Glorious past, uncertain present, bad future? Assessing effects of land-use changes on habitat suitability for a threatened farmland bird species

Abstract

Land-use changes have strong impacts on biological communities. Among them, land abandonment is threatening a large number of conservation-concern species associated with semi-natural habitats shaped by ‘traditional’ farming. We focused on the red-backed shrike as a model for investigating the effect of land abandonment on a threatened bird species, and used historical data to model dynamic scenarios. We explored variations in habitat suitability from the 1950s to the present and predicted possible future variations. After investigating local habitat preferences of the species, we formulated a spatially explicit model of habitat suitability for shrikes according to current land-use types; then, we evaluated past habitat suitability, by applying the model to three known past scenarios, and simulated the habitat changes after land abandonment. By combining a habitat-association approach with past and future land use scenarios, we assessed and predicted the effects of habitat changes caused by abandonment. Shrike occurrence was favoured by the cover of four types of grassland and of shrubland with trees, and negatively affected by broadleaved woodlands. The current average habitat suitability is less than half of what it was in the 1950s. Future predictions in a complete abandonment scenario suggest that important decrease could be expected 10 or 20 years after abandonment, and that after 30 years the red-backed shrike would be completely extinct. Alternative scenarios involving partial abandonment suggested that subsidy policies may mitigate the effects of abandonment. Knowing land-use dynamics allowed the exploration of effects of land-use changes and corroborated the importance of low-intensity farming for conservation.

Introduction

Changes in land-use have a global effect (Leemans and Zuidema, 1995, Hurtt et al., 2006) and, interacting with other components of global change, have a strong impact on the structure of biological communities (Sala et al., 2000, Thomas et al., 2004, Rosenzweig et al., 2008, Gil-Tena et al., 2009). Mediterranean ecosystems appear to be especially susceptible to the impacts of global change (Lavorel et al., 1998), because of the high exposure to human activities (Blondel, 2006), the sensitivity to climatic conditions (Peñuelas et al., 2002) and the simultaneous impact of several factors (Sala et al., 2000). The effects on biodiversity of such impacts are largely unknown (Sala et al., 2000), may be non-linear and may display opposite directions (Gil-Tena et al., 2009).

Land abandonment has been an important land-use change in recent decades, being a major consequence of socio-economic changes in highly industrialized countries (Ostermann, 1998). The decrease in farming primarily affects the least productive agricultural land and triggers a recovery of semi-natural vegetation (Sirami et al., 2008). In most of the Mediterranean region, land abandonment has occurred during the last century. This has led to the naturalization and vegetation closure of many areas, thus favouring the spread of forests (e.g. Debussche et al., 1999, Poyatos et al., 2003, Roura-Pascual et al., 2005). The Northern Mediterranean basin has a long-term and complex history of human exploitation, which has largely shaped its landscapes and associated fauna (Blondel, 2006). It has recently experienced fast and widespread agricultural abandonment (Mazzoleni et al., 2004), which appears to be an important cause of species declines, especially in bird communities (Farina, 1995, Farina, 1997). The recovery of semi-natural vegetation that results from agricultural abandonment causes a decrease of open grassland-like habitats and an increase in shrubland and, especially on the long-term, woodland cover (Debussche et al., 1999, Romero-Calcerrada and Perry, 2004). This causes a decline of species tied to open habitats (Preiss et al., 1997, Suárez-Seoane et al., 2002, Sirami et al., 2007), with migrant species associated with open farmland habitats being particularly affected (Sirami et al., 2008). In general, many species of conservation concern in Europe are associated with traditional farm landscapes and the semi-natural habitats they produce and maintain (see e.g. Tucker and Evans, 1997). Such types of landscapes have been increasingly studied in order to understand mechanisms underlying the decline of farmland birds (Benton et al., 2003) and the main steps to be taken for their conservation (Donald et al., 2002).

On the basis of the above, assessments are urgently required on the impact (past, current, and likely future) of environmental change, on animal communities and especially on threatened bird species in the Mediterranean region. Assessments should especially include land abandonment and the consequential fast recovery of forest, further boosted by climate change and frequency of fires (Grigulis et al., 2005).

However, as it is rare to have enough information on past land-uses: (i) it is difficult to assess previous ecological gain or loss, and (ii) future predictions are often built up without a historical base. Although notable exceptions exist (see e.g. Domenech et al., 2005, Ficetola et al., 2009, Fouquet et al., 2010), none of them consider farmland birds, one of the main concerns for conservationists at the beginning of the third millennium.

