis tarda</italic>) across Spain to create: (1) a yearly series of correlative Eltonian models based on satellite data (1‐km pixel Normalized Difference Vegetation Index derived from NOAA‐AVHRR) as a proxy for resource availability and (2) a static topo‐climatic Grinnellian model. We applied two modelling algorithms (MaxEnt and Bootstrapped Generalized Mixed‐Effects Models) calibrated and validated across different distributional ranges to investigate breeding success in abiotically suitable areas and recent extinctions within the accessible range.</p> </sec> <sec id="ddi12256-sec-0004" sec-type="section"> <title>Results</title> <p>Eltonian models showed higher performance and reliability than Grinnellian models at geographical scale and allowed for comparisons of breeding habitat suitability through the study period. Then, satellite data were a more successful predictor of the species breeding success than topography and climate. Breeding was possible out of the range currently occupied by the species, but habitat suitability was significantly lower in non‐occupied areas, on where local extinctions have occurred in recent historical times.</p> </sec> <sec id="ddi12256-sec-0005" sec-type="section"> <title>Main conclusions</title> <p>This study highlights the advisability of using remote‐sensing explanatory variables at high temporal resolution matched with population fitness data in Eltonian approaches to improve the knowledge of the relationship between realized and potential distributions of the species at geographical scale. This is a fundamental question in biogeography and a key issue for designing comprehensive conservation strategies.</p> </sec> </abstract>