Printer Friendly

River dynamics as a forest process: interaction between fluvial systems and alluvial forests in large European river plants.

II. Introduction

In alluvial plains, river dynamics is recognized as the major force for landscape patterning and ecosystem distribution. Very often the nomenclature for forests is restricted to detailed aspects of species composition, vegetation type, biogeochemical cycles, or forest management. There are fewer comprehensive works concerning the long-term natural forest dynamics (i.e., gap dynamics, successional processes or cycles of mature stands). Understanding the complex relationships between river dynamics and forest processes is critical in both applied and basic ecology. A thorough knowledge is necessary for the correct interpretation both of forest landscapes in big river plains and of the multiple roles that rivers play in maintaining them.

This paper, an overview of the multiple impacts of the recurrence of disturbance dynamics on the functioning of gallery forests, is based on scientific works published in European and North American reviews and on data and figures from more than 10 years of studies carried out in the upper Rhine valley, France.

III. The Geomorphic Pattern of Big River Plains

A floodplain is an accumulation of deposits carried and then set down by flowing water. Water is hence the dynamic component of the alluvial landscape that distributes energy, material, and information across and through the plain. Fluxes of water, sediments, and energy are both longitudinal (from upstream to downstream) and transversal (from the river to the edges of the plain and vice versa). Water movements are also vertical, with infiltration within the sediments of the plain and circulation through them (Amoros & Petts, 1993).

Water kinetic energy decreases with the decreasing slope of the valley. Along their courses from the high valleys to the coastal lowland, rivers present a wide diversity of geomorphological and hydrological patterns, in relation to the decreasing slope and various other factors such as subsidence, diversity of the geological structure of valleys, and effects of Pleistocene and Holocene history. Moor (1958) distinguished at least five sections characteristic of a river that has its source in the high mountains:

1. Near its source, a narrow stream including small channels into the rock

2. An upstream, erosive river sector with a wandering gravel-bed river split into small fast-flowing channels (braids)

3. A wide middle sector with a river divided into braids and anastomosed side-arms. Sedimentation and erosion alternate in space and time. Sediments are mainly sand and silt. Erosion, migration, and sedimentation constantly rework the landscape, which explains the complex sedimentological pattern with terraces, hollows, backwater depressions, wetlands, secondary arms, and old channels periodically connected with the river. Exchanges with groundwaters are highly active. In the Rhine valley, for example, the river permanently feeds the water table with about 3 [m.sup.3] [s.sup.-1], and abandoned channels form a cold, oligotrophic running network of groundwater streams (Carbiener & Tremolieres, 1990).

4. A wide low sector characterized by a gradient of less than 0.3%. The river presents large-sized meanders, and sedimentation is predominant. Deposits are fine-textured (clay).

5. Estuaries, with special conditions due to the influence of the tide.

Flooding periods vary according to the longitudinal sector. In sectors 1 and 2, flooding occurs in summer after the melting of snow and glaciers. In sector 3, floods occur in early summer but also in winter, and the influence of tributaries becomes important. Downstream, flood periods occur increasingly later in spring and winter (Ellenberg, 1988).

In middle and downstream sectors, tributaries entering through the valley create new ecological gradients, from very wet sites (swamps) to low-altitude terraces. In these two sectors, a transversal section of the floodplain from the main river to its edges shows increase of hydromorphy and enrichment in clay, due to the increasing influence of tributaries. These gradients are similar to those of the longitudinal profile. This homology between longitudinal and transversal profiles was emphasized for large river plains in Europe by Wendelberger (1952), for the Danube by Wendelberger (1984), and for the Rhine by Carbiener (1984) and Carbiener and Schnitzler (1990).

Floodplain pattern depends also on vegetation (Tricart, 1977; Pautou et al., 1986). It is well known that bare soil, newly deposited by rivers, very quickly becomes invaded by plants. Bushes of therophytes or woody species (essentially Salix and Populus) contribute, by their rapid growth, to raising the level of the ground by retaining sand, gravel [up to 60 cm in one season according to observations of Grelon (1971) in the middle Loire], woody debris, and leaves. A new sand layer covers these organic deposits during the next flood. Except in the case of a great flood, the island in formation grows regularly in width and height, with continuing accumulation of coarse debris and trapped sediment on its head and an encroachment of willow thickets. When the extension of the islands is lowered, the extension of the islands and the spread of plant cover can be dramatic: in the Loire valley the islands have doubled in the last century (Bomer, 1971). Coarse debris (dead trunks, large branches or roots) therefore play a major role in stream channel geomorphology (Tricart, 1977; Sedell & Frogatt, 1984; Triska, 1984; Harmon et al., 1986), by accumulating at sites such as heads of islands and mouths of secondary channels and bridges, causing local movements of colluvium and sediments. Coarse debris also creates backwaters, stores inorganic sediments, and maintains complex aquatic habitats. Dead trees originate from the river banks; they are frequently pulled down by the river when it undercuts the ground of the edges. Large portions of trees can also be swept away when big channels migrate laterally during great floods. In large rivers, coarse debris is transported downstream and out of channels onto floodplains. This transfer can reach high volumes: up to 5.2 [m.sup.3] [ha.sup.-1] [year.sup.-1] for five streams of different sizes in Oregon (Harmon et al., 1986).

IV. Particularities of the Nutrient Cycle

Exceptionally fertile conditions occur in floodplains more than in any other type of habitat, because of favorable temperature, favorable air and soil moisture, and regular input of nutrients and water through flooding. Temperatures are often higher in floodplains than in upland adjacent sites for various reasons, particularly because of protection by nearby mountains against cold winds and influence of warm-air masses when valleys are oriented north-south (e.g., in the case of the Rhine and the Rhone). Air moisture is also constantly higher than upland adjacent sites [up to 20% more (Wendelberger, 1952; Corillion, 1995)] because of the permanence of wide surfaces of evaporating water, particularly during flooding periods. High air moisture coupled with thermophilic mesoclimate enhance evapotranspiration, which probably reaches potential values (Carbiener, 1970). In large floodplains subject to summer floods, soil moisture is provided in hot months by the rising of groundwater coupled with floodings. In upstream sectors, high soil moisture does not induce below-hydromorphic horizons such as gley, thanks to the permeability of the sediments and the high amplitude of groundwater levels, which ventilate the soil and bring an immediate renewal of soil to flooded surfaces. If natural drainage quickly removes flood waters, flooding is hence an energy source bringing sediments and nutrients. Surface water is rich in nutrient salts, fine-particulated sediments [20-25 [micro]m (Decamps & Naiman, 1989)], and litter or coarse debris. One major effect of the floodwaters when invading the plain is therefore to manure soils in communities. Along the Garonne, it was assessed that a willow community gave to the river up to five tons per year and per kilometer (Tabacchi & Planty-Tabacchi, 1990). In February 1990, inputs of a Rhine inundation 1 m high and with 2 cm of sediment deposit were 27 kg [ha.sup.-1] 1 phosphorus, 8.9 kg [ha.sup.1] nitrogen, 86 kg [ha.sup.-1] sulphur, 51 kg [ha.sup.-1] magnesium, and 34 kg [ha.sup.-1] potassium (Sanchez-Perez et al., 1993). Ions and gases are transported within the sediments by vertical movements of flowing waters.

Manuring conditions, however, are not similar all over the floodplain. When the soil level is high above the average water level, inundations are rarer and plants therefore evolve in a less nutrient-rich environment Besides, they may have to cope with temporary dry periods, particularly when they grow on high terraces. When the soil level is barely above the average water level (essentially, old channels), soils are waterlogged and plants suffer from excess of hydromorphy and slowing of the nutrient cycle. The most productive zone is found around the center of the floodplain, where inundations are regular and calm, the soils coarse-textured, and the amplitudes of the groundwaters high.

Chemical transfers between the river and the plain present a spiraling dynamics from upstream to downstream (Amoros & Petts, 1993). In the upper sectors, C, N, and P are assimilated under organic form by invertebrates, then degraded, mineralized, transported, and reassimilated downstream with resilience times that vary according to the size of the debris. Storage in water of big dead trunks can be very long [up to several years (Franklin et al., 1981)], leading to a considerable slowing down of nutrient flux toward downstream, in spite of physical fragmentation by erosive effects of current and activity of aquatic invertebrates and bacteria. However, the importance of forested systems to consumer metabolism is very high. It was estimated that over 50% of the carbon consumed in estuarine waters originated in upstream forests. In conditions of high moisture, warmth, and richness in aerobic bacteria and mushroom populations, decomposition rate of organic matter is rapid--e.g., leaves of Fraxinus excelsior, Cornus sanguinea, Ulmus minor, Clematis vitalba, and Prunus padus (decomposition within 7 months, according to Penka et al., 1985). Nutrient supplies from litter are therefore available at the beginning of spring. In early summer there exists another source of nutrients, from ivy litter and the geophyte leaves of Allium ursinum. These species, abundant in flooded forests, decompose easily and rapidly (Tremolieres & Carbiener, 1985; Tremolieres et al.,1988).

