Home About Indicator Development Partners Resources

Biodiversity Habitat Index

Key indicator facts

Indicator type

Pressure

Applicable for national use

Yes (find out more)

Indicator classification

Operational and included in the CBD's list of indicators

Indicator type

Pressure

Applicable for national use

Yes (find out more)

Indicator classification

Operational and included in the CBD's list of indicators

Last update

2018

Coverage

Global

Availability

Not freely available

Partners

Csiro logo

Commonwealth Scientific and Industrial Research Organisation (CSIRO)

Contact point

Dr Simon Ferrier - simon.ferrier@csiro.au

Indicator description

The Biodiversity Habitat Index (BHI) has been developed by CSIRO (Australia’s national science agency), working in partnership with GEO BON, GBIF, Map of Life and the PREDICTS project. This indicator is intended to add value to existing assessments of the “rate of loss [and degradation and fragmentation] of all natural habitats, including forests”, under Aichi Target 5, by translating the observed spatial distribution of habitat loss and degradation into expected impacts on retention of terrestrial biodiversity.

This assessment is performed using a fine-scaled grid covering the entire terrestrial surface of the planet. For each cell in this grid an estimate is derived of the proportion of habitat remaining across all cells that are ecologically similar to this cell of interest. Ecological similarity between cells is predicted as a function of abiotic environmental surfaces (describing climate, terrain, and soils) scaled using generalised dissimilarity modelling to reflect observed patterns of spatial turnover in species composition, based on best-available occurrence records for plants, vertebrates and invertebrates globally. The BHI for any given spatial reporting unit (e.g. IPBES region, country) is then derived as a weighted geometric mean of the scores obtained for all cells within that unit, with the contribution of each cell weighted according to its ecological uniqueness. This aggregate score therefore indicates the proportional retention of habitat across finely-mapped environments supporting relatively distinct assemblages of species within a given reporting unit.

The current version of the BHI covers forest biomes only, but the indicator will soon be available across all terrestrial biomes.

Related Aichi Targets

Primary target

5

Target 5:

By 2020, the rate of loss of all natural habitats, including forests, is at least halved and where feasible brought close to zero, and degradation and fragmentation is significantly reduced.

Primary target

5

Target 5:

By 2020, the rate of loss of all natural habitats, including forests, is at least halved and where feasible brought close to zero, and degradation and fragmentation is significantly reduced.

5

Related SDGs

E sdg goals icons individual rgb 15

GOAL 15 - Protect, restore and promote sustainable use of terrestrial ecosystems, sustainably manage forests, combat desertification, and halt and reverse land degradation and halt biodiversity loss.

Target 15.1| Relevant indicator

By 2020, ensure the conservation, restoration and sustainable use of terrestrial and inland freshwater ecosystems and their services, in particular forests, wetlands, mountains and drylands, in line with obligations under international agreements.

Target 15.2| Relevant indicator

By 2020, promote the implementation of sustainable management of all types of forests, halt deforestation, restore degraded forests and substantially increase afforestation and reforestation globally.

Target 15.3| Relevant indicator

By 2030, combat desertification, restore degraded land and soil, including land affected by desertification, drought and floods, and strive to achieve a land degradation-neutral world.

Target 15.4| Relevant indicator

By 2030, ensure the conservation of mountain ecosystems, including their biodiversity, in order to enhance their capacity to provide benefits that are essential for sustainable development.

Target 15.5| Relevant indicator

Take urgent and significant action to reduce the degradation of natural habitats, halt the loss of biodiversity and, by 2020, protect and prevent the extinction of threatened species.

E sdg goals icons individual rgb 15

GOAL 15 - Protect, restore and promote sustainable use of terrestrial ecosystems, sustainably manage forests, combat desertification, and halt and reverse land degradation and halt biodiversity loss.

