• Open Access

Biodiversity impacts of bioenergy crop production: a state-of-the-art review

Authors

  • DESIRÉE J. Immerzeel,

    Corresponding author
    1. Group Energy & Resources, Faculty of Geosciences, Copernicus Institute of Sustainable Development, Utrecht University, Utrecht, The Netherlands
    Search for more papers by this author
  • PITA A. Verweij,

    1. Group Energy & Resources, Faculty of Geosciences, Copernicus Institute of Sustainable Development, Utrecht University, Utrecht, The Netherlands
    Search for more papers by this author
  • FLOOR van der Hilst,

    1. Group Energy & Resources, Faculty of Geosciences, Copernicus Institute of Sustainable Development, Utrecht University, Utrecht, The Netherlands
    Search for more papers by this author
  • ANDRÉ P. C. Faaij

    1. Group Energy & Resources, Faculty of Geosciences, Copernicus Institute of Sustainable Development, Utrecht University, Utrecht, The Netherlands
    Search for more papers by this author

Abstract

The use of biomass as feedstock for energy is often associated with increased claims on land, competition with food production and impacts on other ecosystem services. Studies on sustainability aspects of bioenergy production often indicate biodiversity as a key concern. This article presents a first comprehensive review of published impacts of bioenergy crop production on biodiversity, evaluates the drivers and pressures of biodiversity change and summarizes current trends and impacts. The review provides insight into the types of biodiversity indicators applied under a range of conditions and the mitigating measures proposed to minimize negative impacts or realize biodiversity benefits. The 53 selected publications give diverse results that are explained by the various spatial scales, production systems and regions, time horizons, methodologies and biodiversity indicators used. Reported impacts depend on initial land use and are mostly negative, especially in tropical regions. The impacts of second generation bioenergy crops tend to be less negative than first generation ones, and are in some cases positive (at the field level), in particular in temperate regions. Land-use change appears as the key driver of biodiversity change, whereas the associated habitat loss, alterations in species richness and abundance are the main impacts addressed. Such changes are often paired with the (initiation of a) process of biological homogenization. The article confirms that concerns about the expansion of bioenergy crop production not only relate to the direct effects on biodiversity by replacing natural vegetation, but increasingly to indirect effects as well. These effects have, however, shown to be difficult to quantify. At the same time, the land sparing vs. land sharing debate receives growing attention, whereas little evidence exists in bioenergy literature on the impacts of large-scale application of these strategies on (agro)biodiversity. Following the findings of the review, the article finally provides recommendations for future research.

Introduction

The use of biomass for energy can replace fossil fuels, and contribute to greenhouse gas mitigation, security of energy supply and rural development. Currently, renewable energy sources contribute 13% to the total primary energy supply, of which almost 80% (50 EJ) is supplied by biomass. Estimates indicate that the global deployment of energy from biomass is expected to increase to 100 – 300 EJ by 2050, a two to sixfold increase compared with current figures. In these estimates, a growing proportion of second generation biofuels is projected (Chum et al., 2011). The production of bioenergy crops can result, however, in increased claims on land, competition with food production and impacts on other ecosystem services. One of the key sustainability concerns related to the growing demand for bioenergy refers to the impacts of large scale biofuel feedstock cultivation on biodiversity (Dornburg et al., 2010; Chum et al., 2011; Wicke et al., 2012). The article follows the formal definition of biodiversity of the Convention on Biological Diversity (CBD) stating that “biological diversity means the variability among living organisms from all sources including, inter alia, terrestrial, marine and other aquatic ecosystems and the ecological complexes of which they are part; this includes diversity within species, between species and of ecosystems.” Biodiversity is fundamental to ecosystem functioning and contributes directly (through provisioning, regulating and cultural ecosystem services) and indirectly (through supporting ecosystem services) to human well-being (CBD, 2010).

Biofuel feedstock cultivation may affect biodiversity through a variety of driving forces including land-use change, overexploitation, pollution, invasive species and climate change (MEA, 2005; Dornburg et al., 2010). Biodiversity gains and losses can both occur as a result of bioenergy crop production depending on, for example, the spatial scale, type of production system, geographical region or time horizon considered. Besides, the effects of biomass production may be either direct (i.e., natural or non-natural habitats being converted into energy crops) or indirect (i.e., natural vegetation being converted into land-use types that are displaced by energy crop production elsewhere) (Smeets et al., 2008; Wicke et al., 2012).

Several national and international studies have been published that evaluate or study the sustainability of bioenergy crop production (e.g., van Dam et al., 2008; Smeets et al., 2008; Dornburg et al., 2010; Beringer et al., 2011; Chum et al., 2011; Walter et al., 2011; Wicke et al., 2012). Publications on the biodiversity effects of bioenergy crop production give diverse results and show that currently no widely accepted methodology is available to quantify these effects (van Dam et al., 2008; Smeets et al., 2008; Dornburg et al., 2010). Some reviews typically focus on biodiversity impacts in a particular region and for specific crop(s) (e.g., Fitzherbert et al., 2008; Dauber et al., 2010; Hartman et al., 2011). To date, a comprehensive review including different geographical regions and crops has not been published. Such a review is, however, considered important to (1) inform decision makers about the biodiversity effects from bioenergy crop production as published in scientific literature, (2) report to sustainability initiatives and certification processes on these effects and provide guidelines for monitoring and (3) develop a research agenda based on identified gaps in knowledge.

This article aims to cover these issues and presents a review of scientific publications on the impacts of biofuel feedstock cultivation on biodiversity. The article evaluates the drivers of biodiversity change associated with bioenergy crop production and summarizes current trends and impacts. In addition, the article provides insight into the types of biodiversity indicators used and the conditions under which these indicators are applied. The review also presents an overview of proposed mitigating measures to minimize negative impacts or realize biodiversity benefits. Following the findings of the review, the article finally provides recommendations for future research in this field of bioenergy science.

Methodology

The literature on biodiversity impacts of bioenergy crop production is subject to a wide range of definitions and methodologies that are difficult to compare. To structure the findings systematically, the review uses the DPSIR framework (Smeets & Weterings, 1999). This allows for a clear description of the relationships between increasing demands for bioenergy, expansion of bioenergy cropland, impacts on biodiversity and (proposed) mitigating measures. Following the DPSIR framework, social and economic developments (Driving Forces, D) exert Pressures (P) on the environment and, as a consequence, the State (S) of the environment changes. This leads to Impacts (I) on ecosystems which may elicit a societal Response (R) that feeds back on Driving Forces, State or Impacts via various mitigation or adaptation actions (Smeets & Weterings, 1999; Maxim et al., 2009). The framework has been widely used for systematic reviews of environmental impact assessments and scientific literature (see Maxim et al., 2009; Haberl et al., 2009; Spangenberg et al., 2009 for some recent examples in the field of biodiversity).