The specificity of the Rhine fungi flora [high species richness, with a particular abundance in calciphilous and thermophilous species, (Carbiener, 1981)] is a direct response to the particular environment of the floodplain. When flooded, the forests are rich in lignicolous species (32 species). But the number of mycorrhiza is moderate (only 6 species, essentially Cortinariaceae), because of the importance of phosphorous and nitrogen inputs from floodwaters. Noteworthy is the absence of the Russulaceae (Table I). The relative poverty of the saprophyte population is another particularity of the flooded sites (17), due to the very rapid decomposition of the dead leaves. In unflooded sites, the mycorrhizal flora increases to 19 species with a great number of rare, basic species of Boletus, Tricholoma, and Russula genera. Common saprophytes (29 species) of the litter, humus, and buried dead wood appear. Lignicolous, highly specialized species are nearly as numerous as in flooded sites (31 species). Increase in mycorrhiza and saprophyte fungi in unflooded sites improves the nutrient turnover and the P uptake by plants, which partly compensates the loss in nutrient inputs from floodwaters.

Table I Mushroom species richness in flooded and unflooded stands of the Rhine forests (data from Carbiener, 1981; flora from Michael & Hennig, 1964-1970)
 Period of Flooded Unflooded
Species fructification stand stand

 Mycorrhizal species

Amanita echinocephala Aug-Sep (**)
Amanita strobiliformis Jun-Sep (*) (**)
Boletus luridus Jun-Sep (*)
Boletus radicans Jun-Sep (*)
Boletus rhodopurpureus Sep (*)
Boletus satanas Aug-Sep (*)
Gyrodon lividus Sep-Oct (*) (**)
Hebeloma sacchariolens Sep-Oct (*) (*)
Hebeloma sinuosum Sep-Oct (**)
Inocybe fastigiata Sep-Oct (*)
Inocybe maculata Aug-Oct (**) (*)
Lactarius acerimus Aug-Oct (*) (*)
Lactarius pubescens Aug-Oct (*)
Russula delica Jun-Feb (*)
Russula maculata Jul-Sep (*)
Russula pulchella Sep-Oct (*)
Russula vitellina Oct-Nov (*)
Scleroderma verrucosum Aug-Oct (*) (*)
Tricholoma scalpturatum Nov-Dec (**)
 Total number of species 6 19

 Saprophytic species of litter,
 humus and buried dead wood

Agaricus campestre Sep-Oct (**) (*)
Agaricus gennadii Jun-Oct (*) (*)
Agaricus subfloccusus Jun-Sep (*)
Calocybe ionides Aug-Oct (*)
Calvatia exepuliformis Jul-Oct (*)
Candidus albutiscorticus Aug-Nov (**) (*)
Chamaernyces fracidus Aug-Sep (*) (**)
Clitocybe candidans Aug-Oct (*) (**)
Clitocybe phyllophila Oct-Nov (*)
Clitocybe suaveolens Oct-Nov (*)
Collybia confluens Oct-Nov (*)
Collybia dryophylla Oct-Nov (*)
Cystolepiota sistrata Aug-Oct (*) (*)
Disciotis venosa Mar-Apr (**) (*)
Entoloma aprilis Apr (*)
Entoloma saundersii Mar-Apr (*)
Helvella crispa (*)
Hemimycena candida (*)
Lepiota acutesquamosa Jul-Oct (*) (**)
Lepiota cristata Sep-Nov (**) (**)
Lepiota sordida Sep-Nov (*)
Lycopordon perlatum (*)
Marasmius wynnei Sep-Oct (*)
Mitrophora semilibera Apr-May (**) (*)

Morchella esculenta Nov (*)
Morchella rotunda Nov (**) (*)
Mycena roseipalleus Jul-Oct (**)
Paxina acetabulum Apr (*)
Pholiotina foetidus (**) (*)
Psathyrella melanthina Sep-Oct (*) (*)
Psathyrella vernalis Apr-May (*) (*)
Verpa bohemica Apr-May (*)
Verpa digitaliformis Apr-May (*)
 Total number of species 17 29

 Lignicolous species

Armillaria bulbosa (*) (**)
Auricularia auricula judae Oct-Mar (**) (*)
Airocimara mesenterica (**) (*)
Chondrostereum purpureum (*) (*)
Coprinus domesticus Apr-May (*)
Coprinus radians Apr-May (*)
Funalia gallica (*) (*)
Funalia trogii (*) (*)
Laetiporus sultureus (**) (*)
Lentinus tigrinus May-Sep (**) (*)
Lyophyllum ulmarium (*) (*)
Marasmius candidus (**) (*)
Marasmius rotula (*) (**)
Micromphala foetidum (**) (**)
Mycena haematopoda (*) (*)
Phellinus conchatus (**) (*)
Phellinus pomaceus (*) (*)
Phellinus ribi (*)
Pleurotus cornucopiae Jun-Sep (**) (*)
Pluteus selicinus (*) (*)
Polyporus brumalis (*) (*)
Polyporus lentus (*) (*)
Polyporus lepideus Apr-Jun (*) (*)
Polyporus mori (*) (*)
Polyporus picipes Apr-Oct (*) (*)
Polyporus squamosus May-Sep (*) (**)
Stereum subtomentosum (**) (*)
Trametes confragora (**) (*)
Trametes suaveolens (**) (*)
Trametes tricolor (*) (**)
Tubaria hiemalis Oct-Mar (**) (*)
Tyromyces fissilis Sep-Oct (*) (*)
Tyromyces subcaesius (*) (*)
Total number of species 32 31


(*) Present; (**) abundant.

For many microbes, invertebrates, vertebrates, and plants, coarse debris represents a habitat and food source. Vascular plants may send their roots into rotting wood to extract water and nutrients. In the Rhine valley, logs are frequently colonized by shrubs and even trees (Rosa canina, Prunus padus)--which may emerge from rotting hearts of old, senescent trees, particularly in forests prone to long flooding (willow or alder forests)--because rotting trees are the only site on which mesophytic species can grow. Seedlings of some woody species (Hedera helix, Fraxinus excelsior, Acer pseudoplatanus) and grasses (Impatiens noli-tangere, Glechoma hederacea, Carex remota) are found on logs when they rot in full light of wide gaps, but they generally die after one year (pets. obs. along the Rhine).

Nutrient inputs are partly stored in the surface humus and partly taken up by plants. Very few nutrients reach the groundwaters, particularly under highly complex and species-rich forest communities (Sanchez-Perez et al., 1991). Topsoil of floodplain forests is an important and dynamic component, characterized by marked temporal and spatial changes in nutrient concentrations, due to the heterogeneous species composition and the periodic inundations (Penka et al., 1985; Sanchez-Perez et al., 1993). In basic environments, the soil retains phosphorus, potassium, and ammonium. When they are not taken up by plants, magnesium, sodium, sulfate, and nitrate are leached into the groundwaters. Phosphorus may be available to plants during flooding in the form of soluble ferrous phosphate. The regularly flooded sites therefore offer good conditions for P nutrition, in spite of calcareous substrates (Weiss et al., 1991). Concentrations and reserves of nutrients were estimated in woody components of the floodplain forest ecosystem of South Moravia. From this study, Penka et al. (1985) concluded that the highest supplies of nutrients were accumulated in tree stems, and the order of elements was as follows: Ca, N, K, Mg, and P. Leaves are also rich in N and P, as shown by a comparative foliar analysis of dominant woody species between sites prone to eutrophic floods and sites where floods were eliminated by dykes (Schnitzler et al., submitted).

The litter was also nutrient-rich: a total of about 5.6 kg ha yr in south Moravia. For some species, floods influence nutrient retention via the phloem during senescence in autumn, as proved in experiments carried out in the Rhine forests (Table II). The study (Schnitzler et al., submitted) indicates a higher N nutrient retention in flooded sites for Fraxinus excelsior, Clematis vitalba, and Ulmus minor, in spite of a high N nutrient supply in soils. This behavior does not concern P translocation, which may be either higher or similar in flooded sites and unflooded sites (Weiss et al., 1991).

Table II Mean values for concentrations (mg [g.sup.-1] on a dry weight basis of major nutrients in mature and dead leaves of three species in the Rhine valley; data from Schnitzler et al., submitted)
 Nitrogen Phosphorus
Fraxinus excelsior
 Summer leaves
 Flooded stands 36.0 1.5
 Unflooded stands 24.0 1.3
 Significance (**) (***)
 Autumn leaves
 Flooded stands 14.0 0.7
 Unflooded stands 13.4 0.5
 Significance (**) (***)

Ulmus minor
 Summer leaves
 Flooded stands 36.7 1.9
 Unflooded stands 26.5 1.4
 Significance (**) (***)

 Autumn leaves
 Flooded stands 9.9 0.7
 Unflooded stands 9.5 0.6
 Significance (***)

Clematis vitalba
 Summer leaves
 Flooded stands 42.8 2.1
 Unflooded stands 23.4 1.4
 Significance (**) (**)

 Autumn leaves
 Flooded stands 18.2 1.1
 Unflooded stands 12.9 0.8
 Significance (**) (**)

 Potassium Magnesium
Fraxinus excelsior
 Summer leaves
 Flooded stands 15.2 5.6
 Unflooded stands 12.3 6.8
 Significance (*) (*)
 Autumn leaves
 Flooded stands 7.1 3.7
 Unflooded stands 4.2 3.9
 Significance

Ulmus minor
 Summer leaves
 Flooded stands 19.8 3.2
 Unflooded stands 19.6 2.9
 Significance

 Autumn leaves
 Flooded stands 10.5 2.3
 Unflooded stands 8.7 2.2
 Significance

Clematis vitalba
 Summer leaves
 Flooded stands 19.1 4.1
 Unflooded stands 23.5 2.4
 Significance (**)

 Autumn leaves
 Flooded stands 9.6 3.2
 Unflooded stands 12.3
 Significance (**)


(*) P <0.05.

