Assessment of Drought-Tolerant Provenances of Austria’s Indigenous Tree Species

Abstract

(1) Background: Forestry will have to react to climate change because many tree species suffer. Mitigation can be realized either by planting non-native trees from regions with high climatic stress or by utilizing native tree provenances already adapted to stressful environments. Non-native trees have often generated problems in the past due to uncontrolled invasiveness. The use of native trees pre-adapted to the prospective climatic conditions is far less risky for the respective ecosystems. We offer a tool for selecting ecotypes of native trees as provenances for future forestry. (2) Methods: We propose the selection of tree species native to Middle Europe from a database of vegetation relevés of ± natural forest stands. By calculating the mean ecological indicator values of stands from their vegetation, cover sites can be elected that can provide seeds of provenances well adapted to future climatic conditions. (3) Results: By selecting the 10% partition of the most extreme stands of European tree species, seeds can be sampled and propagated for re-cultivating forests fit for future climate. (4) Conclusions: One can expect ecotypes of tree species that grow well on dry sites, since generations have faced evolutionary selection, for survival under stressful environments. This approach helps to avoid ecological risks of non-native trees.

1. Introduction

Climate change dynamics will have a strong impact on forest ecosystems and timber production [1,2]. Due to the most recent IPCC report, hot extremes as well as heavy precipitation will cause considerable damages to the world’s ecosystems [3]. In managed as well as natural forests, increased temperature and lower rainfall will induce problems for the survival of several tree species [4,5,6]. As an example, the managed forests at lower and medium altitudes of Europe dominated by Norway spruce or European beech suffer seriously from the consequences of more frequent droughts [7,8]. Consequently, individual trees are more prone to being attacked by pest insects (i.e., bark beetle) or pathogens [9,10].

Specific problems arise for European trees with increasing mean annual temperatures (plus 1.7–1.9 °C, compared to pre-industrial times [11]) and raising water stress. In the last two decades, summer droughts have led to serious impacts on European forests that ended up with extensive dieback in spruce, pine and beech forests [2,12,13]. The main problem of the trees under drought stress is hydraulic dysfunction and subsequent deficit of carbon storage [14]. The capacity to manage drought periods might be increased in populations that experience drought relative to their species specific normal physiological range. Tree species that arrived to Middle Europe in the postglacial period have occupied various habitats that represent the range of site conditions that the respective species were able to conquer under the given climate and the given competitors. Human influence on this process was rather indirect in that man changed habitat structure and soil by logging but ignoring the need of regeneration [15]. Reforestation—if ever strived for in former times—was not so much aiming at high quality timber. During the period of optimizing the quantity and quality of wood higher attention was led on reforestation with provenances that grew fast and gave high yield [16]. The general use of very productive provenances on sites of various water supply has led to the problem of extensive diebacks of cultivated forests along with dry summers during the last two decades [12,13].

In some tropic regions, the strategies of forestry tend to change by fostering native tree regeneration and gaining finally a rather multifunctional forest ecosystem with differentiated options of human utilization [17,18]. The idea to preferably use native tree species for the transformation of forests according to climate change was also considered in temperate regions [19]. The use of native species is ecologically sound and does not include risks of planting non-native trees that might become invasive and alter ecosystem services [20,21]. Native trees have co-evolved over thousands of years and contribute to the ecosystem functioning in a balanced way.

In European countries, forest management underlies legal regulations, e.g., with respect to reforestation after logging or keeping forests at healthy state [22,23,24]. Unfortunately, in Austria, the regulations for afforestation by use of seed material and saplings from local to regional tree nurseries [18] are nowadays not as strict as they were in the second half of the 20th century [25]. Reforestation was only allowed by use of locally produced saplings [18,26]. This regulation guaranteed that forestry should use generally trees that were adapted to the local climate. In practice, these rules were not always based on the most up-to-date knowledge of forest ecology. In Middle Europe specifically, Norway spruce was planted also far beyond its natural range. In consequence, this species suffers a lot from the recent climatic stressors, especially in case of non-natural stands.