Among depleted and declining farmland birds, the red-backed shrike Lanius collurio is a typical inhabitant of semi-open habitats created and maintained by ‘traditional’, low-intensity agricultural practices (Fornasari et al., 1997). A recent study revealed that shrike conservation in southern European farmed landscapes depends on a trade-off between ‘intense’ agricultural land use and ‘partial’ land abandonment, which provide complementary resources (Brambilla et al., 2007a). Other research in different European countries has highlighted a similar pattern: red-backed shrikes occupy mostly farmed landscapes characterised by a matrix of fields and grasslands, within which marginal features are interspersed, such as shrubs, hedgerows and small trees (Kuzniak and Tryjanowski, 2000, Vanhinsbergh and Evans, 2002, Karlsson, 2004, Tryjanowski et al., 2007, Golawski and Golawska, 2008, Golawski and Meissner, 2008, Brambilla et al., 2009a). These widely converging findings, coming from different geographical contexts, emphasise how red-backed shrikes are principally tied to an ‘intermediate’ ecological disturbance created by agro-pastoral activities over biological communities and environment. Red-backed shrikes need open areas with low grass provided by pastures, mown grasslands or grassy arable-fields; they also need shrubs or small trees for perching and nesting sites, which are found in areas not directly used for agricultural purposes, or by portions of pasture with a lower grazing pressure. Finally, untilled herbaceous plants favour shrike occurrence by positively affecting the abundance and availability of invertebrate prey; untilled grasslands are often ensured by fallow land or set-aside. When a similar mosaic pattern (which in most European countries does not occur as a ‘natural’ feature but only as a by-product of traditional agricultural land use) is created ‘ad hoc’ by habitat management, shrikes respond positively to this improved habitat quality with a steady population increase, stressing the link between this species and the habitat mosaic created by low-intensity agriculture (Casale et al., 2007, and references therein).

Red-backed shrikes are quite easy to note and locate, thanks to their conspicuous behaviour, and previous spatially explicit models of habitat suitability for this species, based on land use features, lead to valid predictions of species abundance (Brambilla et al., 2009a). The red-backed shrike in the northern Mediterranean basin thus appears to be an ideal model for investigating the effect of habitat changes, caused by land abandonment, on threatened farmland birds.

With this work we model habitat suitability for the study species according to dynamic land-use contexts. We analyse the species’ habitat requirements in a landscape widely shaped by traditional agricultural use, where historical changes in land use and associated vegetation structure are known; then, we explore variations in habitat suitability from the 1950s to the present and predict future habitat suitability in a land-abandonment scenario.

The simultaneous occurrence of a range of neighbouring land-types allows us to investigate how strictly red-backed shrikes are dependent on different types of land use. Such types include land for non-intensive crop production and livestock grazing as well as arable-fields, pastures and woodland that have been abandoned recently or in the past. We formulate a spatially explicit model of habitat suitability for the species according to the current pattern of land-use types. Then we evaluate past habitat suitability, by applying our model to three known past scenarios (dating back to the 1950s), and simulate the development of vegetation communities after complete land abandonment, i.e. the cessation of all agricultural and pastoral activities. We simulate the progressive substitution of fields and pastures by shrubland and tree communities, setting the time and order of habitat succession on the basis of the historical evidence of land-use changes. By estimating past habitat suitability and by simulating the future evolution of land-use cover in an abandonment scenario (quite likely to occur in future decades), we further test the effect of land abandonment on shrike occurrence. Land abandonment has been the major force driving landscape changes in recent years in the study area (Bogliani et al., 2003; see also Fig. 1) and in other mountainous regions.

The analysis of possible future scenarios, calibrated using historical data, allowed us to predict the effects of land abandonment on this declining species. Red-backed shrike could also be used as a model for entire biological communities of low-intensity agricultural areas, at least in northern Italy, where it is an indicator of high biodiversity (Brambilla et al., 2009a, Casale and Brambilla, 2009). Since land-use changes (like the ones we consider here), represent a common phenomenon in most of southern European and Mediterranean mountainous areas, we believe that our study species/case may have a more general interest, and may often be applicable to the species-rich (and conservation priority) biological communities of traditionally farmed landscapes.