(**) P <0.01.

(***) P <0.001.

The high organic productivity of floodplains explains the high primary productivity and the high wood productivity. Many trees, shrubs, and woody lianas reach optimum sizes. Along the Danube valley, Salix alba and Populus nigra can reach up to 40 m in height and 5-6 m in circumference in less than 100 years. In the Rhine valley, Ulmus minor reaches 36-40 m. Understory trees or shrubs are often giant: 24 m for Prunus padus, 17 m for Crataegus monogyna. In Lanzhot Natural Park in Moravia, Korpel (1995) quoted about 907 [m.sup.3] pro ha in the optimum stage of oak-elm forest with ashes (Fraxinus excelsior) 47.5 m high and 139 cm dbh (diameter at breast height). From 1958 to 1987, the wood productivity increased from 600 [m.sup.3] to 728 [m.sup.3] per hectare.

Rapid growth and rapid decomposition explain why the forests have high nutrient recycling capacity, which in turn is why generation changes take less time than upland adjacent areas, the successional processes from bush pioneer communities to young mature hardwoods take less than 100 years (Koop, 1989), and the regeneration units are fragmented much more quickly (Frye & Quinn, 1979; Koop, 1989).

V. Specific Particularities

The flow of water in terrestrial ecosystems has many effects on plants and animals. Water acts as a resource because floods act as migration routes for hydrochorous seeds. Schwabe (1991) estimated the persistent seed bank in the soils of a forest community dominated by young Alnus incana trees at 935 viable seeds per [m.sup.2]. After a high-flood winter, new sediments downstream of the forest included about 38 viable seeds per [m.sup.2]. Transported seeds are also of allochtonous origin (exotic plants, e.g., those along the Loire as described in Loiseau & Felzines, 1992).

Floods also favor vegetative propagation through the transport of broken but living trunks of pioneer trees or bushes, which can anchor downstream (Grelon, 1971; Koop, 1987; Amoros & Petts, 1993). But flood water is also a constraint. It can damage plants when water speed slows rapidly. Accumulation of sediment frequently covers low vegetation. Plants react against physical damage by high reiteration capacity. Salix and Populus species, as well as Alnus, Tilia, Ulmus, and Fraxinus can build reiterative sprouts that can replace the vertical axis of a broken or fallen tree. When buried under sediments, species can develop lateral sprouts which allow them to survive. Another constraint is the temporal anoxia caused by standing waters during flooding. At the plant level, many morphological and anatomical adaptations have been described (Lugo et al., 1990). Flood tolerance of trees was investigated by direct observations of scientists when particularly long floodings occurred [i.e., along the Tennessee River, by Hall and Smith (1955); along the Rhine, by Dister (1983)] but also by experiments [along the Mississippi River, by Hosner (1960)]. Adaptations to anoxia include rapid seedling growth enabling them to emerge above the inundation level (this is the case for all tree species from regularly inundated forests), a rough and thick bark for keeping air (e.g., Quercus robur, Fraxinus excelsior, and Ulmus minor), and aerial roots (e.g., Alnus glutinosa, Salix species).

Alluvial species present a gradient in flood tolerance. In Europe, Salix alba shows a particularly high flood tolerance because its seedlings can survive up to 30 days of total submersion. Salix alba adults can tolerate up to 300 days of inundation; Quercus robur, Ulmus minor, and Ulmus laevis, up to 151 days. Fraxinus excelsior dies after 102 days; Acer campestre, Carpinus betulus, and Tilia cordata, after 13 days. Acer pseudoplatanus and Fagus sylvatica are not tolerant at all, particularly when floods occur in summer. These species are relegated to high-elevated terraces or are eliminated. In North America, the decreasing tolerance to complete inundation begins with Salix nigra, Fraxinus pennsylvanica, Liquidambar styraciflua, Populus deltoides, and Acer saccharinum (Hosner, 1960).

Resource availability and trees' adaptations to disturbance constraints play a part in the development of different life-strategies, which include the quantity and seasonality of propagule production, propagule dispersal, growth rates, vulnerability to death or injury from enemies or disturbances, the viability of recruitment from dormant propagules, and the potential for vegetative regrowth from damaged tissues. The model of Grime (1977), analyzed by Brzeziecki and Kienast (1994) using multivariate statistical methods, distinguishes three major groups and intermediates: the ruderal (R), the stress tolerant (S), and the competitive (C) strategies. Ruderal strategy is the strategy of species adapted to colonize intensively disturbed sites. In alluvial environments, the dominant pioneer species are Salix and Populus. They have short lifespan (100-150 years), high potential growth rate, low wood density (330-410 kg [m.sup.-3]), and early reproduction through wind and water dispersion (February to April) to allow reestablishment in bare sediments. The costs paid for these abilities involve a lack of seed reserves, which causes their vulnerability to shade or drought, and vulnerability of wood to attacks by insects and fungi. Alnus glutinosa, Alnus incana, and Betula pendula, also ruderal strategists, are present in the floodplains as scattered small populations because of less competitivity with Salix and Populus species.

Pure stress-tolerant species [mainly conifers, according to Bzreziecki and Kienast (1994)], are not very rare in the alluvial forests. Pinus sylvestris can germinate from seedlings coming from plantations on dry, coarse terraces along the Rhine valley. Juniperus exists naturally in situations of air and soil drought under the Mediterranean climate, on sandy dunes 3-5 m above the mean water level (dunes of the quaternary Danube river bed, Hungary, or dunes in the Rhone delta in Camargue, France). In the Rhone delta, Juniperus phonicea reaches large sizes: 6-8 m high, 30 cm dbh).

But the intermediate group, stress-tolerant ruderals (S-R), is represented by Acer campestre, Malus sylvestris, Pyrus communis, Prunus padus, and Prunus avium. These species, like ruderals, have short life-span (50-150 years), relatively high potential growth rate, and low age of sexual maturity; but they are stress-tolerant and have high wood density (500-700 kg [m.sup.-3]), small stature (15-25 m), large seeds dispersed by birds and small mammals, and a dominated position which explains why these species are scattered in end-successional stages.

Another intermediate group, competitive stress tolerators (C-S), is here represented by Quercus robur. In its establishment phase, this species is tolerant of low availability of water and mineral nutrients. That is why oaks are found along a wide range of moisture gradient, as long as floodwaters are not too frequent or too violent. Oak is able to grow rapidly, can attain large sizes at maturity (42 m, 160 cm dbh), and can reach great age (400-500 years). It is resistant to biotic and abiotic agents and plays the major role among end-successional species.

The third group, of pure competitive strategy, is composed of Ulmus minor and Fraxinus excelsior, which are most successful on the productive sites under conditions of high resource availability and low disturbance, in well-drained, silty, and fertile low floodplain terraces. Fraxinus excelsior and Ulmus minor can attain large sizes and can live to about 300 years. They, along with oak, form the canopy of the end-successional stage in the most fertile sites. Fraxinus excelsior, which is of neutrophilous tendency, is not dominant in all floodplains of central Europe. On subacidic sediments (e.g., along the Loire), it is easily dominated by elm and oak.

Dutch elm disease (vehiculated by Ceratocystus ulmi) and phloem necrosis have contributed to low biomass by eliminating all the large trees of Ulmus minor in Europe and Ulmus americana in the United States which dominated the canopy of fertile forested sites prior to 1960. However, the elm has not disappeared from the alluvial forests, remaining abundant in underlayers, thanks to an active vegetative regeneration.

The alluvial forests include other European tree species defined as competitive in upland forests, such as Acer pseudoplatanus, Tilia cordata, Carpinus betulus, Acer platanoides, and Tilia cordata. In an alluvial environment these species are rarely dominant (except Tilia cordata on high, gravelly terraces), essentially because of intolerance to summer floods.

A. SPECIES ABUNDANCE AND DIVERSITY

Alluvial forests of large river plains have a large array of species, particularly in sectors of high fluvial dynamics, when the plain becomes large and multiplies landforms and gradients: up to 157 species, including 56 woody elements (trees, shrubs, and lianas), in the upper Rhine valley (Schnitzler, 1988). Alpha diversity in the warm temperate floodplain forests is higher, with 59 woody species in the end-successional forests of the Mississippi alluvial plain (Robertson et al., 1978). Species richness decreases from upstream to downstream with changes in channel slope, topography, and fluvial forms. As an example, woody species richness is only 34 species of hardwoods in the downstream sectors of the Rhine (Dister, 1980), 19 in the Danube delta (Simon, 1960), and 19 in the Po delta (Piccoli & Gerdol, 1984). In the lower sectors of the Rhine, increasing hydromorphy and acidity of the substrates account for the fact that some species are progressively disappearing (Alnus incana, Cornus mas, Daphne mezereum, Populus nigra, Tilia platyphyllos, Euphorbia amygdaloides, and Helleborus foetidus) while others are appearing (Salix viminalis, Allium scorodoprasum, Asperula odorata, and Maianthemum bifolium).

