Biodiversity in TFs
Earthwatch data
Does Tiny Forest age influence biodiversity
Alpha and beta diversity
Biodiversity (BD) is a deceptively simple concept, but its measurement is far from straightforward and for which numerous metrics and indices have been devised. There are two basic building blocks for all biodiversity metrics - the number of species, and the abundance or frequency of each species. In an ideal world, biodiversity estimates would be based on exhaustive samples which identified every species in an area and counted the frequency of each, but this is virtually impossible.
Even in small areas intensively surveyed, fully capturing all species across the vast complexity of biological life is a herculean task. Assessing the presence of some species or taxa will require specialist techniques Ike trapping or laboratory services, or rely on skilled identification. Species may come and go as a result of migration, habitat change, dispersal, colonisation or extinction, so continuous monitoring over long periods would be necessary to fully inventory a site, and bidiversity is likely to be dynamic. It is unlikely that resources exist to achieve this even in the smallest areas, and if appropriate resources could be deployed, necessity dictates they are focused on locations or taxa of higher conservation priority.
Instead, biodiversity is increasingly estimated from samples and surveys, and from large scale citizens databases like NBN atlas, GBIF and eBIrd. The growing scale, speed and sophistication of citizen science projects …
Their scope, scale and size, large BD databases creates promise and challenges for BD estimation. There are generally two kinds of observation submitted to citizen science databases (CSDs) - opportunistic, and as part of organised survey programmes, often organised by national bodies like the BTO, Centre for Ecology and Hydrology (CEH) or specialist societies. Examples include the British Breeding Bird Survey, National Plant Monitor Survey, and the Botanical Society of Britain and Ireland atlas project. These kinds of surveys follow standardised protocols like random sampling in site selection, prescribed methods for capturing data, measurement of effort, and training of participants. By contrast, opportunistic submissions may be less accurately identified and recorded. Nevertheless, the inclusion of geolocated and time stamped multi-taxa data at scale provides opportunities for understanding variation in biodiversity and change over time, and relationships with environmental and climate change.
The most widely used BD metric is species richness (SR) which is a count of the number of different species in a sample or area. This is known as alpha diversity. Comparing species richness between samples or sites, or over time is essential for establishing baselines, assessing the effectiveness of conservation measures and so on. There is no single way of measuring SR - the simplest is a count of the number of distinct species observed (species number) but this is generally held to be a minimum value because of the problem of under-detection discussed below. Further discussion….
Analytically, calculating species richness is based on creating site-species matrices (ssm) - tabular data with species as columns, sample sites as rows, and counts of observations as cells. Two problems arise. Firstly, observation counts at a site are not necessarily estimates of abundance. Taking examples from bird observations, a single rare bird may attract considerable attention be entered many times by many observers generating a cell count in an ssm much greater than one; some species are difficult to detect unless conditions are right (for example dartford warblers Sylvia undata, spend much of their time deep in gorse bushes); flock sizes can be vast (e.g. knots Calidris canutus on the Wash). Secondly, a ‘zero’ in an ssm does not mean that a species was not present at that site - it may not have been detected, or may have been misidentified. This means that making inferences about BD variation or trends from CSD needs careful consideration to avoid biased estimates. Techniques for dealing with these issues are are discussed further in the methods section.
Estimating change in alpha and beta BD in TFs
Establishment of control sites for existing Tiny Forests (e.g. adjacent unchanged grassland, or nearby wooded areas), sampled to the same intensity and protocol was not possible - data collection was only possible within TF boundaries. Independent field work by the researcher was not possible as many of the TFs were on private land or school grounds without public access.
Three alternative approaches to assess the impact on biodiversity of TFs are firstly to compare species richness and turnover (changes in community composition) between TFs of different ages or for the same TFs over time; secondly to derived ‘expected’ estimates of species richness for TF and the surrounding area and compare with observed TF variation; and thirdly ….as a function of forest age at the time of survey, controlling for potential confounders, for example , latitude, and planted tree diversity.