The Web of Science and Science Direct databases were queried for studies published in literature between 2007 and May 2012. An internet search was also performed using the meta-search engine Google scholar. The search terms included combinations of biomass, bioenergy, biodiversity, impacts, effects, benefits and the major biofuel feedstocks (oil-, sugar-, starch- and lignocellulosic crops following Chum et al., 2011). Only impact studies with quantified impacts were selected to allow for evaluation of the types of indicators used and related changes. This implies that the more descriptive or qualitative studies were excluded from the review. To be able to assess changes in biodiversity, studies should include at least one (potential) biofuel crop with a ‘reference’ or initial habitat, consisting of a natural (e.g., primary forest), low-intensity (e.g., pasture) land-use or traditional (food) crop; or a comparison of at least two (potential) biofuel crops. In addition, meta-analyses and published reviews of biodiversity impacts of certain crops (in specific regions) resulting from the search were evaluated. The overall results of these studies were included and their references were scanned for additional publications fitting the above criteria.

Part of the divergence in methodologies and indicators used in literature is caused by the fact that there is no single indicator that captures the full complexity of biodiversity. This complicates aggregation and comparison and hampers quantitative analysis. Therefore biodiversity impacts are categorized into negative (−), positive (+) and neutral or both negative and positive impacts (+/−).

The spectrum of biodiversity impact studies in bioenergy literature

The 53 selected publications with quantified biodiversity effects of bioenergy crop production give diverse results (see Table 1 for a full overview). Figure 1 shows the spectrum of studies including the number publications per spatial scale, geographical region, type of crop and methodologies applied.

Table 1. Published biodiversity impacts of biomass production
ReferenceScaleCountryCropDriverImpactsInitial or reference land useMethod/indicatorResultsStudied taxonomic groups
Eggers et al. (2009)ContinentEUFirst and second generation cropsLand-use changeHabitat loss, changes in species richness and distributionSpecies habitat, based on habitat suitability of land-use classesScenario-based assessment of three biofuel policy options using species-specific data on habitat suitability and habitat sizeMammals, reptiles, amphibians and birds
Hellmann & Verburg (2010)ContinentEUFirst generation cropsLand-use changeHabitat lossSemi natural vegetation, forest and High Nature Value farmlandScenario-based assessment of the EU biofuel directiven.a.
Dauber et al. (2010)ContinentTemperate regions, with emphasis on Europe and North AmericaSecond generation cropsLand-use changeHabitat loss, changes in species richness, abundance, compositionArable land, natural woodland or managed forestsReview of impacts at three scales+/−n.a.
Louette et al. (2010)ContinentEuropeSecond generation woody cropsLand-use changeChanges in species distributionOpen agricultural and abandoned landSensitivity scores using species' ecological characteristics such as optimal habitat type and dispersal capacity and based on expert judgmentMammals, reptiles, amphibians, birds, butterflies, vascular plants, freshwater fish, and aquatic macrobenthos
Nagy et al. (2009)FieldHungaryCorn, oilseed rape, sunflowerLand-use changeChanges in species abundanceTraditional small-scale farming systems, larger single crop fields (cereal, fallow, alfalfa), abandoned areas with bush encroachment, grasslands, forest, waters and settlementsField surveys and statistical analyses of presence data-Birds
Fletcher et al. (2010)FieldUSACorn, Pinus and Populus species, switchgrassLand-use changeChanges in species richness, abundance and compositionNatural (e.g., coniferous forest) or low-intensity (e.g., pasture) land useMeta-analysisVertebrates, focus on birds
Robertson et al. (2011b)FieldUSACorn, switchgrass and mixed-grass prairieLand-use changeChanges in species richness and abundanceComparison between three cropsField surveys and statistical analyses of presence data and landscape variables+/−Birds
Werling et al. (2011)FieldUSACorn, switchgrass and mixed-grass prairieLand-use changeChanges in species richness and abundance; loss of biocontrol servicesComparison between three cropsField surveys and statistical analyses of presence data+/−Beetles
Barlow et al. (2007a)FieldBrazilEucalyptusLand-use changeChanges in species richness, abundance and compositionPrimary and secondary forestField surveys and statistical analyses of presence dataTrees and lianas, birds, amphibians, lizards, beetles, mammals, butterflies, bats, moths, grasshoppers, epigeic arachnics, fruit flies, scavenger flies, orchid bees
Barlow et al. (2007b)FieldBrazilEucalyptusLand-use changeChanges in species richness, abundance and compositionPrimary and secondary forestField surveys and statistical analyses of presence dataBirds
Calviño-Cancela et al. (2012)FieldSpainEucalyptusLand-use changeChanges in habitat and species richnessNative forests and shrublandsField surveys and statistical analyses of presence data (using a.o. Shannon index)Understory plants
Hsu et al. (2010)FieldAustraliaEucalyptusLand-use changesChanges in species richness and abundanceDryland and riparian forests, riparian pasturesField surveys and statistical analyses of presence dataBirds
Loyn et al. (2007)FieldAustraliaEucalyptusLand-use changeChanges in species abundanceNative forest and cleared farmlandField surveys and statistical analyses of presence data+/−Birds
Nájera (2010)FieldWorldwide (tropical and temperate)First and second generation cropsLand-use changeChanges in species richness and abundanceForests and other types of natural vegetation or plantationMeta-analysis+/−Birds
Bellamy et al. (2009)FieldUKMiscanthusLand-use changeChanges in species richness and abundanceConventional crop (wheat)Field surveys and statistical analyses of presence data+Birds
Felten & Emmerling (2011)FieldGermanyMiscanthusLand-use changeChanges in species richness and abundanceCereal, grassland, fallow landsField surveys and statistical analyses of presence data+/−Earthworms
Semere & Slater (2007a)FieldUKMiscanthus and reed-canary grassLand-use changeChanges in species richness and abundanceArable fieldsField surveys and statistical analyses of presence data using a.o. Shannon-Wiener index+Ground flora, small mammal and bird species
Semere & Slater (2007b)FieldUKMiscanthus and reed-canary grassLand-use changeChanges in species richness and abundanceComparison between miscanthus and reed-canary grassField surveys and statistical analyses of presence data using a.o. Shannon-Wiener index+Ground beetles, butterflies and arboreal invertebrates
Sage et al. (2010)FieldUKMiscanthus and SRCLand-use changeChanges in species richness and abundanceArable crops and grasslandField surveys and statistical analyses of presence data+/−Birds
Haughton et al. (2009)FieldUKMiscanthus and SRC (willow)Land-use changeChanges in species abundanceArable field marginsField surveys and statistical analyses of presence data+Butterflies
Barney et al. (2012)FieldUSAMiscanthus and switchgrassInvasive speciesHabitat changeRiparian and dryland vegetationField experiment and statistical analyses+/−n.a.
Azhar et al. (2011)FieldMalaysiaOil palmLand-use changeChanges in species richness and abundancePeat swamp forestField surveys and statistical analyses of presence data (using a.o. Shannon–Wiener index (H) and Simpson index (D))Birds
Brühl & Eltz (2010)FieldMalaysiaOil palmLand-use changeChanges in species richnessNatural forestField surveys and statistical analyses of presence dataAnts
Danielsen et al. (2009)FieldIndonesiaOil palmLand-use changeChanges in species richness, abundance and compositionForestsMeta-analysis of published faunal studies and field sampling of flora speciesFlora species (including a.o. forest trees, lianas, epiphytic orchids, and indigenous palms) and fauna species (including a.o. birds, lizards, and mammals)
Fayle et al. (2010)FieldMalaysiaOil palmLand-use changeChanges in species richness and abundanceLowland dipterocarp rainforestField surveys and statistical analyses of presence dataAnts
Rajaratnam et al. (2007)FieldMalaysiaOil palmLand-use changeChange in species distribution and abundanceSelectively logged primary forest and secondary forestField survey (including a.o. prey distribution and abundance)+/−Mammals
Turner & Foster (2009)FieldMalaysiaOil palmLand-use changeChanges in species abundancePrimary and logged forestField surveys and statistical analyses of presence dataArthropods
Berndt & Hölzel (2012)FieldGermanyOilseed rapeLand-use changeHabitat changePrimary wetland habitat (reed)Field surveys, habitat mapping and statistical analyses of presence data+Birds
Volpato et al. (2010)FieldBrazilPineLand-use changeChanges in species richness and abundanceNatural forest and Araucaria plantationsField surveys and statistical analyses of presence dataBirds
Codesido et al. (2011)FieldArgentinaSoyLand-use changeChanges in species richnessNatural grassland and wetlandField surveys and data comparison with previous study in the regionBirds
Schrag et al. (2009)FieldArgentinaSoyLand-use changeChanges in species richness and abundanceNative woodlandsField surveys and statistical analyses of presence data and landscape variablesBirds
Riffel et al. (2011)FieldNorth AmericaSRCLand-use changeChanges in species richness, abundance and compositionWoodlands