Forest ecology has promoted the establishment of mixed forests to overcome the problems of monospecific stands for decades [27]. Mixed forests favoring native trees have been pointed out as most appropriate to fulfill the multiple functions of productive ecosystems under future climatic conditions [28,29,30,31]. Due to the accelerated climate change dynamics, traditional forest management will not have time to let the natural regeneration work until the fittest individuals made it [32,33,34]. In the tropics, reforestation of a mixed forest with native species is already applied to overcome environmental problems and to guarantee forest multifunctionality [35]. Few attempts to favor native trees were also made in parts of Europe [36,37]. Spatial limitations (natural or cultural land-use types) often prohibit the spread of suitable ecotypes of trees that were locally selected on regional substrate and dry habitat conditions. Forest managers might have problems finding such preadapted populations of native trees at the local or regional scale [38].

Therefore, we aim in this study to provide a tool for detecting drought-resistant ecotypes of the desired native trees at the local or regional scale. Specifically, we (1) compile a data pool that unites knowledge from forest site ecology and vegetation ecology (vegetation surveys), we (2) perform data mining by extracting site ecological information from the vegetation cover, we (3) select the percentile of the driest stands covered by the more common tree species and we (4) provide maps and metadata of stands of Austrian tree species capable of surviving under relatively dry conditions.

2. Materials and Methods

The basic idea in this paper is to select tree species provenances that are preadapted to future (dryer) site conditions by vegetation metadata analysis. Information on site conditions of forest stands is manifold but far-scattered. However, there is a high number of relevés published in vegetation ecology that can be used as indirect estimates of the respective site conditions.

Austrian forests were classified by the last author [39] based on a database comprising more than 20,000 relevés, with 19,058 of these relevés selected after a floristic quality-check for those that contain significant coverage of the more common Austrian tree species in the tree layer. The selected relevés represent plot-wise species lists of tree and herb layers, including data on coverage of species quantified by the Braun-Blanquet scale [40]. The geographical localization, aspect and altitude of all plots were stored in the metadata of this database.

Out of the 164 woody plant species in Austria, 73 taxa show tree-like shape. Some of them are rare and were not represented in the database. In order to get statistically sound results, we considered only those relevés where the given tree species covered at least 25% of the ground in one of the tree layers. This process resulted in 46 tree species to be represented in at least one relevé (

Table 1. Number of relevés for each tree species present at coverages of 25% minimum in at least one vegetation relevé.

Species	Number of Relevés	Species	Number of Relevés
Picea abies	4700	Ulmus glabra	95
Fagus sylvatica	2707	Salix fragilis	86
Pinus sylvatica	1145	Pinus cembra	83
Fraxinus excelsior	1133	Populus × canadensis	62
Alnus incana	890	Prunus padus	42
Quercus petraea s.lat.	809	Betula pendula	37
Alnus glutinosa	799	Populus nigra	35
Carpinus betulus	681	Acer platanoides	32
Abies alba	636	Prunus avium	29
Acer pseudoplatanus	475	Betula pubescens	25
Quercus robur	395	Robinia pseudacacia	24
Larix decidua	323	Ulmus minor	17
Pinus nigra	227	Fraxinus ornus	15
Salix alba	210	Populus × canescens	14
Tilia cordata	206	Ulmus laevis	10
Ostrya carpinifolia	183	Quercus dalechampii	5
Qercus pubescens s. lat.	175	Castanea sativa	4
Populus alba	170	Populus tremula	4
Quercus cerris	156	Salix caprea	2
Tilia platyphyllos	123	Acer tataricum	1
Acer campestre	113	Sorbus aucuparia	1
Fraxinus angustifolia	109	Sorbus torminalis	1
Pinus uncinata	105	Tilia x vulgaris	1

). Finally, 16 species were selected for detailed analysis due to their relevancy for silviculture under recent policy constraints.

Several trees in Table 1 are widespread in Austria and ecotypes are mentioned for some of them [38]. Some tree species can grow well under a very broad amplitude of climatic or edaphic factors and are cultivated in various habitats [27], e.g., Pinus sylvestris, Fagus sylvatica and Quercus robur. In this paper, we focus only on the most relevant tree species for Austrian forestry. In the full report of our StartClim20011D-project, even more tree species are discussed [38].

Ellenberg’s ecological indicator values (EIV) have been well-established in ecology for decades [41] and serve as an indirect estimate of the site conditions under which a plant species preferentially grows [42,43,44]. EIVs characterize the mean conditions at the species’ optimal performance and are defined for distinct geographical regions. The EIVs for Austria are available from BOKU [45]. They are based on the ratings by Ellenberg [46] and adapted for Austria [47,48,49].