## Determinants of urban biodiversity
There are numerous ‘lenses’ for viewing potential determinants of BD for Tiny Forests but given the vast literature it is helpful to narrow the focus. TFs could be viewed as individual restoration projects or as a contribution to a wider programme of urban regeneration. This regeneration process has many names - ecological engineering, ecological restoration, the Miyawaki method or rewilding (reintroducing native species) to name but a few. TFs are largely situated in urban or suburban areas on improved grassland, changing the land cover to a potentially more bio diverse woodland. This regeneration process has many names - ecological engineering, ecological restoration, the Miyawaki method and rewilding (reintroducing native species) to name but a few.
More recent literature about urban biodiversity introduces the concept of green infrastructure defined by … as
the..
Small is beautiful?
Given their size (200m2 or 0.02 Ha per TF with a collective size of UK TFs of 4 Ha) it is tempting to ask if TFs could make a significant contribution to biodiversity gain. Recent conservation policy has been focused at much larger scale [], SLOSS, however there is growing interest in the value of small areas. Wintle et al. (2019), arguing that island biogeography theory, which has influenced thinking about enlarging and joining habitats for conservation, reviewed the global literature on the conservation value of small and isolated habitat patches. This is highly relevant
To achieve this EW survey data was assembled into a longitudinal format for each taxa, and a variable age_at_survey was calculated as the difference between survey data and date of TF planting in days. Species richness was calculated for each survey and modelled as a function of age_at_survey using generalised linear mixed modelling (GLMM) with TF site as a random variable. Differences in community composition were estimated using generalised dissimilarity modelling (GDM), which models the effect of predictor variables on the differences in pairwise comparisons of Bray-Curtis distance matrices between TF sites. The primary hypothesis tested in these analyses was “does forest age influence alpha and beta diversity?”. It was anticipated that as forests mature, species richness for butterflies, ground dwellers and pollinators would increase due to understorey growth, soil maturation and colonisation by pollinating plants or butterfly food plants. Similarly, more mature sites were expected to have different communities than younger sites, and there would be within-site changes in community composition.
Rows: 82
Columns: 18
Groups: site, year.y, month.y, survey_id [82]
$ site <dbl> 85, 85, 86, 87, 87, 87, 89, 92, 92, 92, 92, 92, 92, 92, 92, …
$ year.y <dbl> 2021, 2022, 2023, 2022, 2022, 2022, 2022, 2021, 2022, 2022, …
$ month.y <dbl> 7, 5, 7, 5, 5, 6, 5, 7, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 6, 7, …
$ survey_id <int> 824, 2268, 12911, 2442, 3006, 4112, 2429, 636, 2672, 2673, 2…
$ Species.x <int> 1, 2, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 2, …
$ chao <dbl> 1, 2, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 2, …
$ chao.se <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, …
$ jack1 <dbl> 1, 2, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 2, …
$ jack1.se <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, …
$ jack2 <int> 1, 2, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 2, …
$ boot <dbl> 1, 2, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 2, …
$ boot.se <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, …
$ n <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, …
$ Species.y <int> 1, 2, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 2, …
$ sample <int> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, …
$ Y <dbl> 51.77789, 51.77789, 51.76914, 51.73478, 51.73478, 51.73478, …
$ age <dbl> 498, 789, 901, 471, 481, 504, 454, 135, 451, 451, 451, 451, …
$ trees <int> 12, 12, 21, 19, 19, 19, 16, 21, 21, 21, 21, 21, 21, 21, 21, …
References
Wintle, Brendan A., Heini Kujala, Amy Whitehead, Alison Cameron, Sam Veloz, Aija Kukkala, Atte Moilanen, et al. 2019. “Global Synthesis of Conservation Studies Reveals the Importance of Small Habitat Patches for Biodiversity.” Proceedings of the National Academy of Sciences 116 (3): 909–14. https://doi.org/10.1073/pnas.1813051115.