A Case Study: Farmland Abandonment in Peneda and Its Effects on Macro-moths
Coming back to the European marginal land context, where farmland abandonment is currently perceived as a major threat to the diversity of specialist butterflies from open biotopes and to other open-biotope invertebrates (Marini et al. 2009a; van Swaay et al. 2010), we recently carried out a two-year project (2011–2012) in Portugal to study the response of moths to farmland abandonment. There are a total of 2583 Lepidoptera species currently recognised from Portugal (Corley et al. 2013). We light-trapped macro-moths at 84 fixed sampling sites, each of which was repeatedly sampled six times. Sampling sites were part of a semi-nested sampling design (Proença and Pereira 2013) in three study landscapes that represented a farmland abandonment gradient within the Peneda mountain range (NW-Portugal; Fig. 6.1; elevation 750–1155 m). Here, an on-going rural exodus since the 1950s has led to farmland abandonment and regeneration of native woodland cover, although significant amounts of land, especially close to villages, are still being farmed (Rodrigues 2010). This situation is common to many other rural areas in Europe, and the Peneda area is considered to be representative of other traditional agricultural landscapes in mountainous areas of Southern Europe (Queiroz et al. submitted). We were able to analyse the species composition response to farmland abandonment and forest regeneration of both open-biotope and closed-biotope species across spatial scales, from local up to landscape scales. Nice add-ons to our research were the discovery of a micro-moth species new to science, namely Isotrias penedana, currently only known from the Peneda mountain range (Trematerra 2013; Fig. 6.2a), as well as the discovery of 12 macro-moth species new for Portugal (Corley et al. 2012, 2013; Fig. 6.2b).
Fig. 6.2 a Notable moth discoveries Isotrias penedana, a tortricid (micro-moth) species new to science, discovered during 2012 within the Peneda mountain range (Trematerra 2013) (photo credit: Martin Corley). b Watsonarctia deserta, an arctiid (macro-moth) species added to the Portuguese list in 2012. In Portugal currently only known from four locations, all within the Peneda mountain range (Corley et al. 2013). (Photo credit: Eduardo Marabuto)
In a first analysis of the ca. 23.000 individuals and ca. 380 species rich dataset, we lumped the data from the six sampling rounds for each of the 84 local sites (20 × 20 m, as light traps have relatively small attraction radii: Merckx and Slade 2014). This analysis shows that overall macro-moth species richness is locally richest at woodland sites (mean ± SE: 77.1 ± 4.2), intermediate at meadows (63.2 ± 3.8) and tall shrub (61.7 ± 5.9), and poorest at low shrub sites (51.0 ± 4.3). Not only species richness, but functional diversity too was significantly higher within woodland than in meadows or shrubland for macro-moths (Queiroz et al. submitted). Although this overall pattern among the four main biotopes is basically present in the meadow-dominated, shrub-dominated and forest-dominated landscapes, species richness is locally, at all four biotope types, consistently lower within the shrubdominated landscape than in meadow-dominated and forest-dominated landscapes (Fig. 6.3a). Possible confounding factors, such as altitude, slope and soil fertility, were not analysed. Although they may have played a role in the patterns observed, their role is likely to have been small given the large amount of sites sampled within the same region, and given the limited altitudinal differences between sites.
These observations suggest that local abandonment of relatively species-rich, semi-natural meadows may reduce (−20 %) local richness levels of nocturnal macro-moths, when meadows turn into low shrub biotopes, but that richness levels should locally increase again when these abandoned fields gradually turn into native forest, eventually reaching diversity levels well above those associated with meadows (+ 22 %). The above results may also mean that when the resulting shrubby vegetation from farmland abandonment becomes the dominant land cover within a landscape (for instance due to simultaneous abandonment of large areas within the same landscape and/or due to the arrested development of the resulting low shrub by overgrazing/shepherding and fire), that this landscape effect lowers local species richness levels. This would happen within remaining meadows and woodlands due to fragmentation effects, and also within the scrubland itself due to lower species inputs from different habitats (i.e. reduced spillover effects because of lower heterogeneity). Eventually, such a process may lead to an impoverished species composition of mainly shrub specialists (e.g. the horse chestnut moth Pachycnemia hippocastanaria) and ubiquitous species, and this at both the local and landscape-scale.