The indicator for water supply (‘moisture’ = M) considers the availability of water for the plants on site irrespective of the source of water either from rainfall, air moisture or soil water. The ordinal scale for M spans from level 1 to level 12 [46]. Almost all of the plant species from Austrian forests are assigned an EIV for moisture [45]. Species that have no clear optimum along the gradient are classified as “indifferent” and do not contribute to the mean.

By calculating the arithmetic means of the EIVs of the species growing at a site a rough estimate of the site conditions can be given [42,44]. Calculating means of EIVs for relevés is sensitive to the number of species present in the plot and contributing to the means [50,51]. Such relevés often represent sites where light conditions at the forest floor are very poor, with only very shade tolerant species occurring. Therefore, means were calculated only for relevés that contained more than five vascular plant species. Furthermore, the species of interest was excluded from the calculation of the mean moisture values (mM).

The mM-values of relevés of a respective species vary around a mean for all relevés. We decided to affiliate the 10th percentile of the relevés with the relatively lowest mM-values as useful to pick drought-resistant populations of the respective tree species. We illustrate the distribution of all relevés comprising the respective species along environmental gradients such as altitude and aspect in comparison with the 10th percentile of the driest relevés of those species. Practitioners in forestry will find helpful the spatial positioning of the 10% relevés, where the most drought-resistant populations of tree species can be found for future forestry under climate change conditions (Appendix A).

The relevés were linked to climatic factors (mean annual precipitation and mean annual temperature for the period 1971–2000). We aligned the floristic mapping grids [52] of the corresponding relevés (if coordinates for the relevé were available) with grid-based means of annual precipitation and annual mean temperature data using ARCGIS 10.2 (Figure 1).

3. Results

3.1. Means and Amplitudes of Drought Tolerance of Selected Austrian Tree Species

In this subsection, we demonstrate the amplitude and means of ecological factors of plots where the respective tree species plays a significant role in the tree layer. mM-means of relevés indicate the average water availability for the tree species on the respective sites and can be used to define the 10% driest relevés of the respective tree species (e.g., Figure 2). In addition, the distribution of the 10% driest and the remaining 90% relevés along the gradient of environmental factors such as altitude and aspect confirm the selection of just the populations with the most drought-resistant populations of the respective tree species (e.g., Figure 3 and Figure 4). Furthermore, we projected the distribution of the 10% driest and the remaining 90% relevés of the analyzed tree species against the factors annual sum of precipitation and annual mean temperature at the respective sites of the relevés comprising our selected tree species (e.g., Figure 5). In this way, we can expect to provide sites with tree provenances that would fit to the expected future temperature and precipitation situation.

The first group of tree species consists of six Conifers with high silvicultural relevancy for Austria. The second group comprises of ten Angiosperm species that are also most relevant for actual as well as future silviculture from the ecological point of view. The species are ranked within the group according to the number of relevés in which they are reported (Table 1).

We calculated a mean mM-value of 5.41 for Norway spruce (Picea abies) and, similar to Silver fir (Abies alba), its optimum appeared to be represented by sites with an mM between 5 and 6. However, the wider distribution of mM-values over the moisture scale (3.6–8.8) suggests a larger amplitude in drought tolerance than that of Fir (Figure 2).

While the altitudes of the Spruce relevés range from 120 to 2000 m a.s.l., a substantial proportion of relevés were located between 1300 to 1500 m. The driest relevés, however, showed a more restricted range and more uniform distribution along the altitudinal gradient (Figure 3). While the general pattern was similar to Fir, the dominance of SE-, S- and SW-exposed sites in the dry relevés was more pronounced for Spruce (Figure 4). Furthermore, relevés belonging to the driest 10% demonstrated significantly higher mean annual mean temperatures (6.3 °C compared to 5.6 °C) and lower mean annual precipitation (1289 mm compared to 1379 mm) than the residual relevés (Figure 5).

Scots pine (Pinus sylvestris) is in general a light demanding species, although in terms of water requirements this species can be found at both ends of the moisture spectrum. This is represented to some extent by the wide distribution observed here: mM between 2.7 and 8.5. The mM-value of 10th percentile was, at 3.37, very low (Figure 2).

