Florian Hartig

← Publications

A comparison of methods for estimating plant population size

Christoph Reisch, Christoph Schmid & Florian Hartig

Biodiversity and Conservation, 27(8), 2021–2028 (2018)
Cite this
@article{reisch2018comparison,
  author = {Christoph Reisch and Christoph Schmid and Florian Hartig},
  title = {A comparison of methods for estimating plant population size},
  journal = {Biodiversity and Conservation},
  volume = {27},
  number = {8},
  pages = {2021--2028},
  year = {2018},
  doi = {10.1007/s10531-018-1522-1},
}

DOI: 10.1007/s10531-018-1522-1
Citation statistics: 26 (Google Scholar) · 19 (OpenAlex), as of 29 October 2026

Google Scholar citations per year, as of 21 October 2026 (yearly counts are fetched separately from the total)

View article (DOI)

Abstract

Assessing local population size is one of the most common tasks in biodiversity monitoring. Population size estimates are not only important for conservation management and population threat assessment; they also enter many other analyses in landscape ecology and conservation. It is therefore concerning that methods for estimating plant population sizes are not standardized. We surveyed the literature and found that the most commonly used methods are counting either all or only flowering individuals on a site, as well as counting individuals in random plots or plots on transects. Sometimes, these methods are combined in the same study, without assurance that they produce comparable results. We therefore conducted a field study, in which we obtained population size estimates from all four methods for six different calcareous grassland species at 18 study sites. Our results demonstrate not only substantial differences between overall count rates generated by the different methods, but methods that surveyed the whole population also systematically yielded lower counts when species were less visible and when the area was larger, suggesting that these methods suffer from biases that could distort species and site comparisons. We conclude that estimates from different methods should not be mixed, and that plot or transect based surveys have likely smaller biases for large areas or poorly visible individuals, and are therefore preferable.

What the paper shows and why it matters

(AI-generated)

Counting the individuals of a plant population is a routine task in ecology and conservation. Population size is linked to genetic variation and extinction risk and feeds into monitoring and conservation law. Yet no standard method exists. A survey of 143 studies found that a third gave no details of their method, that nearly half counted or estimated the individuals on the whole site (all plants, or only flowering ones), and that a few counted in sample plots or along transects and scaled up to the site area. Studies sometimes mix methods without assurance that they give comparable results. Such mixing is only safe if the methods agree, and the paper asks whether they do, and whether their errors depend on the species or the site, which would distort comparisons even within one study.

The paper argues that the common methods for estimating plant population size are not interchangeable and that whole-site counts are probably biased, in ways that depend on the species and the site. The study estimated the population size of six grassland species at 18 sites near Regensburg with four methods: counting all individuals or only flowering ones on the whole site, and counting in random squares or along transects and multiplying by the site area. Whole-site counts were roughly ten times lower than plot-based estimates, lower for less conspicuous species, and estimated density fell as sites got larger, whereas for plot-based estimates density showed no significant change with area (estimates were on average proportional to it). Because true population sizes were unknown, the analysis can show only that the methods differ. The authors judge it highly likely that the whole-site counts are the biased ones, in line with Keith (2000), who had suggested that counting all individuals underestimates population size at large sites.

Estimates from different methods should not be mixed, since this can lead to wrong conclusions. Whole-site counts can distort comparisons between species and sites, so the paper recommends plot-based methods for situations like the one studied. It adds that in less dense habitats with easily recognised plants, whole-site counts could do as well, which it leaves for further study. Perret et al. (2023), a study of how often observers miss plants when counting in quadrats, cite the paper.

References

  • Keith (2000). Sampling designs, field techniques and analytical methods for systematic plant population surveys. Ecological Management & Restoration 1, 125–139. doi:10.1046/j.1442-8903.2000.00034.x
  • Perret et al. (2023). Plants stand still but hide: Imperfect and heterogeneous detection is the rule when counting plants. Journal of Ecology 111, 1483–1496. doi:10.1111/1365-2745.14110