Meta-analysis of existing studies to assess potential bird and small mammal response at the stand level to conversion of existing, intensively managed

forests to SRWCs

+/−Birds and mammals
Baum et al. (2012)FieldSweden and GermanySRC (willow and poplar)Land-use changeChanges in species richnessArable land and grasslandsField surveys and statistical analyses of presence data+Vascular plants
Dhondt et al. (2007)FieldUSASRC (willow and poplar)Land-use changeChanges in species richness and breeding successThirty-nine species regularly used in SRWC plantations, shrublands and successional habitats (census data)Field surveys and statistical analyses of presence data+/−Birds
Schulz et al. (2009)FieldEurope and North AmericaSRC (willow and poplar)Land-use changeChanges in species richness and abundanceForest ecosystems, agricultural cropland, SRC cropsReview+/−Birds and ground beetles
Rowe et al. (2011)FieldUKSRC (willow)Land-use changeChanges in species richness and abundanceArable crops and set-aside landField surveys and statistical analyses of presence data+Ground flora and winged invertebrates
Robertson et al. (2011a)FieldUSASwitchgrass and mixed-grass prairieLand-use changeChanges in species richness and abundanceComparison between switchgrass and mixed-grass prairieField surveys and statistical analyses of presence data and landscape variables, and predictions based on established bird-habitat relationships+/−Birds
Woodcock et al. (2010)FieldUKWinter wheat and oilseed rapeLand-use changeChanges in species richness and loss of biocontrol servicesAgri-environment scheme field marginsField surveys and statistical analyses of presence data and landscape variables using a.o. Shannon–Wiener landscape habitat diversity indexBeetles
Felton et al. (2010)GlobalGlobalForest plantationsLand-use changeChanges in species richness and abundancePastureMeta-analysis of differences between timber plantations and pasture lands+/−Plants, invertebrates, reptiles/amphibians, mammals, and birds
Beringer et al. (2011)GlobalGlobalSecond generation cropsLand-use changeHabitat lossNatural grasslands, shrublands, forest and semi-natural vegetation

Scenario-based assessment of the global bio-energy potential using various pristine wilderness areas and areas with exceptional

concentrations of biodiversity

n.a.
Koh & Wilcove (2008)NationalMalaysia and IndonesiaOil palmLand-use changeHabitat loss an changes in species richnessPrimary and secondary forestsAnalysis of land-cover data and comparison of species richness in different land-use typesBirds and butterflies
Gevers et al. (2011)RegionGermanyCornLand-use changeChanges in species richness and abundanceCash cropsScenario-based assessment and simulated species responses+/−Birds, mammals and insects
Landis et al. (2008)RegionUSACornLand-use changeLoss of biocontrol services(diversified) agricultural landscapes

Estimated value of natural

biological control and the economic impacts of reduced biocontrol using (a.o.) a biocontrol services index (BSI)

Arthropods
Meehan et al. (2010)RegionUSACorn & soy; mixed grasses & forbsLand-use changeChanges in species richnessMarginal landScenario-based assessment (landscape-scale bird species richness modelled as a function of land cover)+/−Birds
Brockerhoff et al. (2008)RegionWorldwide (tropical and temperate)Forest plantationsLand-use changeHabitat lossNatural forestReview+/−n.a.
van der Hilst et al. (2012)RegionNetherlandsMiscanthus and sugarbeetLand-use changeChanges in mean species abundanceArable land, pasture, HNV areasScenario based-assessment using HNV and MSAn.a.
Abdullah & Nakagoshi (2007)RegionMalaysiaOil palmLand-use changeFragmentationForest and wetlandFragmentation indexn.a.
Buchanan et al. (2008)RegionPapua New GuineaOil palmLand-use changeHabitat loss, changes in conservation statusNatural forestDeforestation estimates, measured through satellite imagery, combined with an assessment of conservation status based on IUCN Red List criteriaBirds
Fitzherbert et al. (2008)RegionIndonesia and MalaysiaOil palmLand-use changeHabitat loss, changes in species richness, abundance, compositionPrimary forestsReview of literature comparing biodiversity in oil palm with forestBirds
van Dam et al. (2009)RegionArgentinaSoy and switchgrassLand-use changeChanges in species abundanceAbandoned cropland, degraded and non-degraded grasslandScenario-based assessment using MSA+/−n.a.
Kessler et al. (2007)RegionArgentina and Brazil; Indonesia and MalaysiaSoy; oil palmLand-use changeHabitat loss, changes in species abundanceAgricultural lands and natural ecosystemsFive indicators used to assess biodiversity impacts, all based on the NCI (or mean species abundance)n.a.
Stoms et al. (2012)RegionUSASugar beet, bermudagrass, and canolaLand-use changeHabitat lossAgricultural landScenario-based assessment based on the suitability of agricultural landscapes for wildlife species and using a.o. conservation planning tool Marxan+/−Birds, amphibians, reptiles, and mammals
Hartman et al. (2011)RegionUSASwitchgrassLand-use change, invasive speciesHabitat changeMarginal and CRP lands (Conservation Reserve Program)Review of literature on the ecological implications of changing use of marginal and CRP land+/−Birds and insects
Figure 1.

The spectrum of biodiversity impact studies in bioenergy literature. *More studies focus on second generation crops than first generation crops (26 vs. 17 respectively), whereas 10 studies analyze both. Other feedstocks addressed include oilseed rape, sugarbeet, sunflower, wheat, prairie, reed canary grass and other grasses.