Similar species and genera recur in the different alluvial forest communities of Europe. This confirms the importance of common edaphic conditions and physiological stresses that plants of riverine ecosystems share regardless of climatic differences (Brinson, 1990). On a broader scale, similar genera recur throughout the temperate area (Table III). The reference to the fossil species lists established by Geissert (1984) for the Pliocene Rhine alluvial forests shows their specific analogies with present-day alluvial forests of Europe, Asia Minor, and North America.

Table III Common genera of the alluvial forests in North America, Asia Minor, and Europe, in comparison with the alluvial flora of the Rhine during the Pliocene
 North America(a) Asia Minor(b)

Trees and shrubs
Acer negundo, rubrum, laetum, rubra,
 saccharinum insigna
Aesculus discolor, glabra
Alnus serrulata, glutinosa,
 rhombifolia, subcordata
 tenuifolia
Betula nigra
Carpinus caroliana betulus
Carya aquatica,
 cordiformis,
 glabra, laciniosa,
 ovalis, ovata,
 tomentosa
Celtis occidentalis, laevi australis
 gata, canadensis
Cornus amomum, florida, australis
 foemina, racemosa
Corylus avellana
Crataegus monogyna,
 ambigua
Diospyros virginiana lotus
Euonymus atropurpurea
Fagus grandifolia(*)
Fraxinus americana, angustifolia,
 caroliana, excelsior, ornus
 latifolia,
 pennsylvtica,
 tomentosa
Ilex decidua, opaca
Juglans capsica, hindsii
Liquidambar styraciflua
Liriodendron tulipifera
Morus nigra alba
Nyssa aquatica, sylvatica
Parrotia persica
Pinus taeda
Platanus occidentalis,
 racemosa
Populus angustifolia, capsica
 deltoides,
 fremontii,
 heterophyUa,
 sargentii
Prunus occidentalis, capsica,
 heterophylla, laurocerasus
 deltoides
Pterocarya serotina fraxinifolia
Quercus alba, lobata, castaneifolia
 lyrata,
 pagodaefolia,
 phellos, michauxii,
 bergii,
 mulhenbergi,
 nigra, palustris,
 shumardi, bicolor,
 imbricaria,
 velutina,
 macrocarpa,
 virginiana, rubra
Rhamnus grandifolia,
 frangula
Robinia pseudacacia
Rubus ulmifolius
Salix nigra, interior fragilis, micans,
 aegyptica
Sambucus canadensis ebulus
Sassafras albidum
Staphylea
Tamarix pentandra
Taxodium distichum
Tilia americana rubra, caucasica
Ulmus americana, rubra minor

Zelkova crenata

Woody lianas
 Hedera colchica
 Vitis aestivalis, hissarica,
 riparia, cinerea, silvestris
 rupestris

 Fossil genera in
 the Rhine valley
 Europe(c) (Pliocene)(d)

Trees and shrubs
Acer campestre, campestre,
 negundo,(**) platanoides,
 tataricum, monspessulanum
 monspessulanum,
 pseudoplatanus
Aesculus hippocastaneum spinosissimo
Alnus glutinosa, incana viridis, glutinosa
Betula pendula
Carpinus betulus betulus, minima
Carya angulata, moenana
Celtis
Cornus mas, sanguinea mas, sanguinea
Corylus avellana
Crataegus monogyna, nigra, sp.
 laevigata, degenii
Diospyros
Euonymus europaeus, latifolia
Fagus sylvatica(*) decurrens
Fraxinus excelsior
Ilex aquifolium aquifolium, cornuta
Juglans regia bergomensis
Liquidambar europaea
Liriodendron germinata
Morus
Nyssa disseminata
Parrotia persica, fagifolia,
 reidiana

Pinus sylvestris
Platanus occidentalis
Populus alba, nigra, 7 species
 tremula
Prunus avium, spinosa, girardii, spinosa,
 mahaleb, padus lansdorfii,
 laurocerasus
Pterocarya limburgensis
Quercus robur, pubescens hickelii, pubescens
Rhamnus catharticus,
 frangula, alaternus
Robinia pseudacacia(**)
Rubus ulmifolius, laticostatus
 caesius, fruticosus
Salix alba, eleagnos, 2-3 species
 daphnoides,
 triandra,
 purpurea,
 viminalis,
 nigricans, cinerea
Sambucus nigra, ebulus pulchella
Sassafras 1 species
Staphylea pinnata pliocaenica
Tamarix gallica(*) ramossisima, laxa
Taxodium dubium
Tilia cordata, platyphyllos
 platyphyllos
Ulmus minor, laevis, minor, longifolia
 glabra
Zelkova ungeri

Woody lianas
 Hedera helix helix
 Vitis silvestris parasilvestris,
 teutonica, ludwigii


(a) Brinson, 1990.

(b) Emberger & Sabeti, 1962.

(c) Schnitzler, 1988; Karpati & Toth, 1961.

(d) Geissert, 1984.

(*) Not regenerating.

(**) Introduced.

The species richness allows the coexistence of many species of relatively few genera (Schnitzler et al., 1992). In the Rhine valley, the 157 species occur within 113 genera. In the Mississippi valley, the genus Quercus includes 12 species; Acer, 3; Fraxinus, 2; and Ulmus, 2 in the end-successional stage (Robertson et al., 1978).

In complex land-forms in upstream and middle sectors of large river plains, the spatial segregation of closely related species becomes particularly fine-grained. In the upper Rhine valley, spatial segregation of the genus Salix (9 species without hybrids) provides an excellent opportunity to visualize the environmental gradients (Schnitzler et al., 1992). For Salix, selection between species occurs during germination (Ellenberg, 1988). All seeds brought to a new open environment by wind or floods can germinate, but they are quickly selected according to the geochemistry: the basic deposits favor Salix elaeagnos, S. daphnoides, and S. alba, and the subacidic deposits favor Salix triandra, S. viminalis, and S. fragilis. On the same deposit, then, species segregate along a textural gradient: bushes of narrow-leaved S. daphnoides and S. elaeagnos grow on gravels and face the current, whereas bushes of broad-leaved S. alba grow on the opposite face of the same deposit, where texture is enriched in sand or silt. In subacidic deposits, S. viminalis and S. triandra segregate from S. fragilis in the same way. Moreover, geochemistry and groundwater levels segregate the species close to the main river channel from those of stabilized areas. Along the main river, the distribution of Salix species gradually changes downstream because of the increasing hydromorphy and decreasing amount of limestone. Subacidic species replace basiphilous species. Hybrids of S. alba and S. fragilis (S. x rubens) are found in a broader range of sites, both in calcareous and subacidic sites.

In the same valley, Alnus glutinosa and A. incana and their hybrid A. x pubescens, and Populus alba and P. tremula and their hybrid P. x canescens, segregate mainly on the basis of geochemistry (Schalin, 1967). Acer and Ulmus genera, which both have three species in the Rhine valley, have strongly marked biogeographical characters. Acer campestre and Ulmus minor are thermophilous species, frequent in forests along the Rhine. Acer pseudoplatanus and Ulmus glabra are subatlantic, submontaneous species, sparse in all alluvial phytocoenoses. Acer platanoides and Ulmus laevis are both continental and central European species, which also are sparse in the Rhine forest communities. The six species coexist in the Rhine rift thanks to its climatic particularities (subatlantic with a continental tendency), segregating in the same genera according to a double gradient of fertility and texture (Schnitzler et al., 1992).

High total diversity is closely coupled with the river system. First, on a large scale, alluvial landscape presents a continuum from flooding-prone sites to particularly protected ones. Both extremes act on refuges and seed sources when the conditions change (White, 1979). That is why there is no regional species extinction but, rather, local elimination of species, either pioneer or climax; and additional allochtonous species are also brought by floods. Second, the instability of the landscape is a factor eliminating intense interspecific competition, because the temporal dynamics of spatial environmental gradients, plus chance events, maintain population variability. Third, the high fertility occurring in most habitats of the plain due to recurrent flooding is a positive factor for high species diversity and productivity as well as high coexistence potential.

B. COMMUNITY PATTERN AND FOREST DYNAMICS

The majority of forest communities described in large European plains occur in geographically separated forested basin wetlands, with minor floristic differences. Many detailed phytosociological and floristic descriptions of the alluvial forests of the large European rivers are found for the Danube forests in Wendelberger (1952) and Karpati and Toth (1961), for the Po in Sartori (1984), for the Rhine in Dister (1980), Carbiener (1983), and Schnitzler (1988, 1994a), and for the Loire in Schnitzler (1995a). Landscape diversity is high in large river plains such as the upper Rhine (Tables IV, V).

