1 INTRODUCTION

The use of permanent plots has a long tradition in ecology (Callahan, 1984; Wildi and Schültz, 2000; Lindenmayer et al., 2012; Hughes et al., 2017) and vegetation science (Bakker et al., 1996a). Recently, permanent‐plot studies were considered among the six most important developments in vegetation science (Chytrý et al., 2019). As the present Special Feature demonstrates, the value of permanent plots is becoming ever more evident as a growing number of available time series highlights the variability inherent in plant communities and the non‐linear ways in which community composition and function respond to global change. In a previous Special Feature in Journal of Vegetation Science edited by Bakker et al. (1996a), different contributors showed the importance of permanent plots in understanding the mechanisms underlying vegetation changes, particularly following succession. Bakker et al. (1996a) used the term ‘permanent plots’ broadly to ‘include studies in which a series of randomly located plots or transects have been described at certain time intervals within a fixed area’. Such permanent plots are thus based on regular observation of the temporal dynamics of vegetation using sampling units with a fixed location in time, while the sampling approach is kept consistent.

A similar approach is the resurvey of vegetation plots. The topic of vegetation resurvey was well covered recently in a stimulating Special Feature in the sister journal Applied Vegetation Science (Hédl et al., 2017). In this approach, historical vegetation plots (usually older than two decades) are resampled using the same or similar sampling method, though not always using the same exact geographical location (Alstad et al., 2016). Vegetation resurveys are often conducted when contemporary researchers wish to capture trends in vegetation response to environmental changes — such as climate change — that the original researchers did not anticipate (Harrison et al., 2010). These opportunistic studies aim to make the best use of existing data and allow earlier observations than many permanent‐plot studies, but a downside of this approach is the potential risk of relocation and sampling biases (Kapfer et al., 2017). The distinction between permanent plots and semi‐permanent (or quasi‐permanent) plots used in the vegetation resurvey approach is often not definite, and both approaches can be very useful to assess medium‐ to long‐term trends in vegetation (Figure 1). At the same time, as further discussed in this Special Feature, frequent and regular sampling using permanent plots allows the assessment of species dynamics and community stability (Figure 1 and this issue) in addition to longer‐term trends. Within permanent plots, we can further differentiate between observational plots and experimental plots, where natural or semi‐natural vegetation is sampled after the application of experimental treatments in the latter. The difference between these types of plots is that experimental set‐ups are affected by both experimental treatments and natural variability. Establishing and, when needed, maintaining the treatments can require additional effort.

Sampling with a long‐term view takes effort. As summarized by Bakker et al. (1996a) ‘it needs a great deal of discipline to maintain a series of permanent plots and analyse them yearly over a period long enough to answer relevant (ecological) questions’. The commitment of individual researchers to permanent‐plot sampling have significantly advanced the field of ecology. For example, permanent‐plot studies spearheaded by a few individuals have elucidated the cyclical nature of population dynamics (The Portal Project; Morgan Ernest et al., 2016), and the role of disturbance for diversity (Jasper Ridge; Hobbs et al., 2007). Such discipline, however essential, is likely not the only trait required of researchers who successfully undertake the challenge of establishing and maintaining permanent plots for many years. Researchers also need to be able to secure support from institutions, either academic or governmental, including continuous funding, special agreement with landowners, security at the sampling sites and safe and stable data storage. Institutional support is likely a major bottleneck, particularly in the context of predominantly short‐term scientific support from most existing grant agencies. Because of this, most field observations and experiments are conducted only over short periods, despite the fact that environmental drivers work over long time periods, the response of vegetation could be delayed in time (see extinction debt; Helm et al., 2006) and that the effect of management may have long‐term legacies (e.g., short‐term fertilization effects detectable after 70 years; Spiegelberger et al., 2006). With some notable exceptions (e.g., Crawley et al., 2005; Silvertown et al., 2006), many permanent‐plot sampling schemes do not exceed a few decades, often overlapping with the career of a few dedicated researchers. Developing funding mechanisms to support the long‐term work of individual research teams provides the missing support needed.