Drivers and pressures

In the past 50 years, the most important direct driver of biodiversity change in terrestrial ecosystems has been land use and land cover change (MEA, 2005; CBD, 2010; WWF, 2012). Not surprisingly, almost all research on biodiversity impacts of bioenergy crop production focuses on land-use change. The associated habitat loss and fragmentation are the major pressures on biodiversity. Various studies show that concerns about the expansion of biomass production not only relate to the direct effects from converting natural or non-natural habitats into bioenergy crop plantations, but also to the indirect land-use change effects of biofuel feedstock cultivation (e.g., Kessler et al., 2007; Eggers et al., 2009; Hellmann & Verburg, 2010). Indirect land-use changes (iLUC) can occur when a) direct displacement of pasture or cropland results in livestock or crops being produced elsewhere to continue to meet demand and b) the diversion of crops to other uses triggers higher crop prices, which results in more land being taken into agricultural production elsewhere (Wicke et al., 2012). Although literature estimating iLUC-induced GHG emissions is growing (e.g., Dornburg et al., 2010; Chum et al., 2011; Wicke et al., 2012), quantification of the indirect effects of biomass production on biodiversity is difficult and has been attempted by few of the selected studies. For example, Hellmann & Verburg (2010) indicate on a European level that the area of semi natural vegetation, forest and High Nature Value (HNV) farmland converted directly is small in all scenarios variants related to the biofuel directive. The indirect effects of the directive on European land use and biodiversity are, however, much larger. The area of semi natural vegetation is found to be 3–8% smaller in scenarios with the directive as compared with scenarios without the directive. This is due to replacement of these areas by grassland or arable land compensating for the cultivation of bioenergy crops in other areas. Kessler et al. (2007) calculated the ‘multiplier effect’ in various tropical countries from commodity development until 2005, which means the additional land-use change as a collateral effect outside the actual cropping areas. They expressed this effect in terms of decrease in Natural Capital Index (an indicator for changes in original species composition). The highest multiplier effects occur for soy production in Brazil (87% decrease in NCI caused by soy production) and for palm oil in Indonesia (35% decrease in NCI caused by palm oil production). Figure 2 shows the mechanisms involved in the multiplier effect for the case of Brazil. Although such results are paired with large uncertainties, they give an indication of iLUC expressed in terms of an established biodiversity indicator.

Figure 2.

An example of the multiplier effect as analyzed by Kessler et al. (2007). The Figure is based on Verweij et al. (2009) and schematically shows the complex interactions between agriculture expansion and cattle ranching in the agricultural frontier region between the Amazon and Cerrado biomes in Brazil. Historically, the dominant pattern of forest conversion begun with small-scale exploration for timber or subsistence agriculture, followed by consolidation into large scale cattle ranching operations or abandonment to secondary forest (Morton et al., 2006). The expansion of large-scale mechanized agriculture (mainly soy and increasingly sugarcane) has introduced a new pathway for loss of natural vegetation, generating debate on the contribution of (bioenergy) cropland expansion to deforestation dynamics (Verweij et al., 2009; Sparovek et al., 2012).

Two studies address the introduction of perennial bioenergy crops as a driver of change in biodiversity due to invasive traits (Hartman et al., 2011; Barney et al., 2012). The crops are being cultivated from non-native taxa to have few resident pests, to tolerate poor growing conditions and to produce competitive bioenergy crops (Barney & Ditomaso, 2008). The invasive potential of most bioenergy crops and the susceptibility to the invasion of certain habitat types to bioenergy crop species are, however, largely unknown (Barney et al., 2012). Only one field-based assessment by Barney et al. (2012) aimed at analyzing habitat susceptibility and resistance to invasion by perennial grasses in potential receiving environments.

An important driver that is not addressed as such by the selected studies is the intensification of production systems. Intensification is one of the focal points in the bioenergy sector to reduce the competition between food, feed and fuels, and therefore iLUC (Dornburg et al., 2010; Wicke et al., 2012). The amount of bioenergy that can be produced without expanding the total agricultural land-use area depends on the rate of intensification of the agricultural sector and the suitability of the land that becomes available for energy crop production (van der Hilst et al., Forthcoming). Growing attention goes therefore to production of dedicated bioenergy crops on so-called surplus land (two different origins of surplus land are distinguished by Dauber et al. (2012): (1) land currently not in use for the production of food, animal feed, fibre or other renewable resources due to poor soil fertility or abiotic stress and (2) land currently no longer needed for food and feed production because of intensification and rationalization of production, resulting in yield increases and thus a reduced requirement for land), whereas the production of food is continued on established agricultural land (Dauber et al., 2012). This might reduce pressure on sensitive habitats. However, current estimates of availability and suitability of surplus land for development of energy crop production or use of existing biomass potentials are uncertain.

Although the consequences of such developments on biodiversity are only discussed to a limited extent in the selected studies, intensification of agricultural production receives considerable attention in biodiversity science (outside bioenergy literature). Besides, the current land sparing vs. land sharing debate is growing as well (e.g., Fischer et al., 2008; Phalan et al., 2011; Anderson-Teixeira et al., 2012; Tscharntke et al., 2012). Land sparing is the segregation of land for nature and for production and land sharing (or wildlife-friendly farming) is the integration of both objectives on the same land (Tscharntke et al., 2012).

Although intensification of agricultural production systems and the use of surplus land might alleviate pressure on natural ecosystems, it can increase pressure on (agro)biodiversity through a growing share of high-yielding monocultures in the landscape and pollution resulting from increased use of fertilizers and pesticides (Sala et al., 2009). This article follows Jackson et al. (2007) stating that “Agrobiodiversity refers broadly to the biota, e.g., soil microbes and fauna, weeds, herbivores, carnivores, etc., colonizing the agroecosystem and surviving according to the local management and environment.” Contrary, positive changes can be expected when an intensive land use is replaced by a less intensive one. Introduction of perennial habitats into intensive agricultural systems, also through, for example, agroforestry systems such as alley cropping of SRC trees, might lead to biodiversity benefits such as gains in habitat and higher connectivity (e.g., Porter et al., 2009; Dauber et al., 2010; Holzmeuller & Jose, 2012).

Results in this debate are still contradictory. Authors increasingly argue, however, that so far the land sparing approach has failed to take into account the multiple ecosystem services provided by extensively managed agricultural landscapes (Porter et al., 2009; Tscharntke et al., 2012). This shows that in the field of bioenergy, major scientific challenges remain regarding agricultural intensification, suitability of (surplus) land and the associated pressures on biodiversity.

A long-term driver that has not been assessed by the selected studies includes climate change effects from bioenergy use. In the long term, bioenergy crop production is expected to contribute to a reduction in GHG emissions and it could therefore lead to a reduction in the effects of climate change on biodiversity (Brink et al., 2007; CBD, 2010; Dornburg et al., 2010). The balance of climate change effects was analyzed by CBD (2010) in a global scenario study in which the target of mitigating climate change by limiting CO2eq concentration to 450 ppm is met by using mainly woody biomass for large scale bioenergy production. In this scenario, the ‘positive’ effect is that biodiversity, as indicated by the MSA, would decrease by 10% between 2000 and 2050, compared with 11% in the baseline scenario.

Drivers and pressures summarized

Figure 3 presents an overview of the relationships between drivers and pressures of biodiversity changes discussed above. Habitat loss and fragmentation are presented here as pressures on biodiversity resulting from land-use change (also following the formal definition of the Millennium Ecosystem Assessment (MEA) (MEA, 2005)). Most studies use habitat loss (and in a few cases fragmentation), however, as an indicator for biodiversity impacts. Reported results on habitat loss and fragmentation are therefore further discussed in the impacts section below.

Figure 3.

Relationships between drivers and pressures of biodiversity change resulting from bioenergy crop production. Negative influence: image Positive influence: image.