E sdg goals icons individual rgb 15

Other related MEAs and processes

Indicator icon

IPBES Regional Assessment Chapters

Chapter 3| Official indicator

Status, trends and future dynamics of biodiversity and ecosystems underpinning nature’s benefits to people

Chapter 4| Official indicator

Direct and indirect drivers of change in the context of different perspectives of quality of life

Indicator icon

IPBES Regional Assessment Chapters

Indicator icon

Themes

Bip terrestrial

Terrestrial habitats

View related indicators >
Bip terrestrial

Partners

Csiro logo

Key indicator facts

Indicator type

Pressure

Applicable for national use

Yes (find out more)

Indicator classification

Operational and included in the CBD's list of indicators

Indicator type

Pressure

Applicable for national use

Yes (find out more)

Indicator classification

Operational and included in the CBD's list of indicators

Last update

2018

Coverage

Global

Availability

Not freely available

Indicator description

The Biodiversity Habitat Index (BHI) has been developed by CSIRO (Australia’s national science agency), working in partnership with GEO BON, GBIF, Map of Life and the PREDICTS project. This indicator is intended to add value to existing assessments of the “rate of loss [and degradation and fragmentation] of all natural habitats, including forests”, under Aichi Target 5, by translating the observed spatial distribution of habitat loss and degradation into expected impacts on retention of terrestrial biodiversity.

This assessment is performed using a fine-scaled grid covering the entire terrestrial surface of the planet. For each cell in this grid an estimate is derived of the proportion of habitat remaining across all cells that are ecologically similar to this cell of interest. Ecological similarity between cells is predicted as a function of abiotic environmental surfaces (describing climate, terrain, and soils) scaled using generalised dissimilarity modelling to reflect observed patterns of spatial turnover in species composition, based on best-available occurrence records for plants, vertebrates and invertebrates globally. The BHI for any given spatial reporting unit (e.g. IPBES region, country) is then derived as a weighted geometric mean of the scores obtained for all cells within that unit, with the contribution of each cell weighted according to its ecological uniqueness. This aggregate score therefore indicates the proportional retention of habitat across finely-mapped environments supporting relatively distinct assemblages of species within a given reporting unit.

The current version of the BHI covers forest biomes only, but the indicator will soon be available across all terrestrial biomes.

Contact point

Dr Simon Ferrier - simon.ferrier@csiro.au

Graphs / Diagrams

Figure. Change in global Biodiversity Habitat Index between 2000 and 2014 (averaged across plants, vertebrates and invertebrates), for forest biomes only.

Current storyline

The global trend in the Biodiversity Habitat Index for forest biomes (see Graphs and diagrams) indicates a gradual linear decline from 2000 to 2014. This suggests that finely-mapped environments supporting relatively distinct assemblages of species within broader forest biomes have, on average, a lower proportion of their extent covered by forest in 2014 than they did in 2000.

Indicator relationship to Aichi Target 5

Target 5: “By 2020, the rate of loss of all natural habitats, including forests, is at least halved and where feasible brought close to zero, and degradation and fragmentation is significantly reduced.”

The BHI indicator assesses the “rate of loss [and degradation and fragmentation] of all natural habitats, including forests” by superimposing the spatial distribution of habitat loss and degradation on fine-scaled modelling and mapping of spatial turnover in biodiversity composition, to estimate expected impacts on retention of terrestrial biodiversity.

Data and methodology

Coverage: Global/Sub-global/Regional/National. The current version of the indicator covers forest biomes only, based on WWF’s ecoregionalisation (http://www.worldwildlife.org/biomes) – i.e. Tropical and subtropical moist broadleaf forests; Tropical and subtropical dry broadleaf forests; Tropical and subtropical coniferous forests; Temperate broadleaf and mixed forests; Temperate coniferous forests; and Boreal forests. The indicator is derived from data and models covering the entire global extent of these biomes, as mapped by WWF, at 30-arcsecond (approximately 1km) grid resolution.

Scale: Global data. As above – the indicator covers the entire global extent of forest biomes at 1km grid resolution.

Time series available: 2000, 2005, 2010, 2011, 2012, 2013 and 2014.

Next planned update: 2018. The next update in 2018 will generate indicator results up to 2016 (this update will also include results for 2015, and will cover both forest and non-forest biomes, from 2000 onwards)

Possible disaggregations: The indicator can potentially be reported at any desired level of spatial disaggregation including individual 1km grid-cells, countries, any defined regional classification (e.g. IPBES regions), or the entire planet.

It can also be reported separately for each of three biological groups – plants, invertebrates, vertebrates – or as a combined average across these groups.