Table IV Patterns of landscape diversity and community richness in the Rhine and Loire valleys

 Total
 Community Family Genera species
 richness richness richness richness

RHINE VALLEY 12 56 113 157
 Rhine floodplain 8 53 107 143
 Ill floodplain 4 29 83 106
LOIRE VALLEY 8 47 98 110
 Loire floodplain 7 47 98 110
 Allier floodplain 7 47 98 110

 Total
 woody
 species Beta (1)(a) Beta (2)(b)
 richness diversity diversity

RHINE VALLEY 56 1.07 0.66
 Rhine floodplain 52 0.26 0.26
 Ill floodplain 37 0.98 0.98
LOIRE VALLEY 46 0.49 0.41
 Loire floodplain 46 0.48 0.41
 Allier floodplain 46 0.47 0.41


(a) Beta 1 = s/a-1 where s = total number of species and a = number of species in each of the samples.

(b) Beta 2 = g(H) + I(H)/2a where g is the gain of species and I is loss.

Table V Ecology (physiography and hydrology) and species diversity per vegetation type in the upper Rhine valley (Rhine and its tributary the Ill)
Phytosociology Physiography

RHINE
 Salici-Populetum nigrae river bank of low elevation

 Ligustro-Populetum nigrae gravelly terraces of high
 elevation

 Fraxino-Populetum albae silty terraces

 Querco-Ulmetum popule- silty-sandy low terraces
 tosum

 Querco-Ulmetum typicum silty-sandy low terraces

 Querco-Ulmetu tilietosum sandy high terraces

 Querco-Ulmetum sandy low terraces
 carpinetosum

Pruno-Quercetum silty Holocene terraces

ILL
 Carici elongatae-Alnetum clayey swamps

 Pruno-Praxinetum old channels

 Alno-Carpinetum clayey-silty low terraces

 Stellario-Carpinetum gravelly Wiirmian terraces

Phytosociology Flooding regime

RHINE
 Salici-Populetum nigrae 3-8 days several times a year in
 summer

 Ligustro-Populetum nigrae 1-2 days every 3-5 years in May-June

 Fraxino-Populetum albae 1-3 days every 3-5 years in May-June

 Querco-Ulmetum popule- 1-3 days every 3-5 years in May-June
 tosum

 Querco-Ulmetum typicum unflooded for 30 years

 Querco-Ulmetu tilietosum every 10 years in May-June

 Querco-Ulmetum unflooded for 150 years
 carpinetosum

Pruno-Quercetum major floods in summer

ILL
 Carici elongatae-Alnetum 5-6 months every year in spring/winter

 Pruno-Praxinetum 1-2 weeks in spring/winter

 Alno-Carpinetum 3-8 days in spring/winter

 Stellario-Carpinetum major floods in summer

 Mean Total
 Total species woody
 species richness species
Phytosociology richness in 0.1 ha richness

RHINE
 Salici-Populetum nigrae 81 18.4 31

 Ligustro-Populetum nigrae 63 28.6 35

 Fraxino-Populetum albae 111 28.9 47

 Querco-Ulmetum popule- 99 31.2 44
 tosum

 Querco-Ulmetum typicum 92 33.7 43

 Querco-Ulmetu tilietosum 90 34.4 44

 Querco-Ulmetum 84 34.8 41
 carpinetosum

Pruno-Quercetum 77 37.1 39

ILL
 Carici elongatae-Alnetum 51 18 14

 Pruno-Praxinetum 85 27.1 32

 Alno-Carpinetum 87 35 32

 Stellario-Carpinetum 41 16.6 12

 Mean
 woody
 species
 richness Shannon
Phytosociology in 0.1 ha index H'

RHINE
 Salici-Populetum nigrae 4.1 1.8-2.9(*)

 Ligustro-Populetum nigrae 10.6 2.1-2.7(*)

 Fraxino-Populetum albae 15.9 1.9-2.1(*)

 Querco-Ulmetum popule- 17.8 3.4
 tosum

 Querco-Ulmetum typicum 15.7 3.4

 Querco-Ulmetu tilietosum 17.6 2.6

 Querco-Ulmetum 16.7 3.1
 carpinetosum

Pruno-Quercetum 16.1 3.2

ILL
 Carici elongatae-Alnetum 4.1 0.8

 Pruno-Praxinetum 10 2.7

 Alno-Carpinetum 13.7 2.8

 Stellario-Carpinetum 4.3 0.8


(*) Range of values for young and old forest communities.

The particularities of the microrelief and the local distribution of protected and unprotected sites, as well as the kinetic energy level of floods, generate patches of forest units. Vegetation varies from communities adapted to long hydroperiods, in swales and old channels, to communities located in relatively high terraces on the floodplain, some of which are not regularly flooded. In both extremes of the moisture gradient, forest communities have low structural development and relatively low species richness, attributed to water stress. This is the case for the swamp forests (Alnetum glutinosae) and the hardwoods of the gravelly, unflooded terraces of the center of the plain (Stellario-Carpinetum) defined in the Rhine valley (Table V).

Each of these communities shows a specific sylvatic mosaic, composed of a variable spatial pattern of the different stages of the successions. Oldeman (1990) defined five different steps:

1. The innovation stage develops on a large, empty site caused by a severe disturbance. In alluvial plains, disturbance consists mainly in floods of high kinetic energy. This stage is characterized by patches of pioneer herbs, shrubs, or tree saplings.

2. The pre-equilibrium stage appears with the closure of the canopy, after about 50-100 years in upland, stable environments. In alluvial plains, the closure of the canopy may be shortened to 10-15 years.

3. The equilibrium stage is the "climax" stage, which represents ecosystem maturity. Internal functioning is complex and highly structured, fragmented in complex mosaic units (units of regeneration or gaps, units of maturity, units of decay or death) that are of small dimension and structurally dynamic. The internal organization of the equilibrium stage is clearly separated from the environment. In temperate, stable forests, the average time to reach this level is about 250-300 years (Jones, 1945). This stage may persist a long time (several thousand years) if no severe accident occurs.

4. The elimination stage appears after a stress that eliminates many units at the same time. Rather severe floods, when frequent after a change in the course of the river, may be another common stress in the alluvial plains. The stress can also be of biotic origin, such as the present-day extension of Dutch elm disease, which has caused since the beginning of the century the death of many Ulmus species in alluvial forests [e.g., 8 elms per ha for a total surface of about 40 ha in alluvial forests of the Morava (Korpel, 1995)].

5. The collapse stage, during which nothing is left but bare soil, leads to a new installation stage. Paleogeomorphological studies of the Rhine valley (Striedter, 1988) proved that great floods, which destroyed all the forest formations of the entire valley, occurred about every 500-800 years during the Holocene. This could explain why the oldest trees of fossil trunks of Quercus found in the Rhine valley never exceed 500 years. This relatively short periodicity of destructive floods was still higher along the main river (every century for the big European river plains).

In alluvial environments, the forests present several particularities. First, the forest-environment separation is not very clear, since floodwaters regularly pass through the system, bringing new propagules, sediments, water, and nutrients and transporting organic matter and seeds elsewhere. Secondly this stage may be reached in less than 150 years.

In alluvial forests, self-perpetuation through gap dynamics is promoted by flood action (transport of trees or destabilization and uprooting) as well as by biotic causes (disease, senescence). However, patterns of tree mortality vary regionally in the plain because the relative importance of catastrophic agents varies widely with forest type and the situation in the floodplain (influence of river energy). Unfortunately, the paucity of undisturbed bottomland hardwood stands now existing in Europe have made it difficult for researchers to deduce the relative impacts of uprooting and piecemeal disintegration by physiological factors.

Gap dynamics were studied in alluvial forests of the Rhine valley by Walter (1982), Beekman (1984), and Koop (1989). When small, gaps are rapidly colonized by pioneer trees and shrubs, often extending through vegetative reproduction. Clonal structures are numerous in gaps, with stands of Populus, Ulmus, Alnus, Prunus, Crataegus, Viburnum, Ligustrum, Cornus, and Salix. When larger, gaps are colonized by dense blankets of Clematis vitalba (and Vitis sylvestris before phylloxera) which covers the ground and remnant shrubs and climbs to the edges and into the trees. Regeneration is slowed by the blankets of lianas, which prevent the closure of the canopy for decades.

C. THE PARTICULAR CASE OF RIVERINE ECOSYSTEMS OF HIGH-DYNAMICS RIVERS

Along a big river, vegetation must cope with flooding severity (i.e., frequency and kinetic energy of flowing waters), a factor that depends essentially on distance from the river channel and elevation. In this zone there are two major units: softwoods and hardwoods. This denomination is a direct translation from the German Weichholz and Hartholz (terms related to wood density) as softwoods (for the first steps of the successions: Salici-Populetum, Ligustro-Populetum) and hardwoods (for the equilibrium stage: Querco-Ulmetum).

The installation stage near the river is composed of various patterns of shrubs (Salix viminalis, S. triandra, S. eleagnos, S. purpurea, or tree saplings of Populus nigra and Salix alba or S. fragilis). The pre-equilibrium stage occurs with the establishment of two forest units: Salici-Populetum in moist, low sites, and Ligustro-Populetum nigrae in dry, elevated terraces.