Impacts

In this section, categorized impacts are summarized and for the main indicators, trends and key examples per indicator are described. Reported impacts are mostly negative (26 studies) but particularly depend on spatial scale, geographical region, type of bioenergy crop and initial land use.

Different impacts per spatial scale and geographical region

Figure 4 shows the biodiversity impacts on different spatial scales. On each scale, the majority of the impacts are negative. Positive impacts are only reported at the field scale. It is, however, difficult to draw general conclusions here, as the number of studies is strongly biased towards the field scale. This might indicate that benefits at the larger geographical scales are more difficult to obtain, but more research at these scales is required to underpin such conclusions. This could particularly be the case for second generation crops in temperate regions, where the acreage of those crops is still too small to identify coarse scale impacts on biodiversity.

Figure 4.

Biodiversity impacts on different spatial scales. One national study is included under region in the figure. Regional scale is defined here as a region within a country.

As shown in Fig. 5, the expansion of bioenergy crops in the tropics raises most concerns with 88% of studies reporting negative impacts, especially for first generation crops (mainly oil palm and soy).

Figure 5.

Biodiversity impacts per production region.

The importance of initial land use

The majority of negative impacts refer to the conversion of natural vegetation to first generation crops, in particular in the tropics. Koh & Wilcove (2008) report, for example, that between 1995 and 2005, 55–59% of oil palm expansion in Malaysia and at least 56% in Indonesia occurred at the expense of primary forest. As indicated by Gasparatos et al. (2011), reviews on the biodiversity impacts from oil palm expansion have shown that oil palm plantations contain much fewer species than primary forests (e.g., Fitzherbert et al., 2008; Danielsen et al., 2009). Besides, the subsequent communities were dominated by a few generalist species of low-conservation value (Danielsen et al., 2009). The studies suggest that biodiversity loss in oil palm plantations occurs because plantations are structurally less complex than primary forests, have a shorter life span and are major landscape fragmentation factors (Fitzherbert et al., 2008; Gasparatos et al., 2011).

The evidence of negative impacts not only refers to the conversion of primary or secondary forests but also to the conversion of other natural ecosystems including grasslands, wetlands or other areas of high biodiversity value. However, when abandoned cropland, degraded or marginal lands are converted, biodiversity benefits are reported in both tropical and temperate regions, especially in relation to second generation crops. Caution is, however, needed when making general assumptions about the (potential) biodiversity benefits of using these broadly defined land uses (Felton et al., 2010; Dauber et al., 2012). A clear distinction between degraded, marginal or low-intensity pasture or grassland is often lacking, whereas at the same time such areas might hold important biodiversity values. Various studies have shown that conversion of such lands into plantations of perennial crops may therefore also result in negative impacts on biodiversity (e.g., Felton et al., 2010; Fletcher et al., 2010; Hsu et al., 2010; van der Hilst et al., 2012).

Different crops have different impacts

The impacts of second generation crops tend to be less negative than first generation ones, and are in several cases positive (at the field scale). About 87% of the studies on first generation crops report negative impacts, vs. 27% for second generation crops. Positive effects are shown in 32% of these studies, and are especially true for grasses and SRC crops in Europe. The reported impacts per bioenergy crop are presented in Fig. 6.

Figure 6.

Biodiversity impacts per bioenergy crop.

A total of 20 studies report both positive and negative (or neutral) impacts, the majority for second generation crops. In some cases (five), combined analyses of first and second generation crops showed negative impacts on the first, and positive effects on the latter. Perennial crops have the potential to provide habitat or shelter for specific species (such as migratory birds) and improve connectivity or support restoration of degraded or marginal land, resulting in biodiversity benefits (Dauber et al., 2010; Meehan et al., 2010; Robertson et al., 2011a). These studies are, however, cautious in presenting positive results on second generation crops and indicate that biodiversity benefits depend on (1) different species responses within and between taxonomic groups, (2) different spatial scales at which studies are conducted, and (3) management, age, size and heterogeneity of plantations. This also relates to the increased vulnerability of an area to new invasive species when landscape heterogeneity is reduced (Hartman et al., 2011). Similar conclusions are drawn by other reviews on forest plantations (e.g., Brockerhoff et al., 2008; Schulz et al., 2009). An additional notion here is that except for one scenario-based assessment in Argentina (van Dam et al., 2009), none of the studies evaluate the (potential) impacts from second generation feedstock cultivation in tropical regions.

The large share of studies focused on second generation crops in temperate regions and oil palm in Indonesia and Malaysia, shows important gaps in knowledge particularly in tropical regions. Although a significant proportion of first-generation biofuel feedstocks are produced in the tropics and the demand for these crops are likely to continue at least in the next decade, quantified results are largely biased towards oil palm in South East Asia (Lee et al., 2011). Continued production of bioenergy crops in the tropics is expected to (further) threaten the high concentrations of globally endemic species in these areas, which are located within the world's key biodiversity hotspots such as the Brazilian Cerrado (Myers et al., 2000; Klink & Machado, 2005; Koh & Ghazoul, 2008; Sano et al., 2010). These regions are still relatively unexplored in bioenergy science, whereas uncertainties are likely to be larger (due to, for example, higher levels of endemism as well as lack of data).

Evidence of quantified impacts using different biodiversity indicators

No single indicator covers the totality of the various impacts on biodiversity; consequently the evidence of quantified biodiversity impacts in this review is based on different biodiversity indicators. The main indicators used in the selected studies are (1) change or loss of habitat, ecosystems or other types of valuable nature areas, (2) High Nature Value areas (HNV), (3) Mean Species Abundance (MSA) and (4) indices for species abundance, richness and distribution. Definitions of these indicators are presented in Appendix S1. Table 2 explains the main trends per biodiversity indicator, and indicates key example studies. The results are shown following the quantified effects of land-use change, being the focus of the majority of the studies.

Table 2. Main trends per biodiversity indicator following land-use change as quantified by the selected studies
Main driverIndicatorsTrends 1st gen. cropsTrends 2nd gen. cropsNotesKey example studies
  1. a

    Most studies focus on changes in species abundance, richness and/or distribution in their assessments, generally at the field scale. Smaller spatial scales are often required to assess the impacts on individual species or species groups. A clear comparison between these studies is difficult, because the majority of the studies lacks a clear definition of these indicators and use different methodologies (i.e., various survey techniques and statistical analyses).

  2. b

    This may lead to so-called homogenization of biodiversity which is defined in the MEA as “the process whereby species assemblages become increasingly dominated by a small number of widespread species, representing further losses in biodiversity” (MEA, 2005).

