Understanding properly the `potential natural vegetation' concept



This is a response to critical comments concerning the inappropriate use of the potential natural vegetation (PNV) concept made in a recent contribution to the Commentary section of this journal. We consider that the PNV concept has been misinterpreted. PNV has been used extensively in several European countries since the mid-1950s and was never intended to be used to make a prediction of what vegetation would dominate in an area if human influence were removed. PNV maps express hypothetical assumptions of what corresponds to dominant or natural vegetation in each area. Remnants of the vegetation of the past provided by palaeopalynology and other disciplines provide valuable information for interpreting modern vegetation, but natural changes and anthropogenic influences operating over the last millennia have to be taken into account. Annex I of the Habitats Directive provides a balanced list of habitat types for implementing conservation policies within the European Union.

A recent article in the Commentary section of this journal, entitled ‘The survival of the “natural potential vegetation concept” (or the power of tradition)’ (Carrión & Fernández, 2009), revives an old debate in plant ecology. The debate is so old and has so completely run its course that, in our opinion, it no longer holds any interest for an international audience. Rather, it is a local echo of the old Gleasonian–Clementsian controversy long since resolved, regarding which most plant ecologists adopted an intermediate position some time ago (Gurevitch et al., 2002).

After examining the article, it seems clear that it is based on an inexact interpretation of the potential natural vegetation (PNV) concept. This concept was one of Tüxen’s notable contributions to vegetation science and it dates from the middle of the last century (Tüxen, 1956). Since then, many other authors in various countries have added to and used it, frequently in association with vegetation mapping projects. It is important to point out that PNV units are always constructed on the basis of the modern vegetation present in the area, and, as reasonable as that may be, they represent a projection to the future (not to the past). The potentiality of one particular vegetation type has been formulated more in terms of a hypothesis than a prediction; it has never been intended to be a prophecy of what vegetation would be certain to establish in the absence or removal of human impact (Härdtle, 1995). After the survey of a territory, a vegetation scientist decides which type of vegetation found in the area can be designated as PNV; ordinarily, the largest, most structured and competitive remnant type is chosen. This has meant that many Quercus and Fagus forests have been used as PNV types, as they are among the most competitive trees of the modern native dendroflora. Forest types dominated by other tree species, such as, for example, native pines, have however been used to a lesser extent. This is because many of the modern pine formations are secondary forests associated with disturbance (fire, management, etc.) and cannot be considered to be the most competitive type in many areas. In other cases, these forests are associated with steep slopes and shallow soils, and their primary stations correspond to permanent vegetation in rough stony terrains. Nevertheless, the PNV area attributed to pine forests in the Iberian Peninsula has increased in recent times with respect to that in previous approaches for this region (Rivas-Martínez, 1987, 2007).

Another important point is that, from the very beginning, the PNV concept was established as different from that of primitive vegetation, and it has always been accepted that the potential vegetation cannot correspond exactly to what was there before human disturbance began (Braun-Blanquet, 1964). This is true for many reasons: changes in topography, soil and climate have taken place across large areas of the world in the period since human impact began; in addition, many taxa have become extinct or have migrated. If human pressure were to be removed now, it would take a long time for a potential natural forest to grow; indeed, it would take so long that the climate would probably change again in that time. Thus, PNV is more a theoretical concept than a true prediction of the future.

For that reason, in some cases the term vegetation series (VS), or Sigmetum, has been used, as it is more neutral, less predictive, and suggests that units are being constructed from vegetation complexes that live under homogeneous environmental conditions. In practical terms, the cartography of such VS becomes equivalent to the mapping of ecologically homogeneous units, each of them populated by a particular complex of vegetation types that have a particular system of dynamic relationships. Among these vegetation types the one selected as PNV is interesting because it represents an assessment of the biotic possibilities of the habitat (Braun-Blanquet, 1964).

Unlike the future, which can only be guessed at, the past can be known, at least to a certain extent, and is studied in several scientific disciplines, for example palaeopalynology and anthracology, that identify evidence of the flora of the past. All findings are extremely valuable for determining what vegetation was present at various times in the past, as well as its evolution towards the current situation. However, a direct reconstruction of the vegetation following the relative abundances of the various pollen types appearing in the spectrum is not possible, as each plant species produces different amounts of pollen, which in turn spreads very different distances. It also has to be taken into account that some pollen types can originate from trees that were quite distant from the surveyed site; that is, there is no linear equivalence between pollen abundance and the abundance of mother plants (Rodrigues Tarroso, 2007). Thus, the pollen spectrum of each site has to be interpreted carefully, and the reconstruction of the surrounding vegetation is approximate. Even so, vegetation scientists tend to read findings published from this field of research with interest, as they help in the interpretation of present vegetation. Extant reality must, however, be used to construct PNV units – the past is essential for interpreting the present but should not condition the formulation and definition of modern vegetation units.

The map included in the paper by Carrión & Fernández (2009) shows differences between the data from the past and the situation today in several sites, but the remnants of the past cannot be interpreted as representatives of a modern or future PNV, as conditions have changed considerably during the last millennia. It is also generally accepted by the scientific community that anthropogenic influence has been the dominant shaping factor of the vegetation of modern vegetal landscapes in terms of plant taxa and vegetation distribution and abundance (Loidi, 2007). These human-induced transformations occurred while the vegetation was probably still reacting to previous climatic changes: we must assume that there is a lag between environmental changes and the adaptation of the vegetation to new conditions, particularly among long-living trees and their forests. In many areas, however, it is probable that human influence has catalyzed some changes in vegetation distribution. In fact, predictive distribution models of forest tree species based on current climate (Benito et al., 2006, 2007) do not differ from PNV patterns as markedly as the Carrión & Fernández map seems to imply, and the same is true of remote sensing mapping derived from the normalized difference vegetation index (NDVI) (Lobo et al., 1997), although in both cases discrepancies may be detected in some territories.

Phytosociologists in Spain have recently made notable contributions to the inventories of habitat types, particularly after the implementation of the Habitats Directive of the European Union in this country, and feel proud of having been instrumental in this task, which consisted mostly in the realization of a 1:50,000 vegetation map of the entire country (Loidi, 1999). This inventory, rigorously drawn up, substantially increased the knowledge of the distribution of all vegetation types across the territory, with strong implications for conservation and land planning in favour of the biodiversity and functioning of terrestrial ecosystems in order to ensure their sustainability and services (Rivas-Martínez et al., 2003). It is hardly reasonable to accuse this Directive of a bias against conifer forests or in favour of certain PNV types, when its Annex I includes all natural and semi-natural Mediterranean pine forests (except recent plantations) among the habitat types of European conservation interest, as well as a vast number of semi-natural habitats related to different disturbance regimes, and contains a substantial proportion of the Mediterranean endemic flora, while some beech, oak and birch forests have been excluded.

Potential natural vegetation or VS cartography, widely used in Europe (Härdtle, 1995; Bohn et al., 2000), represents a valuable tool for land management as it provides a reference for ecological restoration in terrestrial ecosystems (Moravec, 1998). It does not produce a map of recommendations exclusively oriented to favouring or restoring the PNV vegetation types because decisions are made by local managers and interests vary according to area. If natural forests are desired, targets for the most appropriate ones are provided, whereas if other seral vegetation types are to be maintained in complex systems, some guidance for management should also be given. Such a map is useful for conservationists and for ecological land management and planning; however, it does not represent a hard and fast rule of what must be done in each place.

The PNV is a scientific concept that requires a detailed knowledge of the flora, the vegetation and the mesological conditions of the territory, of course within the European geobotanical tradition. It has nothing to do with authority.

Editor: John Lambshead