Scots pine also demonstrated a wide range in elevation, from 220 to 1890 m a.s.l., with elevations of around 700 m a.s.l. the most frequent. The range of the dry relevés was more restricted to lower elevations between 200 and 900 m a.s.l. (Figure 3).

Like Austrian pine (Pinus nigra), the distributions between aspect of both the dry and residual relevés were similar although compared to the residual relevés, the dry relevés with Scots pine showed a stronger preference for SE- rather than SW-exposures (Figure 4). However, unlike Austrian pine, Scots pine relevés belonging to the driest 10% demonstrated significantly lower annual precipitation (698 mm compared to 999 mm) than the residual relevés. The difference in annual mean temperature was statistically insignificant (Figure 5).

Silver fir was found occupying sites with an mM between 4 and just above 7, although the distribution indicates optimal sites with mM-values between 5 and 6 and a general avoidance of dry sites (Figure 2). The general avoidance of dry conditions is also indicated by the prevalence of W-, N- and NE-facing locations within the residual relevés (i.e., those not belonging to the 10th driest percentile) (Figure 4), where, due to less intense incident radiation, evaporative demand is typically lower. The driest 10% recorded mM-values of 4.93 and below and were strongly represented by SE- to SW-aspects. However, there was no significant difference in either annual precipitation or annual mean temperature between the driest 10% (the dry relevés) and the remaining 90% (the residual relevés) (Figure 5).

Silver fir is generally classified as a montane/submontane tree species, which is rather well illustrated by the elevation distribution of the relevés, with Silver fir of both the dry and the residual relevés (Figure 3). The fir relevés ranged from 300 to 1700 m a.s.l., with the highest dry relevés found at 1400 m).

The relevés comprising European larch (Larix decidua) showed a slightly skewed distribution along the moisture gradient: a few relevés indicating considerable presence on dry soils, moist sites (mM > 6.25) almost completely avoided and an optimum mM of 5.4 (Figure 2).

Apart from a difference in range, there was no clear distinction with respect to distribution over elevation between the dry and residual relevés. As Larch is an alpine species, often found in forest communities at the timber line, the range of the residual relevés reached just above 2000 m a.s.l. (Figure 3). The dominance of southern exposures was very clear in the dry relevés of larch (almost half were either S-/SW-exposed sites), despite a clear preference of the residual relevés for north-facing sites (Figure 4). Due to insufficient allocation of climate data to the Larch relevés, statistical comparison of temperature and precipitation values of the dry and residual relevés could not be performed.

Austrian pine is a very drought-tolerant species, which is well reflected by the distribution in mM-values: range in mM from 2.6 to 4.8 and dominance of relevés with mM of between 3 and 4. The mM of the 10th percentile was 3.15 (Figure 2). Consequently, the Austrian pine relevés were generally located at lower elevations. The distribution ranged from 250 to 1330 m a.s.l., with the majority lying within submontane elevations. This elevation zone was also dominated by the dry relevés, although their distribution beyond this range, particularly at higher elevations (>900 m a.s.l.), was more limited (Figure 3).

As a drought-resistant species, the distribution between aspect of both the dry and residual relevés demonstrated a clear preference for the S- to W-aspects, where, all other site characters being equal, evaporative demand is generally higher (Figure 4). There was no significant difference in either annual precipitation or annual mean temperature between the dry relevés and the residual relevés (Figure 5), perhaps due to the similar geographical distribution of both the dry and residual relevés (Figure A1 in Appendix A).

The rather narrow range in the mM-values indicates a clear preference of Stone pine (Pinus cembra) for fresh soils: the optimum was an mM of just above 5 (Figure 2).

Together with Larch, Stone pine is a common species found at the timber line of the Alps, although, unlike Larch, its range in elevation is more limited to higher altitudes: the residual relevés ranged from 1500 to 2270 m a.s.l., with the optimum lying between 1800 and 2000 m. The range of the dry relevés was narrower: 1700–2100 m a.s.l. (Figure 3). Of the dry relevés, SW-exposed sites were the most represented, although the general distribution over the different aspects was rather uniform (Figure 4). This was perhaps a result of the small number of dry relevés detailing the sites aspect (n = 83). Due to insufficient allocation of climate data to Larch and Stone pine relevés, statistical comparison of temperature and precipitation values of the dry and residual relevés could not be performed.