A limited number of national and international initiatives have successfully launched and maintained permanent vegetation monitoring schemes worldwide, particularly using forest and grassland plots. For example, the Center for Tropical Forest Science established a global network of forest inventory plots in the 1980s (Anderson‐Teixeira et al., 2015). Currently, together with standardized sampling and data storage (Condit et al., 2014), this evolved into the ForestGEO initiative (https://forestgeo.si.edu/what‐forestgeo). Similar initiatives include, for example, the Chinese Forest Biodiversity Monitoring Network (http://www.cfbiodiv.org/; De Cáceres et al., 2012), the Brazilian Biodiversity Research Program (Magnusson et al., 2018), the Spanish Forest Inventory (Ruiz‐Benito et al., 2013) or New Zealand's Land Use and Carbon Analysis System (Holdaway et al., 2017). In 1980, the US National Science Foundation established the Long‐term Ecological Research (LTER) program, which supports a network of 28 sites to offer a long‐term view on ecological dynamics. Today, research programmes at multiple LTER sites (including in other regions of the world), provide open‐access ecological data to answer a number of pressing ecological questions across taxa. Other national initiatives, such as the Biodiversity Exploratories (BE), a German Science Foundation‐funded project, maintain a very exhaustive standardized sampling of plots along a land‐use intensity gradient in different regions. Similarly, the Environmental Change Network (ECN) focuses on monitoring, data and research to understand environmental change in the United Kingdom. Some initiatives have established a common sampling scheme to follow trends in composition and diversity in specific ecosystems, such as mountain summits (e.g., Pauli et al., 2012) or tundra (Elmendorf et al., 2012), although these sites are not always sampled on an annual basis.

Establishing comparable sampling schemes in different regions and habitats represents an ideal solution to develop robust monitoring schemes. However, this clearly requires a highly coordinated effort, with common and stable funds, which unfortunately is still often unrealistic. Moreover, there is a balance between standardized, comparable designs across systems and long‐term experiments tailored to test key purported dynamics of an individual system, with their specificities. Initiatives such as the ones mentioned above are restricted either to a few countries or to particular habitats and organisms. However, a number of ‘grassroot’ initiatives (Aubin et al., 2020) have developed worldwide to implement distributed, replicated permanent‐plot experiments (e.g., the global Nutrient Network, NutNet, https://nutnet.org/, Borer et al., 2014, or DroughtNet, https://drought‐net.colostate.edu/). At the same time, synthesis efforts have developed to compile permanent‐plot data, irrespective of specific sampling methods, across individual studies for cross‐site comparisons. For example, BioTIME (Dornelas et al., 2018) is an impressive initiative that collects data from existing long‐term sampling schemes for different organisms from independent sources for a minimum of two years, although not necessarily consecutively. This type of data, despite the sampling differences, can be effective to assess large‐scale trends in biodiversity (Dornelas et al., 2014; Blowes et al., 2019).

A particularly interesting example of independent efforts to monitor biodiversity in time is the Park Grass Experiment (e.g., Crawley et al., 2005; Silvertown et al., 2006). The Park Grass Experiment, begun in 1856, is likely the oldest ongoing ecological experiment. Its value to science has changed and grown since it was established to test primarily agricultural questions. Particularly in recent years, the interest in the original experiment has transcended its initial aim and facilitated tests of questions related to the mechanisms governing the relationship between biodiversity and productivity and the response of plant communities to atmospheric nutrient deposition (Storkey et al., 2016). Hence, the Park Grass initiative illustrates how long‐term experiments grow in value with time and how they may be used to investigate scientific questions that were inconceivable at their inception.

The papers in this Special Feature cover a number of long‐term studies that show how permanent plots can be essential to answering a number of important ecological questions. Some papers focus on the unique characteristics of individual sites (Brambila et al., 2020; Burge et al., 2020; Collins et al., 2020; Fischer et al., 2020; Herben et al., 2020) or intensive long‐term experimental manipulations (Hédl and Chudomelová, 2020; Liu et al., 2020; Rychtecká and Lepš, 2020; Ward et al., 2020) to test long‐standing ecological theories. Others combine long‐term data sets to identify general patterns across biomes, e.g., Ward et al. (2020), or Valencia et al. (2020, with the LOng‐Term Vegetation Sampling, LOTVS).

The pressing threat from multiple global change drivers and the need to follow their consequences in different regions and habitats worldwide call for coordinated efforts using repeated monitoring tools such as permanent plots (Borer et al., 2014). For this reason, we think it is important to answer the question: why do we still need to invest time, effort and funding in permanent plots? Following Bakker et al. (1996b), this Special Feature is an attempt to provide answers to this question and illustrate the need for special funding schemes beyond conventional ones that are based on short‐term funding cycles.