Abstract

AbstractAimTo describe the population trend for European turtle doves in Spain. To identify favourable and unfavourable areas for the species and to test whether favourability or land use explain spatial variation in abundance change.LocationMainland Spain.MethodsWe used generalized linear models with extensive abundance data to describe population change for the European turtle dove across Spain. We used breeding distribution (presence/absence) data at 100 km2 resolution to model environmental favourability in relation to topo‐climatic and land use variables. Finally, we tested whether land use and favourability explained spatial variation in abundance trends.ResultsThe large Spanish turtle dove population declined by 37% between 1996 and 2018. Favourability was highest in the south, east and north–west of Spain and lowest in the north and at higher altitudes. Abundance trends were more negative in areas of lower environmental favourability and in localities dominated by arboreal habitats such as forests, “dehesas” (open agro‐forestry landscapes with scattered Quercus trees), transitional woodland shrubs or sclerophyllous vegetation (a mixture of sclerophyllous shrubs with some scattered trees). Trends were more positive in localities dominated by complex cultivation (small parcels of mixed crop types, including woody permanent crops like olive, or almond trees or vineyards).Main conclusionsOur study highlights a substantial recent decline in the numerically important turtle dove breeding population in Spain. Declines in abundance were more strongly associated with arboreal (forest and shrub areas) rather than agricultural habitats, highlighting an urgent need for further research into the ecology of this important quarry species in arboreal breeding habitats in southern Europe.

Highlights

  • To assess temporal changes in abundance across mainland Spain, we fitted a Generalized Linear Model (GLM) to the count data with year as continuous variable (1996 to 2018) and UTM cell identity as additive explanatory variables, specifying a negative binomial error distribution to allow for overdispersion

  • We assessed the classification power of the logistic model by calculating sensitivity, specificity and their correct classification rate (CCR), using a favourability value of F = 0.5 as the classification threshold, and we evaluated the discrimination capacity using the area under the curve (AUC) of the receiver operating characteristic, which is independent of any favourability threshold

  • We aimed to identify which land uses are most strongly associated with population change, in order to provide a land use focus for future diagnostic research and potentially conservation action

Read more

Summary

Introduction

To assess temporal changes in abundance across mainland Spain, we fitted a Generalized Linear Model (GLM) to the count data with year as continuous variable (1996 to 2018) and UTM cell identity as additive explanatory variables, specifying a negative binomial error distribution to allow for overdispersion. We tested whether variation in temporal trends in turtle dove abundance (1996–2018) across UTM cells was related to their favourability.

Objectives
Results
Conclusion
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call