There are several broadleaved tree species in Austria that are the most valuable for sustainable forestry. We illustrate the distribution of the 10th percentile of driest relevés for deciduous trees in Figure 6 and project the relevés against ecological factors in Figure 7, Figure 8, Figure 9 and Figure 10).

Most prominent for ecologically sound forest management is European beech (Fagus sylvatica). With respect to distribution over the moisture gradient (mM-values), Beech shows a similar pattern to Hornbeam (Carpinus betulus) (Figure 6). However, the slightly higher mean mM-value of all relevés (5.2) and that of the 10th percentile (4.73) indicates that Beech is—on average—less tolerant against dry conditions than Hornbeam. The residual relevés of Beech spanned from the lowest elevations up to 1640 m (Figure 7). In the Northern Calcareous Alps, Beech can often be found in the forest communities at the timber line. However, the most drought-tolerant provenances seem to be distributed further down slope, between 300 and 1000 m a.s.l.

North-facing sites were most frequently represented by the residual relevés, while dry sites were more evenly distributed (Figure 8). Furthermore, both mean annual mean temperature (7.8/6.3 °C) and annual precipitation (962/1307 mm) were significantly differ-ent between the dry and the residual sites (Figure 9 and Figure 10).

The distribution of relevés with Ash (Fraxinus excelsior, Figure 6) illustrates a preference for rather fresh site conditions: a fairly wide normal distribution (mM between 3.8 and 7.5) around a mean of 5.57, with a 10th percentile of 4.9. The elevations recorded for the Ash relevés ranged right up to 1340 m a.s.l., although the dry provenances appear to occupy lower elevations (generally below 500 m a.s.l., Figure 7).

While all aspects were well represented by Ash relevés, flat/slightly inclined sites were the most frequent. Flat sites were also frequent among the dry sites, although of the aspects, southern exposed sites were more common (Figure 8). As with Beech, both mean annual mean temperature (9.1/7.7 °C) and annual precipitation (657/1121 mm) were significantly different between the dry and residual sites (Figure 9 and Figure 10).

The slightly skewed distribution of Sessile oak (Quercus petraea s.lat.) relevés means that while most frequently represented mM-values (4.5–4.75) lay close to the moist end of its range (5.7), the dry limit of this species distribution stretched right into the very dry parts of the moisture gradient (mM of 2.5) (Figure 6). Most of the residual Sessile oak relevés and all of the dry sites were located below 500 m, with the majority at around 350 m a.s.l. However, residual relevés from sites as high as 1200 m a.s.l. were also present (Figure 8). Apart from the dominance of the flat sites among the residual relevés, SE- to SW-aspects were most common among both the dry and residual sites. Finally, both mean annual mean temperature (8.2/8.7 °C) and annual precipitation (557/643 mm) were significantly different between the dry and the residual sites of Sessile oak (Figure 9 and Figure 10).

The mean mM-values derived for Hornbeam showed a normal distribution around a mean of 5, ranging from 3.9 to 6.5 (Figure 6). The 10th percentile value of 4.54 demonstrates some degree of drought tolerance within Austrian Hornbeam populations. The residual relevés of Hornbeam ranged from the lowest elevations in Austria up to 1300 m a.s.l., with elevations between 200 m and 400 m the most frequent. The relative distribution of the dry relevés also mirrored this pattern (Figure 7). Generally, Hornbeam relevés on flat or only slightly inclined sites were the most common, although all aspect classes were well represented within the residual relevés. Flat or slightly inclined sites were also common among the dry relevés, though of the inclined sites, southern aspects (SE, S, SW) occurred more frequently (Figure 8).

Of the compared climatic variables, the dry relevés were shown to receive significantly less annual precipitation (686) over the last 30 years than the residual sites (778 mm) (Figure 9 and Figure 10).

In comparison to Field maple (Acer campestre), the distribution of Sycamore (Acer pseudoplatanus) is rather narrower and finds itself more towards the moist part of the gradient. The mM-values ranged from 4.7 to 6.5, with values around 5.5 the most frequent (Figure 6). This species clearly prefers more fresh site conditions, as illustrated by the rather high 10th percentile value (5.26).