  3. Negative: image Moderate negative: image Positive: image Moderate positive: image Positive or negative: image.

Land-use changeChange or loss of habitat, ecosystems or other types of valuable nature areas image image Overall results show a decrease in habitat, ecosystems or other types of valuable nature areas. Some studies indicate habitat benefits in scenarios with second generation crops. Such impacts vary spatially and often mask wide variations in individual species' responsesBeringer et al. (2011); Eggers et al. (2009); Hellmann & Verburg (2010); Koh & Wilcove (2008); Buchanan et al. (2008); Stoms et al. (2012)
High Nature Value areas (HNV) image image Direct impacts are relatively small in all scenario variants. The indirect effects are however much larger than the direct effects. HNV areas where biodiversity is at risk when converted to first generation crops are more extensive than for conversion to perennialsHellmann & Verburg (2010); van der Hilst et al. (2012)
Mean Species Abundance (MSA) image image Most land-use change scenarios show trends of decreasing MSA values in both tropical and temperate regions when first generation crops replace other land uses. In temperate regions, positive MSA values are reported when second generation crops are projectedvan der Hilst et al. (2012); van Dam et al. (2009); Kessler et al. (2007)
Species abundance & compositiona image image In general, abundance tends to decrease with an increase in biofuel crop area, especially with regards to first generation crops (in both tropical and temperate regions). Various studies report a subsequent increase of generalist species and decrease of species with restricted ranges, specific habitat requirements and/or conservation concern. bSuch changes in species composition may affect the provision of ecosystem services by certain species such as crop pollination and biological control. Although literature on the effects of mass-flowering agricultural crops including potential bioenergy crops such as oilseed rape is significant, such studies did not result from the selection and were therefore not includedFitzherbert et al. (2008); Fletcher et al. (2010); Barlow et al. (2007b); Hsu et al. (2010); Azhar et al. (2011); Danielsen et al. (2009); Volpato et al. (2010); Codesido et al. (2011); Aratrakorn et al. (2006); Landis et al. (2008); Woodcock et al. (2010); Gevers et al. (2011)
Species richness image image Birds are clearly the most studied taxonomic group (31 studies included birds). Trends of decreasing bird species richness when native habitat is converted to bioenergy crops are reported by most studies. Some studies indicate that these effects are larger for first generation than for second generation crops, and in some cases positive. Such benefits are however likely to diminish with age of the crop and when crop management intensifies and depend on landscape structureFitzherbert et al. (2008); Barlow et al. (2007b); Azhar et al. (2011); Danielsen et al. (2009); Volpato et al. (2010); Schrag et al. (2009); Aratrakorn et al. (2006); Fletcher et al. (2010); Loyn et al. (2007); Riffel et al. (2011); Schulz et al. (2009); Bellamy et al. (2009); Meehan et al. (2010)
 Species distribution image image Actually few studies consider the impacts of bioenergy cropland expansion on the distribution or distributional ranges of species. Habitat suitability or species'sensitivity is, however, evaluated using species'ditributional ranges. Overall results are negative, but tend to be less in case of second generation crops. Buchanan et al. (2008) estimate bird species altitudinal ranges to assess changes in the conservation status of endemic avifauna following the IUCN Red List criteria. They show increasing numbers of threatened or near threatened species following forest lossEggers et al. (2009); Stoms et al. (2012); Louette et al. (2010); Buchanan et al. (2008)

Conditions under which biodiversity indicators are applied

Table 3 summarizes the conditions under which the main biodiversity indicators are applied in the selected publications including spatial scale, biodiversity level, time horizon and data requirements.

Table 3. Summary biodiversity indicators used in selected publications
 Spatial scaleBiodiversity levelTime horizonData readily availableTypes of data required
LocalRegionContinentGlobalWithin speciesBetween speciesEcosystemsEmpiricalScenario-basedYesNoLand use/cover mapsLand-use specific MSA valuesHabitat preferencesSpecies presence dataSpecies Red List status
Habitat, ecosystems or other types of valuable nature areasXXXX  XXXX X    
HNV XX X XXX XX XXX
MSA XXX X XXX XX   
Fragmentation indicesXX    XXXX X    
Species richnessXX  X  X  X   X 
Species abundanceXX  X  X  X   X 
Species distributionXXX X  XX XX XX 
Conservation statusXXXXX  X  X   XX

The extent of habitat, ecosystems or other types of valuable nature areas is a widely used indicator for biodiversity on each spatial scale, especially in areas outside agricultural lands. Data are readily available (often from satellite images) and can be applied in both empirical and scenario-based assessments. Data are frequently on a coarse scale and would therefore require field verification to improve the accuracy of mapping exercises. Assessments of the impacts on individual species or species groups are difficult solemnly using this type of indicator. Indicators that can be applied to include changes within or between species (groups) at larger spatial scales are the HNV and MSA. The HNV indicator is currently only available for Europe and would require large amounts of data in other regions. This makes it difficult to apply this indicator in (scenario-based) assessments outside Europe. Besides, there are still important challenges in using the HNV indicator in Europe especially due to differences and limitations in data and methodologies (Peppiette, 2011).

The MSA is a less precise, but relatively easily applicable indicator that can provide an indication of (potential) changes in the relative abundance of species, also in future scenarios. It is, however, difficult to apply the indicator at smaller spatial scales because of the low level of detail and considerable uncertainties related to the cause–effect relationships, the drivers considered and underlying data determining the MSA values. Assessments of species distribution can provide more species-specific information at larger spatial scales and can be applied together with indices for species richness and abundance in impact assessments at smaller spatial scales. These indicators allow for analysis of changes in individual species or species groups in both natural areas and agricultural landscapes and provide more accurate information at the species level. The use of these indicators is, however, data intensive and often requires long-term field surveys. Besides, they are difficult to apply in scenario-based assessments as changes in these indicators are hard to project.

Fragmentation indices can add an extra dimension to the above indicators and can provide insight into the effects of fragment size and landscape connectivity on biodiversity, especially if such indices can be related to species-specific indicators at smaller spatial scales. Fragmentation indices can be relatively easily calculated from land use and land cover data. An analysis of the relationship between fragmentation and species-specific indicators requires, however, more data intensive research. Few selected studies have directly referred to or assessed the integrated effects of fragmentation. Fragmentation effects are defined by Fahrig (2003) and include (1) reduction in habitat amount, (2) increase in the number of habitat patches, (3) decrease in sizes of habitat patches and (4) increase in isolation of patches. Added to this can be so-called edge effects, where the area of land at the edge of the habitat patch is altered and less suitable for certain species (MEA, 2005; Verweij et al., 2009). Species that are specialized to particular habitats and those with poor dispersal ability suffer from fragmentation more than generalist species with good dispersal ability. This may affect the extinction risk of the more vulnerable species. Habitat loss and such changes in species composition are addressed by the selected studies, but most authors do not distinguish between habitat loss and habitat fragmentation or consider fragmentation related habitat variables at the field scale, not at the landscape scale [as was also concluded by Fahrig (2003)]. Only one study showed that oil palm is a major contributor to forest fragmentation using a fragmentation index (Abdullah & Nakagoshi, 2007).

Changes in the conservation status of species are also scarcely included in bioenergy literature. At smaller spatial scales, changes in the conservation status are difficult to assess, but at larger spatial scales an increased number of for example vulnerable or (critically) endangered species can provide an important proxy for changes in the state of biodiversity.

Insights from several reviews of biodiversity indicators show that the selected publications exclude several indicators that are often used in biodiversity research (e.g., Reid et al., 1993; Brink, 2000; Dornburg et al., 2008; Butchart et al., 2010; Vačkář et al., 2012). These indicators include, for example, the extent of protected areas, Red List Index and Living Planet Index (LPI). An explanation of these indicators is included in Appendix S1.