Salici-Populetum and Ligustro-Populetum have similar features, the only difference being the dominance of Salix or Populus in the canopy (Schnitzler, 1988, 1995b). These units are characterized by two main strata: the canopy (about 20 m high) and the grass layer (up to 2 m high). The Shannon index varies from 2 to 2.7. Shrubs are relatively rare on wet sites and in young stands. The canopy is even-aged and composed of pioneer species: Salix alba and Populus nigra, sometimes Populus alba or Alnus incana. Trees never exceed 60 cm dbh. The ratios of Salix alba and Populus nigra depend on the degree of moisture. The shrubs are mainly remnant individuals of the previous bush stage (Salix eleagnos or Salix viminalis). In gaps, Humulus lupulus and Convolvulus sepium form small veils. The grass layer is composed of high clones of Urtica dioica, Impatiens glandulifera, Phalaris arundinacea, and other nitratophilous, heliophilous species. The total density of this association varies with flooding conditions. In the Rhine valley, the total density varies from 1212 stems per [ha.sup.-1] when regularly inundated and dominated by Salix alba to 4900 stems per [ha.sup.-1] in terraces that are rarely flooded and are dominated by Populus nigra. Average basal area is 20 [m.sup.3] [ha.sup.-1]. When old, moist Salici-Populetum (dominated by Salix alba) keeps similar total woody densities, with some giant old willows (up to 160 cm dbh). High grasses are still abundant and surround small patches of Rosa canina, Salix purpurea, S. viminalis, or S. eleagnos. In Salici-Populetum dominated by Populus nigra, total density decreases to less than 2000 stems per [ha.sup.-1] and the woody species richness increases. In both stands, the presence of giant old trees explains the high basal area: from 80 to 90 [m.sup.2] [ha.sup.-1].

Querco-Ulmetum is characterized by five well-defined strata. The Shannon index varies from 3 to 3.4. The total density is 5400 stems per [ha.sup.-1], which corresponds to an average along a gradient of flooding recurrence. In regularly flooded sites, the average density is 2919 stems per [ha.sup.-1]. When flooded every 30 years, seedlings are less often killed by floods, and the density increases to as much as 5081 stems per [ha.sup.-1]. After a long period without flooding (about 150 years), the density reaches 6830 stems per [ha.sup.-1] (Schnitzler, 1994b). The average basal area of the Querco-Ulmetum is 32 [m.sup.2] [ha.sup.-1]. The canopy is uneven and relatively open (about 38 stems per [ha.sup.-1]), with trees ranging from 25 to 30 m in height and some emergent individuals of 35 m height and up to 120 cm dbh (Fig. 1). The dominant trees are Quercus robur and Fraxinus excelsior, with 51% and 26% of the forest canopy, respectively. Ulmus minor, the third component of the canopy, is being progressively eliminated because of elm disease. Populus alba and Populus nigra, with a scattered distribution, are often relicts of earlier successional episodes. Unevenness of the canopy, coupled with small leaf size, favors penetration of light along trunks and the ground and accounts for the abundance of giant lianas (Clematis vitalba, Hedera helix, Vitis sylvestris before phylloxera) in the upper tree layers, the higher proportion of giant shrubs in the secondary tree layer (Cornus mas, Cornus sanguinea, Corylus avellana, Crataegus monogyna, Evonymus europaeus, Prunus spinosa, Sambucus nigra, and Viburnum lantana), and high underlayer densities (average of 5000 stems per [ha.sup.-1]). In the Loire valley, Humulus lupulus and Evonymus europaeus reach large sizes in Querco-Ulmetum, whereas all the calcicolous species cited above are rare or absent.

The high species richness and complexity of Querco-Ulmetum explain its strong phenological individuality. This forest community flowers in an interrupted phenological sequence over a period of 9 months, from early January to early September (Schnitzler & Carbiener, submitted), which is a long time in comparison with other forest systems of temperate regions [no more than 7 months, according to Ellenberg (1988)]. Aerial surveys of hardwood forests during flowering and fruiting peaks in April-June and October could provide a crucial overall view of canopy structure from contrasting foliage coloring and irregularity in this degree of openness with species population patchiness.

Another originality of gallery forests is liana richness. Giant lianas are particularly abundant in Querco-Ulmetum mature stands, when regularly flooded (20% of the total individuals of the canopy in the Rhine valley). Old softwoods, when dominated by Populus nigra, are liana-rich as well. In tributaries, the proportion of giant lianas is low because of the presence of below-hydromorphic horizons, to which liana roots are very sensitive, and the abundance of shade species such as Carpinus betulus (Table VI; Schnitzler, 1995c).

Table VI Structural characteristics of canopy lianas in the old stands of the Rhine valley
 Giant
 Number lianas
 of giant (%)
 lianas in the
Forest type ([ha.sup.-1]1) canopy

Flooded Querco-Ulmetum populetosum 28.6 30
Salici-Populetum populetosum 24.9 6.6
Unflooded Querco-Ulmetum populetosum 56 42.7
Alno-Carpinetum 4.4 4.2
Ligustro-Populetum 120 11.4
Pruno-Quercetum 18.9 17.7
Querco-Ulmetum tilietosum 20.8 24
Fraxino-Populetum albae 33.3 15.3
Querco-Populetum carpinetosum 6.3 7.3

 Basal area
Forest type ([m.sup.2] [ha.sup.-1])

Flooded Querco-Ulmetum populetosum 0.03
Salici-Populetum populetosum 0.78
Unflooded Querco-Ulmetum populetosum 0.38
Alno-Carpinetum 0.02
Ligustro-Populetum 1.13
Pruno-Quercetum 0.1
Querco-Ulmetum tilietosum 0.12
Fraxino-Populetum albae 0.1
Querco-Populetum carpinetosum 0.04

Forest type Dominant species

Flooded Querco-Ulmetum populetosum Hedera helix
Salici-Populetum populetosum Clematis vitalba
Unflooded Querco-Ulmetum populetosum Hedera helix
Alno-Carpinetum Hedera helix
Ligustro-Populetum Hedera helix
Pruno-Quercetum Hedera helix
Querco-Ulmetum tilietosum Clematis vitalba
Fraxino-Populetum albae Hedera helix
Querco-Populetum carpinetosum Clematis vitalba


The Querco-Ulmetum area is vast, covering three big climatic areas in Europe: west Mediterranean Europe, continental Europe, and central Europe (for a phytosociological synthesis of central Europe, see Seibert, 1987). Vicariant alluvial hardwood forests evolving in a warmer and moister climate exist in Asia Minor [extending around the Black Sea, the Caspian Sea, and the Aral Sea (Emberger & Sabeti, 1962)] and in North America [e.g., along the Mississippi River (Hosner & Minckler, 1963; Carter-Johnson et al., 1976; Robertson et al., 1978). These forests are more species-rich (including Tertiary relics such as Parrotia, Zelkova, Pterocarya, Celtis, Gleditsia, Diospyros, Liquidarnbar, Liriodendron, and Nyssa genera, as well as more species per genera) than the European alluvial hardwoods for two main reasons: the presence of refuge sites during the Quaternary glaciations, which did not occur in Europe, and the present-day warmer climate (with annual temperatures of 14 [degrees] C and mean annual rainfall of 1000 mm in the Mississippi valley, e.g.). European, Asian, and American alluvial hardwoods all originate from the warm temperate, dense forests of the end of the Pliocene (Table III).

The forested pattern of softwood/hardwood in big river plains has been variously interpreted. For some authors (e.g., Wendelberger, 1952; Moor, 1958; Passarge, 1985), the softwood/hardwood mosaic is guided mainly by textural and moisture gradients (linked to river dynamics) and by flooding frequency (linked to floodplain elevation and distance from the main channel). For others, the pattern depends essentially on temporal processes, and softwoods and hardwoods are interpreted as different steps in the same succession. The basis of this interpretation was the frequent observation that Populus and Salix species predominate in young stands and then are replaced by Ulmus, Acer, and Quercus species over time (for the Missouri River, see Hosner & Minckler, 1963; Wilson, 1970; and Robertson et al., 1978; for the Danube, see Margl, 1973; for the Rhine, see Carbiener et al., 1988; Koop, 1989; and Schnitzler, 1995c) (Fig. 1).

The causes of these different interpretations are due to the fact that there exist old softwood communities in the alluvial floodplains, particularly in downstream sectors. For the proponents of the succession hypothesis, old softwoods are the result of extreme conditions that have disturbed the progress of succession, arresting it at different stages of development. Another cause was that species composition and structure are so dramatically different between the steps of the succession. In Europe, phytosociologists classify them in different associations and even in different higher phytosociological ranks (up to class level).

However, the temporal links between forest units are the key to understanding the forest pattern along the longitudinal profile of large rivers. If inundations are frequent and severe, the time required for developing the succession is too short and forest communities are perpetually in a state of effective non-equilibrium. That is why softwoods predominate the forested landscape in upstream sectors. With the increasing possibilities of protected sites developing downstream, perpetually young softwoods are relegated to the active braids (where they occupy only 10% of the total forested landscape), whereas hardwoods dominate the landscape. In lower sectors, softwoods survive only in extreme conditions of wetness or drought where they can age and become permanent communities. The proportion of old softwoods is about 20-40%.