Our data also show that this negative landscape-effect of a dominance of shrub vegetation within the surrounding landscape does not only reduce local species richness, affecting all biotopes, but translates to reduced species richness at larger scales too (Fig. 6.3b). By contrast, overall macro-moth richness was highest in the forestdominated landscape and intermediate in the meadow-dominated landscape, at all four spatial scales tested (Fig. 6.3b). Moreover, the difference between the forestdominated and meadow-dominated landscape in absolute species richness steadily increases with spatial scale, from an 11 species difference at the local scale, over a 24 and 32 species difference at intermediate spatial scales, to an excess of 41 species for the forest-dominated landscape at the largest scale tested (i.e. 1280 × 1280 m) (Fig. 6.3b). We interpret these results as follows: extensively farmed landscapes do indeed provide high levels of moth diversity, both locally and at the landscapescale. Still, more has to be gained from landscapes with a high amount of forest
Fig. 6.3 Macro-moth species richness (± SE)—Local sites (20 × 20 m; N = 84) were each sampled six times over 2 years (2011–2012). Data from the sampling rounds were aggregated for each site. a Overall richness at local sites for each of four biotopes (Meadow; Low Shrub; Tall Shrub; Woodland), separately for three landscape types (Meadowdominated; Shrub-dominated; Forest-dominated) (see also Fig. 6.1). Note that in the forest-dominated landscape only one meadow site and two tall shrub sites were sampled, explaining the absent and very large error bars, respectively.
b Overall richness at four spatial scales [20 × 20 m: local site ( N = 84); 80 × 80 m: lumping four sites ( N = 12); 320 × 320 m: lumping seven sites ( N = 12); 1280 × 1280 m:
lumping 28 sites ( N = 3)], separately for three landscape types (Meadow-dominated; Shrub-dominated; Forestdominated) (see also Fig. 6.1).
The absence of error bars at the largest spatial scale is because these three landscape types were each represented by one landscape only.c Richness of closed-biotope (i.e. woodland) versus openbiotope species at five spatial scales (see Fig. 6.3b, but with an additional larger scale lumping all 84 sites). For contrast, a majority of species ( N = 223) that occur in mixed or intermediate biotopes are not retained
cover. Although the species richness difference is not that big at the local scale, the difference becomes larger and more notable with increasing scale, which points to larger beta-diversity levels for forest-dominated than for meadow-dominated landscapes/regions. Indeed, it is well known that semi-natural grasslands are able to reach high local species richness (i.e. alpha diversity), at least for certain taxa (e.g. butterflies: van Swaay 2002; flowering plants: Wilson et al. 2012). However, we show, for macro-moths, that those local diversity levels can be even higher within woodlands, and importantly, that beta-diversity levels may be consistently larger at larger spatial scales within landscapes mainly covered by forests compared to landscapes mainly covered by extensively managed agricultural land. So, although the effects on macro-moth diversity of woodlands and forest-dominated landscapes already compare positively with meadows and meadow-dominated landscapes at the local scale, forest-dominated landscapes outcompete meadow-dominated landscapes more strongly at larger spatial scales. We believe this is an important point since rewilding is to be applied at larger, regional scales, whereas biodiversity has traditionally been measured mainly at local scales alone.
A key result linked to this is the interaction we observed between spatial scale and species' biotope characteristics, whereby the species richness of woodland moths shows a steeper increase with spatial scale than the richness of open-biotope moths (Fig. 6.3c). Whilst we found on average more open-biotope species than typical woodland species locally, this difference disappears at the field-scale, and reverses at the largest scales. Thus, it appears that closed-biotope species have higher beta-diversity than open-biotope species, and that the former are indeed responsible for the higher diversity levels within forest-dominated landscapes.
Finally, these results also demonstrate that although forest-dominated landscapes provide advantages over meadow-dominated landscapes, one needs to take good care not to get stuck into a landscape largely dominated by shrubs alone (Fig. 6.3a, b). Rather, we advise to monitor, and if needed to cater for spatial and temporal habitat heterogeneity within landscapes undergoing farmland abandonment, so as to combat these negative effects.