Show Posts
1
Consequences / Re: Decline in insect populations
« on: January 15, 2020, 12:58:39 PM »Quote
Much of the debate surrounding the human-induced biodiversity crisis has focused on vertebrates3, but population declines and extinctions may be even more substantial in small organisms such as terrestrial arthropods4. Recent studies have reported declines in the biomass of flying insects2, and in the diversity of insect pollinators5,6, butterflies and moths1,7,8,9,10, hemipterans11,12 and beetles7,13,14. Owing to the associated negative effects on food webs15, ecosystem functioning and ecosystem services16, this insect loss has spurred an intense public debate. However, time-series data relating to arthropods are limited, and studies have so far focused on a small range of taxa11,13,14, a few types of land use and habitat12—or even on single sites1,17. In addition, many studies lack species information2 or high temporal resolution2,12. It therefore remains unclear whether reported declines in arthropods are a general phenomenon that is driven by similar mechanisms across land-use types, taxa and functional groups.
The reported declines are suspected to be caused mainly by human land use2. Locally, farming practices can affect arthropods directly by application of insecticides18,19, mowing20 or soil disturbance, or indirectly via changes in plant communities through the application of herbicides or fertilizer21. Forestry practices can also affect local arthropod communities via changes in tree species composition or forest structure22. In addition, local arthropod populations can be affected by land use in the surrounding landscape; for example, through the drift and transport of pesticides and nitrogen by air or water23,24, through the effects of habitat loss on meta-communities (source–sink dynamics25) or by hampering dispersal.
To disentangle the local and landscape-level effects of land use on temporal trends in arthropod communities of grasslands and forests, we used data from the ‘Biodiversity Exploratories’ research programme that pertain to more than 1 million individual arthropods (2,675 species) (Extended Data Table 1). Arthropods were collected annually at 150 grassland sites by standardized sweep-net sampling in June and August from 2008 to 2017, and at 30 forest sites with flight-interception traps over the whole growing period from 2008 to 2016. An additional 110 forest sites were sampled in 2008, 2011 and 2014 to test for trends across a larger number of sites. Both the grassland and the forest sites cover gradients in local land-use intensity. Land-use intensity was quantified in the form of compound indices that are based on grazing, mowing and fertilization intensity in grasslands26, and on recent biomass removal, the proportion of non-natural tree species and deadwood origin in forests27. To analyse landscape-level effects, we quantified the cover of arable fields, grassland and forest in circles, with a radius between 250 m and 2 km, around each sampling site. We modelled temporal trends in arthropod biomass (estimated from body size; Methods), abundance and the number of species separately for grasslands and forests, and tested for the effects of local and landscape-scale land-use intensity on these trends, accounting for weather conditions. Analyses were conducted for all species together, and for different dispersal and trophic guilds.
The total number of arthropod species across all sites (gamma diversity) was substantially lower in later than in earlier years in both forests and grasslands (Fig. 1). Gamma diversity, biomass, abundance and number of species fluctuated over time but revealed an overall decrease with strongest declines from 2008 to 2010, especially in grasslands (Fig. 1). Year-to-year fluctuations in arthropod biomass, abundance and number of species were partially explained by weather conditions (Extended Data Fig. 1, Supplementary Table 1-1, Supplementary Information section 2). Accounting for weather, fitted trends from our models showed declines in biomass of 67% for grasslands and 41% for forests, declines in species numbers of 34% for grasslands and 36% for forests, and declines in abundance of 78% for grasslands, with no significant change in abundances for forests (−17%) (Fig. 1, Supplementary Table 3-1). In grasslands, declines occurred consistently across all trophic guilds (herbivores, myceto-detritivores, omnivores and carnivores), although the trend for carnivores was not significant (Supplementary Table 1-1). In forests, the patterns were more complex: herbivores showed an increase in abundance and species number, whereas all other trophic guilds declined. Temporal trends of arthropods on the basis of data recorded in 3-year intervals from all 140 forest sites were similar to the trends based on the 30 sites with annual data (Supplementary Table 1-1). Sensitivity analyses that removed or reshuffled years showed that the decline was influenced by, but not solely dependent on, high numbers of arthropods in 2008. Fluctuations in numbers (including the numbers from 2008) appear to match trends that have been observed in other studies2, which suggests that the recent decline is part of a longer-term trend that had begun by at least the early 1990s (Extended Data Fig. 2, Supplementary Information section 3). Further sensitivity analyses showed consistent declines when data from individual sampling dates were not aggregated per year, and also showed that declines concerned all three regions that we analysed (Supplementary Tables 3-2, 3-3, Supplementary Fig. 3-1).
Fig. 1: Temporal trends in arthropod communities.
figure1
a–d, Gamma diversity (total number of species across all grassland or forest sites) (a), biomass (b), abundance (c) and number of species (d) of arthropods were recorded in 30 forest and 150 grassland sites across Germany. Gamma diversity shows mean incidence-based, bias-corrected diversity estimates (Chao’s BSS, that is, the higher value of the minimum doubled reference sample size and the maximum reference sample size among years29) for q = 0 and 95% confidence intervals derived from bootstrapping (n = 200). Non-overlapping confidence intervals indicate significant difference30. Box plots show raw data per site and year (n = 1,406 (grassland) or 266 (forest) independent samples). Solid lines indicate significant temporal trends (P < 0.05) based on linear mixed models that included weather conditions, and local and landscape-level land-use intensity as covariates. Shaded areas represent confidence intervals. Boxes represent data within the 25th and 75th percentile, black lines show medians, and whiskers show 1.5× the interquartile range. Data points beyond that range (outliers) are not shown for graphical reasons. Plots for biomass and species number have separate y axes for grassland and forest.
