Random population fluctuations bias the Living Planet Index

Abstract

The Living Planet Index (LPI) is a standardised indicator for tracking population trends through time. Due to its ability to aggregate many timeseries in a single metric, the LPI has been proposed as an indicator for the Convention on Biological Diversity's post-2020 Global Biodiversity Strategy. However, here we show that random population fluctuations introduce biases when calculating the LPI. By combining simulated and empirical data, we show how random fluctuations lead to a declining LPI even when overall population trends are stable, and imprecise estimates of the LPI when populations increase or decrease non-linearly. We applied randomisation null models that demonstrate how random fluctuations exaggerate declines in the global LPI by 9.6%. Our results confirm substantial declines in the LPI, but highlight sources of uncertainty in quantitative estimates. Randomisation null models are useful for presenting uncertainty around indicators of progress towards international biodiversity targets. Nations of the world are in the process of negotiating the post-2020 Global Biodiversity Framework under the Convention on Biological Diversity. One of the ambitions of the post-2020 framework is "bending the curve of biodiversity loss" by first slowing down declines by 2030 and then improving the state of biodiversity by mid-century 1-3 . Reliable biodiversity indicators are essential for tracking progress towards global biodiversity targets 4 . A prominent indicator of species abundance over time is the Living Planet Index (LPI) 5-7 . The LPI aggregates population time-series for vertebrates from terrestrial, freshwater and marine systems into a relative index (where the baseline is

Representation of global change drivers across biodiversity datasets

“…2) Statistical weightings have been used to adjust the representativeness of the data sample e.g., by up-weighting under-represented regions or taxa (e.g., as employed by the Living Planet Index 82 and often with citizen science data 31,73 ) but this approach can over emphasize the effect of very small portions of the overall data 83 and potentially inflate errors associated with those data 36,60,83,84 . 3) Bias can be explicitly modelled using fixed effects for continuous variables of driver intensity and random effects to represent geographic, temporal and taxonomic structure (e.g., as in 85 ), but care must be taken to ensure all uncertainties are propagated through to the global mean estimate [86][87][88][89] .…”

Past, present, and future of the Living Planet Index

“…These efforts have also reinforced the LPI as a measure of “relative abundance” rather than “abundance” to help avoid misinterpretations 154 . We have already seen an uptake in the use of the LPR 2020 technical supplement in recent publications and blogs exploring the LPI 155,156 . The analogy of a FTSE index for biodiversity is most commonly used to describe the LPI, but a focus in the future should be on finding other ways to communicate the index that mitigate the use of dramatic narratives but without compromising while retaining the simple message of the LPI that can be broadly understood.…”

Assessing the representation of species included within the Canadian Living Planet Index

“…Recently the discourse in the literature in relation to the LPI has been heightened, with papers exploring biases pertaining to the underlying data and calculation of aggregate measures (Leung et al 2020;Buschke et al 2021;Leung et al 2022a;Leung et al 2022b;Leung et al 2022c;Loreau et al 2022;Murali et al 2022;Puurtinen et al 2022). For instance, random population fluctuations (Buschke et al 2021) and extreme population trends, particularly declines, (i.e., outliers; Leung et al 2020) introduce biases that can exaggerate declines in the global LPI. Despite these biases, the effect of random population fluctuations does not detract from global LPI messaging of substantial declines in average vertebrate population abundance (Buschke et al 2021).…”