The successional softwood/hardwood links explain also why the forest pattern of the big valleys has changed during prehistoric times with the changes in climatic conditions. The history of fluvial dynamics with identification of high-disturbance periods has been studied in the upper Rhine, Main, and Danube through geomorphological profiles of terraces and tree-ring analyses of fossil trunks (Becker, 1982; Striedter, 1988). In the upper Rhine, Striedter (1988) determines six major periods of widespread flooding increase and maximum extension of gravel deposition, from 9000 years b.p. to the beginning of the 20th century, during which a maximum of trees were pulled out and sedimented. The maximum stages of flooding and lateral-changing river courses occur in the middle of Subatlantic times. These periods of intensive disturbance have favored the extension of pioneer communities in a great part of the plain, which explains the high number of fossil trunks of Populus, Salix, and Alnus found in gravel pits. In periods of relatively low disturbance, the main fossil trunks found in the gravel pits were Ulmus, Quercus, and Fraxinus in the Rhine valley.

VI. Conclusion

The role of river dynamics on forest functioning was described through some characteristic elements of the nutrient cycle, species composition, and forest structure and dynamics. The recognition of these fundamental features makes it difficult to apply here the concept of climax, considered traditionally as the presumed result of autogenesis within a stable physical environment. But if we consider the high level reached by hardwood forests, we can admit that this particular forest has reached a climax level.

The problem of separation between strictly endogenous and exogenous categories of disturbance factors was discussed in depth by White (1979), who posited that physical events, when recurrent and predictable, lead to selection of species and communities for adaptation and must hence be considered an intrinsic part of forest dynamics. According to this concept, Querco-Ulmetum is a climax state, reached after deep transformations in its species composition and structure. For White (1979), truly exogenous disturbances are those of very low frequency or those that create new, irreversible situations to which species are not adapted. Human disturbance in alluvial landscape corresponds to the latter case. Man has separated the forests from the river, physically destroyed them, and interfered with the successional and maturation processes. Nowadays all temperate alluvial plains are altered to varying degrees, especially in wealthy countries. As a result, many forests have reached the absolute biological limit of adaptation and have lost many of their original and aesthetic characteristics. Old-growth alluvial stands are nowadays very rare: in Europe, the only primeval alluvial hardwood is situated in the Lanzhot Reserve, situated at the confluence between the Dyje and the Morava Rivers in the Slovak Republic (Korpel, 1995). We are now becoming conscious of this immense loss, and some scientific programs propose restoration measures. Unfortunately, it is rare that we see proposals for conservation, restoration, extension, and re-inundation over wide areas (at least 1000 ha: Schnitzler, 1994b) of complex alluvial ecosystems, evolving in a natural environment with a free course of the river in the floodplain, in order to reconstitute the complete sylvigenesis, with all forest successional steps and all sylvigenetic phases. In France, the best site for creating such a natural forested reserve is the Loire valley, because the river still has a quite natural course, and many alluvial lands are nowadays abandoned by agriculture (Schnitzler, 1995a). But there are probably other adequate sites in Europe and North America.

VII. Literature Cited

Amoros, C. & G. E. Petts. 1993. Hydrosystemes fluviaux. Masson, Paris.

Becker, B. 1982. Dendrochronologie und Palaeookologie subfossiler Baumstamme aus Flussablagerungen. Ein Beitrag zur Nacheiszeitlichen Auenentwicklung im Sudlichen Mitteleuropa. Ph.D. thesis, Stuttgart University.

Beekman F. 1984. La dynamique d'une foret alluviale et le role des lianes. Pages 475-502 in J.-M. Gehu (ed.), La vegetation des forets alluviales. Colloques phytosociologiques 9. J. Cramer, Berlin.

Bomer, B. 1971. Les iles de la Loire, Evolution et stabilite. Etudes Ligeriennes. Pages 70-79 in Symposium Le lit de la Loire et ses iles, 20-21 November 1971. Le courrier d'aquitaine 11.

Brinson, M. M. 1990. Riverine forests. Pages 87-143 in A. E. Lugo, M. Brinson & S. Brown (eds.), Forested wetlands. Ecosystems of the world 15. Elsevier, Amsterdam.

Brzeziecki, B. & F. Kienast. 1994. Classifying the life-history strategies of trees on the basis of the Grimian model. Forest Ecol. Managem. 69(1-3): 167-187.

Carbiener, R. 1970. Un exemple de type forestier exceptionnel pour l'Europe occidentale: la foret du lit majeur du Rhin au niveau du fosse rhenan (Fraxino-Ulmetum Oberd. 53.). Interet ecologique et biogeographique, comparaison a d'autres forets thermophiles. Vegetatio 20(1-4): 97-148.

--. 1981. Der Beitrag der Hutpilze zur soziologischen und synokologischen Gliederung von Feuchtenwaldern. Pages 557-590 in International Symposium, Struktur und Dynamik von Waldern. J. Cramer, Berlin.

--. 1983. Le Grand Ried Central d'Alsace: ecologie et evolution d'une zone humide d'origine fluviale rhenane. Bull. Ecol. 14: 249-277.

--. 1984. Resume de quelques aspects de l'ecologie des complexes forestiers alluviaux d'Europe. Pages c-i in J.-M. Gehu (ed.), La vegetation des forets alluviales. Colloques phytosociologiques 9. J. Cramer, Berlin.

-- & A. Schnitzler. 1990. Evolution of major pattern models and processes of alluvial forests of the Rhine in the rift valley (France/Germany). Vegetatio 88: 115-129.

-- & M. Tremolieres. 1990. The Rhine rift valley groundwater-river interactions: Evolution of their susceptibility to pollution. Regulated Rivers 54(5): 375-389.

--, A. Schnitzler & J. M. Walter. 1988. Problemes de dynamique forestiere et de definition des stations en milieu alluvial. Pages 655-686 in J.-M. Gehu (ed.), Phytosociologie et foresterie. Colloques phytosociologiques 14. J. Cramer, Berlin.

Carter-Johnson, W., L. Robert, R. Burgess, R. Warren & R. Keammerer. 1976. Forest overstory vegetation and environment on the Missouri River floodplain in North Dakota. Ecol. Monogr. 46: 59-84.

Corillion, R. 1995. Le lit mineur de la Loire; Climatologie et vegetation. Technical Report, Academie d'Agriculture, Angers.

Decamps, H. & R. Naiman. 1989. L'ecologie des fleuves. Recherche 20(208): 310-319.

Dister, E. 1980. Zur Struktur und Dynamik alter Hartholzauenwalder (Querco-Ulmetum Issl. 24) am nordlichen OberRhein. Verh. Zool.-Bot. Ges. Osterreich. 123:13-21.

--. 1983. Zur Hochwassertoleranz yon Auenwaldbaumen an lehmigne Standorten. Verh. Ges. Okol. 10: 325-336.

Ellenberg, H. 1988. Vegetation ecology of central Europe. Cambridge University Press, Cambridge.

Emberger, L. & H. Sabeti. 1962. Forets denses intertropicales et forets caspiennes humides. Naturalia Monspel., Ser. Bot. 14: 55-61.

Franklin, J. F., K. Cromack, W. Denison, A. McKee, C. Maser, C. Sedell, F. Swanson & G. Juday. 1981. Ecological characteristics of old-growth Douglas-fir forests. USDA Forest Service General Technical Report. Washington, DC.

Frye, R. J. & J. W. Robinson. 1979. Change in the channel of the Lower Missouri River and effects on fish and wildlife. Aquatic Series 11. Missouri Department of Conservation, Jefferson City.

Geissert, F. 1984. Fossile Floren aus der Rheinebene im nordlichen Elsass (Pliozan-Pleistozan-Holozan). Pages 453-474 in J.-M. Gehu (ed.), La vegetation des forets alluviales. Colloques phytosociologiques 9. J. Cramer, Berlin.

Grelon, J. 1971. Contribution a une etude ecologique et dynamique de la vegetation des greves et des iles de la Loire, a Vouvray (Indre- et Loire) et a St-Jean de-la-Croix (Maine-et-Loire). Ph.D. thesis, Orsay University.

Grime, J. P. 1977. Evidence for the existence of three primary strategies in plants and its relevance to ecological and evolutionary theory. Amer. Naturalist 111: 1169-1194.

Hall, T. H. & G. E. Smith. 1955. Effects of flooding on woody plants, West Sandy Dewatering Project, Kentucky Reservoir. J. Forest. 53: 281-285.

Harmon, M. E., J. F. Franklin, F. J. Swanson, P. Sollins, S. V. Gregory, J. D. Lattin, N. H. Anderson, S. P. Cline, N. G. Aumen, J. R. Sedel, C. W. Llenkaemper, K. Cromack & K. W. Cummins. 1986. Ecology of coarse woody debris in temperate ecosystems. Adv. Ecol. Res. 15: 133-302.

Hosner, J. F. 1960. Relative tolerance to complete inundation of fourteen bottomland tree species. Forest Sci. 6(3): 246-251.

-- & L.S. Minckler. 1963. Bottomland hardwood forests of southern Illinois: Regeneration and succession. Ecology 44: 29-41.

Jones, E. W. 1945. The structure and reproduction of virgin forest of the north temperate area. New Phytol. 13(44): 130-142.