Full size image
Linking changes in biomass, abundance and the number of species to one another enables further inferences regarding the mechanisms that drive arthropod declines. In grasslands, both abundant and less-abundant species declined in abundance (Fig. 2), but loss in the number of species occurred mostly among less-frequent species (Fig. 1, Extended Data Fig. 3, Supplementary Information section 4). This suggests that the decline in the number of species in grasslands was attributable mainly to a loss of individuals among rare species. In forests, species that were initially less abundant decreased in abundance, whereas some of the most abundant species—including invasive species and potential pest species—increased in abundance (Fig. 2, Supplementary Table 5-1). The loss of species was, however, irrespective of their frequency (Fig. 1, Extended Data Fig. 3, Supplementary Information section 4). This suggests that the decline of arthropods in forests is driven by mechanisms that negatively affect the abundances of many species, which leads to an overall decline in biomass and the number of species but favours some species that are able to compensate declines in abundance.
Fig. 2: Changes in the dominance of species.
figure2
Rank abundance curves of arthropod communities for the first two (2008–2009) and final two (2016–2017 for grasslands and 2015–2016 for forests) years of the study, from 150 grassland and 30 forest sites. The insets show enlarged curves for the 30 most-abundant species. Data from the first two and final two study years were pooled (abundances are the total number of individuals of a species observed over two years). Declines in abundance are highlighted by yellow shading, and increases in abundance are shaded in green. The y axes are log-scaled, but show untransformed values.
Full size image
The magnitudes of declines in biomass, abundance and the number of species in arthropod communities were independent of local land-use intensity (Supplementary Table 1-1) as well as changes in plant communities (Supplementary Information section 6) at all sites. However, in forests declines in the number of species were weaker at sites with high natural or anthropogenic tree mortality, possibly owing to increased heterogeneity in local habitats (Extended Data Fig. 4). Landscape composition had no effect on arthropod trends in forests (note that forest sites covered only limited gradients of the landscape variables, Extended Data Fig. 5), but it mediated declines in the number of species in grasslands: the magnitude of the declines increased with increasing cover of arable fields, and marginally increased with cover of grasslands in the surrounding landscape (Fig. 3, Supplementary Table 1-1). This suggests that major drivers of arthropod decline in grasslands are associated with agricultural land use at the landscape scale.
Fig. 3: Landscape effects on arthropod decline in grasslands.
figure3
a, Temporal changes in biomass, abundance and the number of species for all arthropod species. b, c, Temporal change in biomass of species with high (b) or low (c) dispersal ability, conditional on the cover of arable fields in the surrounding landscape (1-km radius). The decline in biomass increased significantly with the cover of arable fields for weak dispersers, but not for strong dispersers. Slopes were derived from models that included weather conditions and local land-use intensity as covariates. The y axes are log-scaled, but show untransformed values.
Full size image
The interaction between a species and the landscape around its habitat depends on its dispersal ability, which ultimately determines its occurrence and persistence28. In grasslands, taxa of high and low dispersal ability (Methods) both declined, but an increasing cover of arable fields—although not of grasslands—in the surroundings amplified declines in the biomass of weak dispersers more strongly than it did declines of strong dispersers (Fig. 3, Supplementary Table 7-1). Weak dispersers may experience higher mortality during dispersal, and thus have a lower chance of (re)colonization of a particular site when arable field cover is high. In forests, strong dispersers declined in biomass, abundance and the number of species, whereas weak dispersers increased in abundance and biomass—but less strongly when grassland cover in the landscape was high (Supplementary Table 7-1). This suggests that the drivers behind arthropod declines in forests also act at landscape-level spatial scales.
We showed that arthropods declined markedly not only in biomass but also in abundance and the number of species, and that this affected taxa of most trophic levels in both grasslands and forests. Declines in gamma diversity suggest that species might disappear across regions. Our results also indicate that the major drivers of arthropod decline in both habitat types act at landscape-level spatial scales, but that declines may be moderated by increases in heterogeneity of local habitats in forests. Although the drivers of arthropod decline in forests remain unclear, in grasslands these drivers are associated with the proportion of agricultural land in the landscape. However, we cannot ascertain whether the observed declines are driven by the legacy effects of historical land-use intensification or by recent agricultural intensification at the landscape level; for example, by the decrease of fallow land and field margins rich in plant species, the increased use of pesticides or use of more potent insecticides (Supplementary Information section 3). Time-series data relating to changes in the use of agrochemicals or the presence of fine-scale arthropod habitats would be necessary to answer this question. Furthermore, the extents to which changes in climate have reinforced the observed trends in arthropod biomass, abundance and number of species is unclear (Supplementary Information section 2). Our results show that widespread arthropod declines have occurred in recent years. Although declines were less pronounced during the second half of our study period, there is no indication that negative trends have been reversed by measures that have been implemented in recent years. This calls for a paradigm shift in land-use policy at national and international levels to counteract species decline in open and forested habitats by implementing measures that are coordinated across landscapes and regions. Such strategies should aim to improve habitat quality for arthropods and to mitigate the negative effects of land-use practices not only at a local scale (within isolated patches embedded in an inhospitable agricultural matrix) but also across large and continuous areas.
3 Likes
And it can cull out invasive species.
).