Recent reports on local extinctions of arthropod species1 and of massive declines in arthropod biomass 2 point to land-use intensification as a major driver of decreasing biodiversity. However, there are no multi-site time-series of arthropod occurrences across land-use intensity gradients to confirm causal relationships. Moreover, it remains unclear which land-use types and arthropod groups are affected and whether the observed declines in biomass and diversity are linked to one another and continue. Here we analyzed arthropod data on more than 1 million individuals and 2,700 species from standardized inventories from 2008 to 2017 at 150 grassland and 140 forest sites in three regions of Germany. Overall gamma diversity in grasslands and forests decreased over time indicating loss of species across sites and regions. In annually sampled grasslands, biomass, abundance and species number of arthropods declined by 67%, 78%, and 34%, respectively. The decline was consistent across trophic levels, mainly affected rare species, and its magnitude was independent of local land-use intensity. However, sites embedded in landscapes with higher cover of agricultural land showed a stronger temporal decline. In 30 forest sites with annual inventories, biomass and species number, but not abundance, decreased by 41% and 36%, respectively. This was supported by analyses of all forest sites sampled in 3year intervals. The decline affected rare and abundant species and trends differed across trophic levels. Our results show that there are widespread declines in arthropods that concern biomass, abundance and diversity across trophic levels. Declines in forests demonstrate that arthropod loss is not restricted to open habitats. Our results 4 suggest that major drivers of arthropod decline act at larger spatial scales, and are, at least for grasslands, associated with agriculture at the landscape level.This implies that land-use relevant policies need to address the landscape scale to mitigate negative effects of land-use practices. Main textMuch of the debate on the human-induced biodiversity crisis has focused on vertebrates 3 , yet population decline and extinctions may be even more substantial in small organisms such as terrestrial arthropods 4 . Recent studies report declines in biomass of flying insects 2 , diversity of insect pollinators 5,6 , butterflies and moths 1,7-10 , hemipterans 11,12 and beetles 7,13,14 . Owing to the associated negative effects on food webs 15 , ecosystem functioning and ecosystem services 16 , the insect loss has spurred an intense public debate. However, time-series data on arthropods are rather limited and studies have so far focused on a small range of taxa 11,13,14 , few land-use and habitat types 12 or even on single sites 1,17 . In addition, many studies lack species information 2 or high temporal resolution 2,12 . Hence, it 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 ...

Machine learning algorithms find frequent application in spatial prediction of biotic and abiotic environmental variables. However, the characteristics of spatial data, especially spatial autocorrelation, are widely ignored. We hypothesize that this is problematic and results in models that can reproduce training data but are unable to make spatial predictions beyond the locations of the training samples. We assume that not only spatial validation strategies but also spatial variable selection is essential for reliable spatial predictions.We introduce two case studies that use remote sensing to predict land cover and the leaf area index for the "Marburg Open Forest", an open research and education site of Marburg University, Germany. We use the machine learning algorithm Random Forests to train models using non-spatial and spatial cross-validation strategies to understand how spatial variable selection affects the predictions.Our findings confirm that spatial cross-validation is essential in preventing overoptimistic model performance. We further show that highly autocorrelated predictors (such as geolocation variables, e.g. latitude, longitude) can lead to considerable overfitting and result in models that can reproduce the training data but fail in making spatial predictions. The problem becomes apparent in the visual assessment of the spatial predictions that show clear artefacts that can be traced back to a misinterpretation of the spatially autocorrelated predictors by the algorithm. Spatial variable selection could automatically detect and remove such variables that lead to overfitting, resulting in reliable spatial prediction patterns and improved statistical spatial model performance.We conclude that in addition to spatial validation, a spatial variable selection must be considered in spatial predictions of ecological data to produce reliable predictions.

The habitat-heterogeneity hypothesis predicts that biodiversity increases with increasing habitat heterogeneity due to greater niche dimensionality. However, recent studies have reported that richness can decrease with high heterogeneity due to stochastic extinctions, creating trade-offs between area and heterogeneity. This suggests that greater complexity in heterogeneity-diversity relationships (HDRs) may exist, with potential for group-specific responses to different facets of heterogeneity that may only be partitioned out by a simultaneous test of HDRs of several species groups and several facets of heterogeneity. Here, we systematically decompose habitat heterogeneity into six major facets on ~500 temperate forest plots across Germany and quantify biodiversity of 12 different species groups, including bats, birds, arthropods, fungi, lichens and plants, representing 2600 species. Heterogeneity in horizontal and vertical forest structure underpinned most HDRs, followed by plant diversity, deadwood and topographic heterogeneity, but the relative importance varied even within the same trophic level. Among significant HDRs, 53% increased monotonically, consistent with the classical habitat-heterogeneity hypothesis, but 21% were humped-shaped, 25% had a monotonically decreasing slope and 1% showed no clear pattern. Overall, we found no evidence of a single generalizable mechanism determining HDR patterns.

Recent progress in remote sensing provides much-needed, large-scale spatio-temporal information on habitat structures important for biodiversity conservation. Here we examine the potential of a newly launched satellite-borne radar system (Sentinel-1) to map the biodiversity of twelve taxa across five temperate forest regions in central Europe. We show that the sensitivity of radar to habitat structure is similar to that of airborne laser scanning (ALS), the current gold standard in the measurement of forest structure. Our models of different facets of biodiversity reveal that radar performs as well as ALS; median R² over twelve taxa by ALS and radar are 0.51 and 0.57 respectively for the first non-metric multidimensional scaling axes representing assemblage composition. We further demonstrate the promising predictive ability of radar-derived data with external validation based on the species composition of birds and saproxylic beetles. Establishing new area-wide biodiversity monitoring by remote sensing will require the coupling of radar data to stratified and standardized collected local species data.

Aim Despite increasing interest in β‐diversity, that is the spatial and temporal turnover of species, the mechanisms underlying species turnover at different spatial scales are not fully understood, although they likely differ among different functional groups. We investigated the relative importance of dispersal limitations and the environmental filtering caused by vegetation for local, multi‐taxa forest communities differing in their dispersal ability, trophic position and body size. Location Temperate forests in five regions across Germany. Methods In the inter‐region analysis, the independent and shared effects of the regional spatial structure (regional species pool), landscape spatial structure (dispersal limitation) and environmental factors on species turnover were quantified with a 1‐ha grain across 11 functional groups in up to 495 plots by variation partitioning. In the intra‐region analysis, the relative importance of three environmental factors related to vegetation (herb and tree layer composition and forest physiognomy) and spatial structure for species turnover was determined. Results In the inter‐region analysis, over half of the explained variation in community composition (23% of the total explained 35%) was explained by the shared effects of several factors, indicative of spatially structured environmental filtering. Among the independent effects, environmental factors were the strongest on average over 11 groups, but the importance of landscape spatial structure increased for less dispersive functional groups. In the intra‐region analysis, the independent effect of plant species composition had a stronger influence on species turnover than forest physiognomy, but the relative importance of the latter increased with increasing trophic position and body size. Main conclusions Our study revealed that the mechanisms structuring assemblage composition are associated with the traits of functional groups. Hence, conservation frameworks targeting biodiversity of multiple groups should cover both environmental and biogeographical gradients. Within regions, forest management can enhance β‐diversity particularly by diversifying tree species composition and forest physiognomy.