Metadata used: The models of ecological similarity used to derive this indicator were fitted using the following input data:

  • Global 30-arcsecond (approximately 1km) gridded environmental surfaces for: Min Monthly Min Temperature, Max Monthly Max Temperature, Max Diurnal Temperature Range, Annual Precipitation, Actual Evaporation, Potential Evaporation, Min Monthly Water Deficit, Max Monthly Water Deficit, Soil pH, Soil Clay Proportion, Soil Silt Proportion, Soil Bulk Density, Soil Depth, Ruggedness Index, Topographic Wetness Index (Sources: http://www.worldclim.org/ https://www.soilgrids.org/ http://www.earthenv.org/ http://www.worldgrids.org/). All temperature, evaporation, and water deficit surfaces were adjusted for the effects of topographic aspect and shading using the approach described by Ferrier et al (2013).
  • Global occurrence records for all terrestrial species within the following taxa: vascular plants, amphibians, reptiles, birds, mammals, ants, bees, beetles, bugs, butterflies, centipedes, dragonflies, flies, grasshoppers, millipedes, snails, moths, spiders, termites, wasps. The records for amphibians, birds and mammals were extracted from data accessible through the Map of Life (https://mol.org/) while records for all other taxa were extracted from data accessible through the Global Biodiversity Information Facility (GBIF http://www.gbif.org/). Records from GBIF underwent basic filtering (for spatial precision) and cleaning (name-matching etc.). The resulting dataset used for model fitting consisted of: 52,489,096 records of 254,145 species of vascular plants; 33,549,534 records of 24,442 species of vertebrates; and 13,244,784 records of 132,761 species of invertebrates.

The input data on habitat condition used in the current version of the indicator are sourced from Hansen et al’s (2013) Landsat-based Global Forest Change dataset (http://earthenginepartners.appspot.com/science-2013-global-forest). This dataset maps annual loss of forest cover from 2000 onwards, at 1-arcsecond grid resolution (approximately 30m at the equator). For any given year the condition of habitat in each 1km cell falling within a forest biome (according to WWF mapping) was calculated as the proportion of 30m cells within that 1km cell covered by forest in that year (according to the Global Forest Change dataset).

Methodology:

Underpinning models of ecological similarity

Generalised dissimilarity modelling (GDM; Ferrier et al 2007) was used to develop a set of statistical models predicting ecological similarity between any pair of 30-arcsecond (approximately 1km) grid-cells, as a function of the abiotic environmental attributes and geographical locations of these cells.

A total of 183 separate GDMs were fitted, one for each possible combination of three broad biological groups – plants, vertebrates and invertebrates – and 61 bio-realms (unique combinations of biomes and biogeographic realms, as per WWF’s ecoregionalisation; http://www.worldwildlife.org/biomes). In a few cases, data from neighbouring or ecologically-related bio-realms were used to supplement the dataset employed in fitting GDMs for more poorly sampled combinations of bio-realm and biological group. To accommodate the ‘presence-only’ nature of much of the assembled biological data, and the diversity of sub-groups encompassed by each of the three broad biological groups, GDMs were fitted to observed matches and mismatches in species identity between pairs of individual occurrence records (drawn from within the same sub-group). The fitted GDMs model spatial turnover in species composition as a function of the environmental variables listed above, the geographical distance between records, and the identity of WWF ecoregions within which they occur.

“Ecological similarity”, as referred to below, equals the predicted proportional overlap in species composition between any given pair of locations (grid cells) – i.e. the mean proportion of species occurring at one of the locations that would be expected to also occur at the other location (in the absence of habitat degradation at both locations). Values for ecological similarity range from 0 (for a pair of cells predicted to have no species in common) through to 1 (for a pair of cells predicted to contain exactly the same species).

Derivation of the indicator

The BHI is derived by combining the above models of ecological similarity, with an appropriate habitat-condition surface for the year of interest. As noted under “Metadata used”, the indicator is initially being derived only for forest biomes, but will soon be generated using condition surfaces covering the entire terrestrial surface of the planet. For each and every cell with condition data an estimate is derived of the average condition of all cells that are ecologically similar to this cell of interest. This is undertaken using the general approach described by Ferrier et al (2004) and Allnutt et al (2008), incorporating enhancements described by Williams et al (2016). This approach is underpinned by the same basic principle as that employed in approaches assessing the proportional loss or conversion (or, conversely, retention) of habitat within discrete ecological classes, such as ecoregions (e.g. Watson et al 2016). When ecoregions are used to assess habitat loss or retention, each location (cell) on the planet is viewed as belonging to a single region, and the proportional retention of habitat in the entire region is assigned to every cell within the region, regardless of whether that particular cell itself contains natural habitat. In other words, all cells within a given ecoregion are mapped in the same colour, indicating the region’s overall level of habitat retention. However, in the approach adopted here, each cell is viewed not as belonging to a homogeneous set of cells forming a discrete region, but rather as sitting within a continuum of ecological variation. The BHI score assigned to a given ‘focal cell’ is therefore calculated as the average condition of all ecologically-similar cells, with the contribution any other cell makes to this calculation weighted according to its predicted level of similarity with the focal cell. If condition is measured on a binary scale (0 = habitat lost, 1 = habitat retained) then the BHI is directly interpretable as a measure of the proportion of habitat retained (across ecologically-similar cells), whereas if condition is measured on a continuous scale between 0 and 1 then the BHI can be interpreted as measuring the ‘effective’ proportion of habitat remaining.

The BHI for any given spatial reporting unit (e.g. IPBES region, country) is then derived as a weighted geometric mean (Buckland et al 2005) of the scores obtained for all cells within that unit, with the contribution of each cell weighted according to its ecological uniqueness – i.e. the inverse of the summed ecological similarity of this cell to all other cells (for further details see Ferrier et al 2004, Allnutt et al 2008, Williams et al 2016). This aggregate score therefore indicates the proportional retention of habitat across the full range of ecological, and therefore biological, diversity occurring within a given reporting unit.

Caveats

It is important to note that the current implementation of this approach, using habitat-condition surfaces derived from Hansen’s Global Forest Change dataset, assumes that all cells within WWF’s mapped forested biomes were originally (naturally) covered by forest. While this will often be a reasonable assumption, exceptions are likely to occur where areas mapped as forest according to WWF’s biome classification contain occurrences of naturally non-forested ecosystems – e.g. patches of montane grasslands at higher elevations within a predominantly forested region. Furthermore, in more open temperate and boreal forest biomes, even an expanse of pristine forest is likely to be mapped as a mix of forest and non-forest pixels, simply because of the wider spacing between trees in such biomes, relative to the fine pixel resolution of the Global Forest Change dataset. Both these phenomena are likely to result in some unknown level of underestimation in BHI scores, particularly for more open forest biomes. Given this caveat we recommend users avoid placing too much emphasis on absolute levels of the BHI, or making comparisons of these levels between different forest biomes, and instead focus more on relative, or proportional, trends in these levels over time.

Future development

CSIRO is nearing completion of a new version of the BHI indicator which will cover the entire terrestrial surface of the planet (including both forested and non-forested biomes). This is employing CSIRO’s recently developed technique for statistically downscaling coarse-resolution global land-use data using finer-resolution covariates (Hoskins et al 2016). CSIRO has extended Hoskins et al’s approach to work with Version 2 of the Land Use Harmonisation product (http://luh.umd.edu/), and with MODIS Vegetation Continuous Fields (http://glcf.umd.edu/data/vcf/) as remote-sensing covariates in place of discrete land-cover classes. The predicted proportions of land-use classes in each 1km cell, for any year of interest, are being translated into ‘habitat condition’ using coefficients fitted in hierarchical mixed-effect modelling of impacts of land use on local biodiversity undertaken by the PREDICTS project (http://www.predicts.org.uk/; Newbold et al 2016).

References

  • Allnutt, T.F., Ferrier, S., Manion, G., Powell, G.V.N., Ricketts, T.H., Fisher, B.L., Harper, G.J., Kremen, C., Labat, J., Lees, D.C., Pearce, T.A., Irwin, M.E., Rakotondrainibe, F. (2008) Quantifying biodiversity loss in Madagascar from a 50-year record of deforestation. Conservation Letters 1: 173-181.
  • Buckland, S.T., Magurran, A.E., Green, R.E. and Fewster, R.M. (2005) Monitoring change in biodiversity through composite indices. Philosophical Transactions of the Royal Society B: Biological Sciences 360: 243-254.
  • Ferrier, S., Harwood, T.D., Williams, K.J. (2013) Assessing refugial potential using compositional-turnover modelling. Pages 51-76 in: Climate change refugia for terrestrial biodiversity: defining areas that promote species persistence and ecosystem resilience in the face of global climate change. National Climate Change Adaptation Research Facility, Gold Coast. https://www.nccarf.edu.au/sites/default/files/attached_files_publications/Reside_2013_Climate_change_refugia_for_terrestrial_biodiversity.pdf
  • Ferrier, S., Manion, G., Elith, J., Richardson, K. (2007) Using generalised dissimilarity modelling to analyse and predict patterns of beta-diversity in regional biodiversity assessment. Diversity and Distributions 13: 252-264.
  • Ferrier, S., Powell, G.V.N., Richardson, K.S., Manion, G., Overton, J.M., Allnutt, T.F., Cameron, S.E., Mantle, K., Burgess, N.D., Faith, D.P., Lamoreux, J.F., Kier, G., Hijmans, R.J., Funk, V.A., Cassis, G.A., Fisher, B.L., Flemons, P., Lees, D., Lovett, J.C., van Rompaey, R.S.A.R (2004) Mapping more of terrestrial biodiversity for global conservation assessment. BioScience 54: 1101-1109.
  • Hansen, M.C., Potapov, P.V., Moore, R., Hancher, M., Turubanova, S.A., Tyukavina, A., Thau, D., Stehman, S.V., Goetz, S.J., Loveland, T.R., Kommareddy, A., Egorov, A., Chini, L., Justice, C.O., Townshend, J.R.G. (2013) High-resolution global maps of 21st-Century forest cover change. Science 342: 850-853.
  • Hoskins, A.J., Bush, A., Gilmore, J., Harwood, T., Hudson, L.N., Ware, C., Williams, K.J., Ferrier, S. (2016) Downscaling land‐use data to provide global 30” estimates of five land‐use classes. Ecology and Evolution 6: 3040-3055.
  • Newbold, T., Hudson, L.N., Arnell, A.P., Contu, S., De Palma, A., Ferrier, S., Hill, S.L.L., Hoskins, A.J., Lysenko, I., Phillips, H.R.P., Burton, V.J., Chng, C.W.T., Emerson, S., Gao, D., Pask-Hale, G., Hutton, J., Jung, M., Sanchez-Ortiz, K., Simmons, B.I., Whitmee, S., Zhang, H., Scharlemann, J.P.W., Purvis, A. (2016) Has land use pushed terrestrial biodiversity beyond the planetary boundary? A global assessment. Science 353: 288-291.
  • Watson, J.E.M., Jones, K.R., Fuller, R.A., Di Marco, M., Segan, D.B., Butchart, S.H.M., Allan, J.R., McDonald-Madden, E., Venter, O. (2016) Disparities between recent rates of habitat conversion and protection and implications for future global conservation targets. Conservation Letters 9: 413–421.
  • Williams, K.J., Harwood, T.D., Ferrier, S. (2016) Assessing the ecological representativeness of Australia’s terrestrial National Reserve System: A community-level modelling approach. Publication Number EP163634. CSIRO Land and Water, Canberra, Australia https://publications.csiro.au/rpr/pub?pid=csiro:EP163634

National use of indicator

Producing this indicator nationally: The BHI indicator is derived using data and models covering the entire global extent of forest biomes at 30-arcsecond (approximately 1km) grid resolution. This relatively fine spatial resolution allows the indicator to be disaggregated, and reported, reliably at national level.

Use of the global method and data at the national level: CSIRO has calculated the BHI indicator for all individual countries, using national subsets of the global 30-arcsecond-resolution data and models. National-level results have been generated for the same years reported globally – i.e. 2000, 2005, 2010, 2011, 2012, 2013, and 2014. CSIRO is currently collaborating with NatureServe to make these results freely accessible via their Biodiversity Indicators Dashboard in the near future.

The methodology used to derive the BHI can potentially be applied to in-country biological, environmental and habitat-condition data in place of, or in combination with, the data employed globally. However this would require expert involvement of CSIRO to undertake additional model-fitting and analysis.

Further resources

Key indicator facts

Indicator type

Pressure

Applicable for national use

Yes (find out more)

Indicator classification

Operational and included in the CBD's list of indicators

Indicator type

Pressure

Applicable for national use

Yes (find out more)

Indicator classification

Operational and included in the CBD's list of indicators

Last update

2018

Coverage

Global

Availability

Not freely available

Partners

Csiro logo

Commonwealth Scientific and Industrial Research Organisation (CSIRO)

Contact point

Dr Simon Ferrier - simon.ferrier@csiro.au