Responses

This review shows that bioenergy crop production leads to various tradeoffs on biodiversity. Reconciling biofuel expansion with biodiversity conservation is therefore not easy (Dornburg et al., 2010; Lee et al., 2011). To mitigate the negative impacts of biofuel feedstock cultivation on biodiversity and to promote biodiversity benefits, various measures are proposed and discussed in literature. These mitigating measures are not focused on the bioenergy sector alone, but should be seen in the context of other associated land uses. As also indicated by Wicke et al. (2012), mitigating the impacts of bioenergy crop production, either directly or indirectly, depends on controlling the extent of land-use changes related to biofuel feedstock expansion, including, for example, forestry, agriculture and livestock sectors. Here, these measures are shortly discussed based on the selected publications, and complemented with additional key references.

Better management practices

The implementation of better management practices can contribute to the enhancement of (agro)biodiversity at the field scale as well as landscape heterogeneity. Such practices have been proposed by many and include crop rotations or multiple cropping systems, use of native species and local varieties, enhancement of understory vegetation, no till farming, minimal irrigation and responsible use of pesticides and fertilizers (e.g., Hennenberg et al., 2010; Nájera, 2010; Lee et al., 2011; Wiens et al., 2011; Calviño-Cancela et al., 2012; Wicke et al., 2012). Increased structural diversity and maintenance or creation of landscape elements within plantations can contribute to landscape heterogeneity. Such elements can contribute to connectivity and reduce fragmentation effects and include, for example, stepping stones of natural habitats, buffer zones around vulnerable areas, protection of riparian areas, slopes and other fragile areas (Fletcher et al., 2010; Hennenberg et al., 2010; Hartman et al., 2011). Improvement of structural diversity might also contribute to limiting the risk of invasion by perennial grasses. Finally, various studies propose a precautionary approach when introducing potentially invasive biofuel feedstocks through pre-introduction screening or risk assessments (e.g., Barney & Ditomaso, 2008; Gordon et al., 2011).

Use of marginal, degraded or abandoned agricultural land

As discussed before, the use of marginal, degraded or abandoned is often promoted as an option to minimize competition with food crops and reduce the need for conversion of natural ecosystems (e.g., Fitzherbert et al., 2008; van Dam et al., 2009; Hartman et al., 2011). Feedstock cultivation on such lands may also provide additional environmental benefits such as erosion control and soil carbon sequestration (Hartman et al., 2011; Wicke et al., 2012). Key uncertainties exist, however, in relation to the definition and identification of such lands, potential yields and their biodiversity and socio-economic values (Dauber et al., 2012; Wicke et al., 2012; see previous sections). Dauber et al. (2012) provide a recent scientific background in support of a reassessment of land available for bioenergy production by clarifying the terminology, identifying constraints and options for efficient bioenergy-use of surplus land.

Land-use planning and landscape design

Conservation gains from avoiding or reducing deforestation of natural or high biodiversity value areas, can only be achieved in combination with land-use planning and landscape design (Fitzherbert et al., 2008). As indicated by Wicke et al.(2012), appropriate zoning of land use and land cover is necessary to minimize negative impacts from land-use changes associated to bioenergy cropland expansion. Various authors indicate the importance of maintaining or enhancing heterogeneity in the landscape (e.g., Hsu et al., 2010; Azhar et al., 2011; Hartman et al., 2011; Baum et al., 2012). This can be performed through the retention and restoration of natural forest or other areas of natural vegetation outside plantations (Azhar et al., 2011). Agroforestry or silvopastoral production systems may also act as corridors to connect surrounding fragments, but also to mitigate further encroachment into these areas and reduce edge effects (Lee et al., 2011; Holzmeuller & Jose, 2012; Wicke et al., 2012). Others point towards combined food and energy (CFE) agro-ecosystems that can provide significantly increased ecosystem services compared with conventional agriculture and require less fossil-based inputs (Porter et al., 2009). Besides, biofuel crops have the potential to stimulate landscape heterogeneity themselves by introducing perennial energy crops into landscapes dominated by annual crops (Wiens et al., 2011). Care should, however, be taken here considering the land-use type that is being replaced (see previous sections and e.g., Dauber et al., 2012).

Policy and regulatory measures

As the most serious impacts on biodiversity result from the conversion of natural vegetation to bioenergy crop plantations, major gains in conserving biodiversity can be achieved by implementing measures aimed at the conservation or exclusion of natural habitats, ecosystems or other areas with high-nature values. These can include not only primary forests but also, for example, valuable secondary forests, natural grasslands and peat lands (e.g., Koh & Wilcove, 2008). Such policy and regulatory measures include the establishment of new protected areas, improvement of protected area networks through, for example, buffer zones and ecological corridors as well as the enhancement of regulation and enforcement to limit deforestation. At the same time, these measures might be most challenging, especially in tropical countries where law enforcement is often weak. Besides, the identification of areas with high-nature values still faces methodological challenges (see Appendix S1).

(Voluntary) certification initiatives

Over the years, a wide range of certification initiatives has started aimed at the development of sustainability standards for bioenergy production and trade. The creation of multi stakeholder initiatives such as the Roundtable of Sustainable Biofuels, the Roundtable of Responsible Soy, Bonsucro and the Roundtable of Sustainable Palm Oil have engaged stakeholders throughout the entire supply chain to work towards the production and trade of bioenergy feedstocks following sets of economic, social and environmental criteria (Lee et al., 2011). Research on the effects of certification on biodiversity is still limited, which makes it difficult to draw conclusions on this issue. Besides, various studies indicate that current initiatives still fail to capture iLUC (e.g., van Dam et al., 2010; Hennenberg et al., 2010; Lee et al., 2011). Most bioenergy sustainability criteria deal with biodiversity by proposing process indicators, for example, by referring to national regulations and by excluding protected areas and land identified as HCVA from bioenergy production. However, none of the current systems is built on a coherent and proven set of methodologies that allows for trustworthy verification of criteria and indicators (van Dam et al., 2008). Following the above challenges, various authors argue that certification systems should be complemented by science-based methodologies and additional mitigating measures at multiple spatial scales (e.g., van Dam et al., 2010; Hennenberg et al., 2010).

Mitigating measures not addressed in the selected studies

The selected studies address to a limited extent the importance of good governance in establishing an enabling environment for the implementation of mitigating measures. Policy frameworks at national and international levels (such as the EU Renewable Energy Directive) can contribute to stimulating sustainable bioenergy supply chains, although it is important that policies in various sectors (including e.g., agriculture, environment and rural development) are aligned. Other (financial) incentives to alleviate the pressure from bioenergy crop production on natural ecosystems that have hardly been addressed include the creation of innovative financing mechanisms such as Payments for Environmental Services. To be effective, however, financing schemes must provide an economically (and socially) attractive alternative production system and fulfil several key requirements to prevent negative consequences such as leakage in other areas (Wunder et al., 2008). Finally, partnerships between multiple stakeholders could fill the gap in these and other mitigating measures where there is a lack of regulations, law enforcement or land-use planning policy by governments for example (Visseren-Hamakers & Glasbergen, 2007; Fitzherbert et al., 2008; Lee et al., 2011).