Karpati, I. & I. Toth. 1961. Die Auenwaldtypen Ungarns. Acta Agron. Acad. Sci. Hung. 11: 421-452.

Koop, H. 1987. Vegetative reproduction of trees in some European natural forests. Vegetatio 72: 103-110.

--. 1989. Forest dynamics: Silvi-Star, a comprehensive monitoring system. Springer-Verlag, New York.

Korpel, S. 1995. Die Urwalder der Westkarpaten. Gustav Fischer, Stuttgart.

Loiseau, J. E. & J. C. Felzines. 1992. Variations du peuplement vegetal alluvial constatees dans la partie moyenne du bassin ligerien en 1990 et 1991. Monde Pl. 445: 14-16.

Lugo, A. E., M. Brinson & S. Brown (eds.). 1990. Forested wetlands. Ecosystems of the world 15. Elsevier, Amsterdam.

Michael, E. & B. Hennig. 1964-1970. Handbuch fur Pilzsfreunde. Fischer, Jena.

Margl, H. 1973. Pflanzengesellschaften un ihre Standortgebundene Verbreitung in teilweise abgedammten Donauauen (unter Lobau). Verh. Zool.-Bot. Ges. Wien 113: 5-51.

Moor, M. 1958. Pflanzengesellschaften schweizerischer Flussauen. Mitt. Schweiz. Anst. Forstl. Versuchswesen 34: 221-360.

Oldeman, R. A. A. 1990. Forests: Elements of silvology. Springer-Verlag, New York.

Passarge, H. 1985. Phanerophyten Vegetation der markischen Oderaue. Phytocoenologia, 13(4): 505-603.

Pautou, G., J. Girel & L. Maman. 1986. Le role des processus allogeniques dans le deroulement des successions vegetales: l'exemple de la plaine alluviale du Rhone entre Geneve et Lyon. Pages 655-689 in J.-M. Gehu (ed.), Vegetation et geomorphologie. Colloques phytosociologiques 13. J. Cramer, Berlin.

Penka, M., M. Vyskot, E. Klimo & F. Vasicek. 1985. Floodplain forest ecosystem. Elsevier, New York.

Piccoli, F. & R. Gerdol. 1984. Typology and dynamics of a wood in the Po plain (North Italy): The "Bosco della Mesola." Pages 161-170 in J.-M. Gehu (ed.), La vegetation des forets alluviales. Colloques phytosociologiques 9. J. Cramer, Berlin.

Robertson, P. A., G. T. Weaver & J. A. Cavanaugh. 1978. Vegetation and tree species patterns near the northern terminus of the southern floodplain forest. Ecol. Monogr. 48: 249-267.

Sanchez-Perez, J. M., M. Tremolieres & R. Carbiener. 1991. Une station d'epuration naturelle des phosphates et nitrates apportes par les eaux de debordement du Rhin: la foret alluviale frene et orme. Compt. Rend. Hebd. Seances Acad. Sci. 312(III): 395-412.

--, --, A. Schnitzler & R. Carbiener. 1993. Nutrient content in alluvial soils submitted to flooding in the Rhine deciduous forest. Acta Oecol. 14(3): 371-387.

Sartori, F. 1984. Les forets alluviales de la basse vallee du Tessin, Italic du Nord. Pages 201-216 in J.-M. Gehu (ed.), La vegetation des forets alluviales. Colloques phytosociologiques 9. J. Cramer, Berlin.

Schalin I. 1967. Germination analysis of Alnus incana (L.) Moench and Alnus glutinosa (L.) Gaertn. seeds. Oikos 18: 253-260.

Schnitzler, A. 1988. Typologie phytosociologique, ecologie et dynamique des forets alluviales du complexe geomorphologique ello-rhenan (plaine rhenane centrale d'Alsace). Ph.D. thesis, Strasbourg University.

--. 1994a. Alluvial hardwood forests of the temperate area. J. Biogeogr. 21: 605-623.

--. 1994b. Conservation of biodiversity in alluvial hardwood forests of the temperate zone. The example of the Rhine valley. Forest Ecol. Managem. 68(2-3): 385-398.

--. 1995a. Les forets alluviales des lits majeurs de l'Allier et de la Loire moyenne entre Villeneuve/Allier et Charite/Loire. Etude phytosociologique, diagnostic de naturalite et propositions de renaturation. Internal report, World Wild Life.

--. 1995b. Successional status of trees in gallery forest along the river Rhine. J. Veg. Sci., 6: 479-486.

--. 1995c. Community ecology of arboreal lianas in gallery forests of the Rhine valley, France. Acta Oecol. 16(2): 219-236.

--. 1996. Comparison of landscape diversity in forests of the upper Rhine and the middle Loire floodplains (France). Biodiversity & Conservation 5: 743-758.

-- & R. Carbiener. Submitted. Flowering patterns of a temperate alluvial hardwood forest over eight years.

--, -- & M. Tremolieres. 1992. Ecological segregations between closely related species in the flooded forests on the upper Rhine plain. New Phytol. 121: 293-301.

--, M. Tremolieres & J. M. Sanchez-Perez. Submitted. Changes in foliar nutrient concentration and translocation efficiency after elimination of floods in Fraxinus excelsior, Ulmus minor and Clematis vitalba.

Schwabe, A. 1991. Zur Wiederbesidlung von Auenwald-Vegetations-komplexen nach Hochwasser-Ereignissen: Bedeutung der Diasporen-Verdriftung, der generativen und vegetativen Etablierung. Phytocoenologia 20: 65-94.

Sedell, J. R. & J. L. Frogatt. 1984. Importance of streamside forests to large rivers: The isolation of the Willamette River, Oregon, USA from its floodplain by snagging and streamside forest removal. Int. Vereinigung Theor. Limnol. Verh. 22: 1828-1834.

Seibert, P. 1987. Der Eichen-Ulmen-Auwald (Querco-Ulmetum Issl. 24) in Suddeutschland. Natur & Landschaft 9: 347-352.

Simon, T. 1960. Contributions a la connaissance de la vegetation du delta du Danube. Ann. Univ. Scient. Budapest. Rolando Eotvos, Sect. Biol. 3: 307-333.

Striedter, K. 1988. Le Rhin en Alsace du Nord au Subboreal. Genese d'une terrasse fluviatile holocene et son importance pour la mise en valeur de la vallee. Bull. Assoc. Quatern. 1: 1-50.

Tabacchi, E. & A. M. Planty-Tabacchi. 1990. Evolution longitudinale de la vegetation du corridor de l'Adour. Botanica pirenaico-cantabrica, 455-468.

Tremolieres, M. & R. Carbiener. 1985. Quelques aspects des interactions entre litieres forestieres et ecosystemes aquatiques. Rev. Ecol. 40: 435-449.

--, --, A. Exinger & J. C. Turlot. 1988. Un exemple d'interaction non competitive entre especes ligneuses: le cas du lierre arborescent (Hedera helix L.) dans la foret alluviale. Acta Oecol. 92: 187-209.

Tricart, J. 1977. Precis de geomorphologie. Sedes, Paris.

Triska, F. J. 1984. Role of wood debris in modifying channel geomorphology and riparian areas of a large lowland river under pristine conditions: A historical case study. Int. Vereinigung Theor. Limnol. Verh. 22: 1876-1892.

Van de Winckel, R. 1984. Le Wyhlerwald, l'architecture et la dynamique d'une foret alluviale rhenane sauvage. Pages 503-542 in J.-M. Gehu (ed.), La vegetation des forets alluviales. Colloques phytosociologiques 9. J. Cramer, Berlin.

Walter, J. M. 1982. Architectural profiles of flood-forests in Alsace. Pages 187-234 in Berichte der Internationalen Symposien der Internationalen Vereinigung fur Vegetationskunde. J. Cramer, Berlin.

Weiss, D., R. Carbiener & M. Tremolieres. 1991. Biodisponibilite comparee du phosphore en fonction des substrats et de la frequence des inondations dans trois forets alluviales rhenanes de la plaine d'Alsace. Compt. Rend. Hebd. Seances Acad. Sci. 313(III): 245-251.

Wendelberger, E. 1952. Die Vegetation der Donauauen bei Wallsee. Ph.D. thesis, Linz University.

Wendelberger, G. 1984. Auenwalder der Donau. Pages 19-54 in J.-M. Gehu (ed.), La vegetation des forets alluviales. Colloques phytosociologiques 9. J. Cramer, Berlin.

White, P. 1979. Pattern, process, and natural disturbance in vegetation. Bot. Rev. (Lancaster) 45: 229-299.

Wilson, R. 1970. Succession in stands of Populus deltoides along the Missouri River in southeastern South Dakota. Amer. Midl. Naturalist 83:331-344.

ANNIK SCHNITZLER Laboratoire de Phytoecologie Faculte des Sciences Ile du Saulcy F-57045 Metz, France
COPYRIGHT 1997 New York Botanical Garden
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1997 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Schnitzler, Annik
Publication:The Botanical Review
Date:Jan 1, 1997
Words:10676
Previous Article:Roles of flavonoids in symbiotic and defense functions in legume roots.
Next Article:Sexual specialization and inbreeding avoidance in the evolution of dioecy.
Topics:

Terms of use | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters