Monitoring Change in Vertebrate Abundance: the Living Planet Index

Abstract

The task of measuring the decline of global biodiversity and instituting changes to halt and reverse this downturn has been taken up in response to the Convention on Biological Diversity's 2010 target. It is an undertaking made more difficult by the complex nature of biodiversity and the consequent difficulty in accurately gauging its depletion. In the Living Planet Index, aggregated population trends among vertebrate species indicate the rate of change in the status of biodiversity, and this index can be used to address the question of whether or not the 2010 target has been achieved. We investigated the use of generalized additive models in aggregating large quantities of population trend data, evaluated potential bias that results from collation of existing trends, and explored the feasibility of disaggregating the data (e.g., geographically, taxonomically, regionally, and by thematic area). Our results show strengths in length and completeness of data, little evidence of bias toward threatened species, and the possibility of disaggregation into meaningful subsets. Limitations of the data set are still apparent, in particular the dominance of bird data and gaps in tropical-species population coverage. Population-trend data complement the longer-term, but more coarse-grained, perspectives gained by evaluating species-level extinction rates. To measure progress toward the 2010 target, indicators must be adapted and strategically supplemented with existing data to generate meaningful indicators in time. Beyond 2010, it is critical a strategy be set out for the future development of indicators that will deal with existing data gaps and that is intricately tied to the goals of future biodiversity targets.

Indicator framework for measuring quantity and quality of biodiversity—Exemplified in the Nordic countries

“…These new policy targets call for strengthening of efforts towards preserving biodiversity but also need to be accompanied by robust and representative methods in order to measuring changes in biodiversity. Much has already been achieved, however, as a number of assessments including our study show; the most important limitation in determining biodiversity trends is the general lack of monitoring data (Collen et al, 2009;Butchart et al, 2010;EEA, 2010; Secretariat of the CBD, 2010). If we want to make a balanced and comprehensive assessment of the development of biodiversity -and hence evaluate the 2020 targets -further efforts towards monitoring programmes to obtain reliable, high quality data on biodiversity at acceptable spatial and temporal resolutions are required.…”

“…This is similar to the European bird indices, which are calculated from trend data for bird species that expert ornithologists agree to be representative of for example farmland or forest (Gregory et al, 2005). In comparison, the Living Planet Index follows a bottom-up approach for which all available species abundance trends are aggregated into a global index (Collen et al, 2009). …”

“…Changes in biodiversity quality are measured as trends in the abundance of selected species within these habitat types and -dissimilar to the NCI -by other key indicators for habitat quality, such as trends in the proportion of old trees or dead wood in forests. Estimating species abundance trends is a broadly accepted approach to generating indicative measures of biodiversity (Mace and Baillie, 2007;Collen et al, 2009;EEA, 2010).…”

Time series data for Canadian arctic vertebrates: IPY contributions to science, management, and policy

“…We calculated an index to display trends of time series data following the same approach used by the living planet index (LPI) (Loh et al 2005;Collen et al 2009). This allowed us to set the variety of our data collections within the same scale and compare them to trends observed at global, Canadian and arctic scales (Arctic Species Trends Index, ASTI) (Arctic Biodiversity Trends 2010; Butchart et al 2010;World Wildlife Fund WWF 2010).…”

“…For the LVI, we set a value of 1 to the first year of data if after 1970. Missing observations for times series data were interpolated using a gam-fitting framework (generalized additive modeling) and those with less than 6 years were estimated with a constant annual rate of change (Fewster et al 2000;Loh et al 2005;Collen et al 2009). The gam models allowed us to derive non-linear trends (e.g.…”