Conclusions and recommendations for future research

This article comprehensively and systematically reviewed publications on the biodiversity impacts of bioenergy crop production. The article evaluated the drivers and pressures of biodiversity change and summarized current trends and impacts. The review provided insight into the types of biodiversity indicators applied under a range of conditions and the mitigating measures proposed to minimize negative impacts or realize biodiversity benefits. The selection of publications is not considered exhaustive. The selection of 53 studies resulted from the applied search method, whereas other combinations of search terms and/or search engines might have resulted in additional studies. Nonetheless, this article presented a state-of-the-art review of biodiversity impacts in literature on bioenergy crop production, identifying important knowledge gaps in this field of science.

The selected publications gave diverse results that were difficult to compare. Clear definitions on drivers, pressures and impacts on biodiversity, and even of biodiversity itself, are often lacking in current literature. The use of the DPSIR framework enabled, however, a structured analysis of the wide ranging publications. The differences were explained by the various spatial scales, production systems and regions, time horizons, methodologies and biodiversity indicators used. Most studies focus on the field level. Reported impacts depend on initial land use and are mostly negative, in particular in tropical regions. The impacts of second generation bioenergy crops tend to be less negative than first generation ones, and are in some cases positive (at the field level), in particular in temperate regions. Land-use change appears as the key driver of biodiversity change, whereas the associated habitat loss, alterations in species richness and abundance are the main impacts addressed. Such changes are often paired with the (initiation of a) process of biological homogenization. The article confirms that concerns about the expansion of bioenergy crop production not only relate to direct land-use change effects on biodiversity by replacing natural vegetation, but increasingly to the indirect effects as well. These effects have, however, shown to be difficult to quantify. At the same time, the land sparing vs. land sharing debate receives growing attention whereas little evidence exists in bioenergy literature on the impacts of large-scale application of these strategies on (agro) biodiversity. Finally, several mitigating measures were identified in bioenergy literature to minimize negative impacts or enhance biodiversity benefits. These measures included better management practices, the use of degraded lands, land-use planning and landscape design, policy and regulatory measures as well as other initiatives promoting a more sustainable bioenergy sector.

This review showed important gaps in knowledge on bioenergy crop production, which are translated here into recommendations for future research. These recommendations are considered essential for the guidance and development of mitigating measures and include the following:

  • Diversification of study regions and target crops to other ‘conversion hotspots' in biodiversity rich tropical ecosystems including, for example, the Cerrado, Atlantic Forest and Chaco in South America, as well as wetlands and other vulnerable arid and semi-arid ecosystems in sub-Saharan Africa.
  • Integration of multiple spatial scales to avoid the current bias towards the field scale and expand the knowledge at larger spatial scales. This requires the use of coarser indicators that can be applied at such spatial scales including, for example, the extent of ecosystems and the MSA, but also indicators that are currently not represented in bioenergy literature such as the Red List Index (RLI). On these coarser spatial scales, this requires modelling approaches on biodiversity responses that are based on accurate projections of future distributions of energy crops. Such projections are, however, challenged by, for example, the time lags of biodiversity responses and low current acreage providing limited baseline information, particularly in case of second generation crops.
  • The complexity of biodiversity requires a combination of complementary indicators on different biodiversity levels. Combinations of various indices are recommended to cover a broader range of the status, trends and changes in biodiversity and the ecosystem services provided by biodiversity (this follows the recommendations of the CBD (CBD, 2010)). For example, the use of habitat, ecosystems or other types of valuable nature areas neglects biodiversity outside these areas. The MSA indicator can then provide complementary information in scenario-based assessments at larger spatial scales. To include more species-specific data, the use of species distribution data can complement the MSA especially on larger spatial scales. To further overcome the limitations of the MSA at smaller spatial scales, indices for species richness and abundance can be applied in impact assessments of changes in biodiversity in both natural areas and agricultural landscapes. The most suitable set of applicable indicators depends, however, on the purpose of the study and data and time availability.
  • Improved insights into changes in species distribution resulting from bioenergy crop production as currently little is known regarding such changes. This can support land-use planning initiatives by identifying, for example, priority areas for conservation that are important in maintaining species' distributional ranges and capacity.
  • Broadening of taxonomic groups to better evaluate differences in responses to bioenergy crop expansion between taxonomic groups. Selecting indicator species (groups) for biodiversity is a challenge, as the findings about the usefulness of certain species groups as indicators are contradictory (Lindenmayer et al., 2000; Dauber et al., 2010). Therefore a broader perspective on how different taxonomic groups are affected can improve the understanding of biodiversity impacts at the species level.
  • Assessments of changes in species extinction risk in response to bioenergy crop expansion can further guide conservation actions and can be an important tool in selecting indicator species for impacts assessments. Depending on the spatial scale, this can include assessments using the RLI or evaluations of changes in the conservation status of particular species.
  • Quantification of pollution effects on biodiversity from the use of fertilizers, pesticides and discharge of waste, especially in aquatic ecosystems. Impact assessments are concentrated on terrestrial ecosystems, partly due to a lack of data on biodiversity in aquatic ecosystems. Increasing concerns have risen on these ecosystems; however, also in the context of bioenergy crop production (where upstream and downstream impacts are often not taken into account).
  • Relating climate change effects from biomass production to biodiversity. One of the main arguments for biomass production for bioenergy is its potential contribution to the reduction in GHG emissions. Global and regional studies indicate possible negative impacts of climate change on biodiversity; therefore insights into the long-term climate change effects from biomass production on biodiversity could assist in minimizing these negative climate change impacts.
  • Studies on the effects of mitigating measures on biodiversity to improve or further guide such measures.
  • This should include specific attention to a better understanding of the effects of intensification of production systems on (agro)biodiversity. When mitigating measures with respect to biodiversity are primarily focused at the conservation of natural areas (as is the basis of most sustainability initiatives), such measures should be complemented with management practices to improve production efficiency and reduce the need for expansion elsewhere. Large uncertainties remain, however, regarding the effects of these ‘land sparing’ options, especially with regards to availability and suitability of land. Assessments of this type of strategy should not only include landscape level effects of intensification on habitat extent, landscape connectivity and fragmentation, but also species-specific indicators such as richness and abundance at the field scale.
  • On the other hand, ‘land sharing’ options via alternative cropping systems such as agroforestry should become research priority. Positive biodiversity impacts from biodiversity friendly cropping systems are presented in literature, but also here large uncertainties remain. A stronger scientific basis in this debate would allow for better (regional) solutions to establish more sustainable bioenergy production systems with optimized production (Dauber et al., 2012).
  • Apart from quantifying impacts from different production systems, detailed assessment and definition of the potential of marginal and degraded land are needed. This calls for a thorough analysis of the state of biodiversity in such areas using more precise and species-specific indicators for species richness, abundance and distribution to capture the full value of these areas and identify under which conditions these areas can be optimally used for bioenergy crop production.
  • Long-term studies are, however, needed to realize most of the above recommendations. Field-based data are often collected within a relatively short time frame covering several years or a decade at the most. Scenario-based assessments project larger time frame of several decades, but have to be based on a wide range of general assumptions. Accurate biodiversity impact assessments require long-term studies to understand the multiple drivers of biodiversity change and their associated impacts.

Acknowledgements

This work was carried out within the BE-Basic R&D Program, which was granted a FES subsidy from the Dutch Ministry of Economic Affairs, Agriculture and Innovation (EL&I) (http://www.be-basic.org).

Ancillary