Both the dry and residual relevés of Sycamore also demonstrate a wide range in elevation (Figure 7). The elevation of the residual relevés ranged from 100 to 1700 m a.s.l., while those of the dry relevés were restricted to lower elevations (200 to 1000 m a.s.l.). Generally, the N- and NE-aspect classes were strongly represented by the Sycamore residual relevés, although the distribution for the dry relevés was comparatively more uniform (Figure 8). Relevés belonging to the driest 10% demonstrated significantly lower mean annual precipitation than the residual relevés (1339 mm compared to 1423 mm). Meanwhile the difference in mean annual mean temperature was statistically insignificant (Figure 9 and Figure 10).

Of the Oaks, Pendunculate oak (Quercus robur) is the most competitive on moist sites, though the wide distribution over the moisture gradient illustrates the wide range in moisture tolerance within its Austrian population. This species demonstrated a normal distribution (between 3.2 and 8.0) around a mean of mM = 5.1, with a 10th percentile of 4.48 (Figure 6). In terms of elevation, Pendunculate oak relevés were found as high as 1100 m a.s.l., though the majority of the distribution of both the dry and residual relevés was restricted to the foothill zones (below 500 m) (Figure 7). Many of the sites were found on rather flat terrain, though the frequency of southern exposed slopes was clearly larger for the dry relevés (Figure 8). Again, as with Sessile oak, only the difference in mean annual precipitation between the dry and residual sites (562/580 mm) was significant (Figure 9 and Figure 10).

The distribution of the Small-leaved lime (Tilia cordata) showed a clear general preference for fresh sites (i.e., soils with a balanced water budget). A rather normal distribution around a mean mM-value of about 5 (5.08) was observed, spanning a range from 3.4 to 6.3. The 10th percentile was 4.59 (Figure 6). While the relevés ranged from the foothills to montane elevations (maximum of 910 m a.s.l.), the dry sites were restricted to the foothills and submontane zones (Figure 7). The residual relevés indicated a tolerance of most aspects and inclinations, though N-exposed sites appear to be preferred. The more drought-adapted provenances seem to be found more often on S- to SW-exposed sites (Figure 8). There was no significant difference in either mean annual precipitation or mean annual mean temperature between the dry relevés and the residual relevés (Figure 9 and Figure 10).

Downy oak (Quercus pubescens s.lat.) exhibited a very similar distribution to that of Turkey oak (Quercus cerris), except for being shifted slightly further towards the drier end of the moisture gradient. The distribution ranged from 3.1 to 4.8, with a mean mM-value and 10th percentile of 3.9 and 3.43, respectively (Figure 6). Both the dry and residual relevés demonstrated very similar ranges in elevation, though the distribution within this range was opposite to one another: relative dominance of lower elevations among the residual sites, and vice-versa for the dry sites (Figure 7). In both cases, the south-exposed sites were most frequent for Downy oak (Figure 8). Although no significant difference in mean annual mean temperature was observed, the difference in mean annual precipitation between the dry and residual sites (562/580 mm) was significant (Figure 9 and Figure 10).

The mM-values of Turkey oak were distributed between 3.2 and 5.6 (Figure 6). The mean mM-value was calculated at 4.4, indicating an optimum on dry to slightly desiccating soils. While elevations ranged from 180 to 570 m a.s.l., the majority of both the dry and residual relevés were found to be 200 and 300 m a.s.l. (Figure 7).

For both the dry and residual relevés, a similar distribution among the aspects was observed; a dominance of flat/slightly inclined sites, with southern exposures the next most frequent (Figure 8). A significant difference in mean annual precipitation was also observed between the dry (561 mm) and residual (628 mm) relevés (Figure 9 and Figure 10).

The mM-values of relevés with Field maple ranged from 3.4 to 6.6, although as the histogram depicts (Figure 6), the lower mM-values are poorly represented. The optimum appears to be around 5, while the mean mM-value of the 10th percentile was 4.5. The distributions between elevation classes revealed that Field maple generally occurs at lower elevations, with elevations of around 300 m most common (Figure 7). Furthermore, although data on aspect was limited for this species, the distribution of the residual relevés between aspect classes indicates that this species prefers ± flat sites with an inclination less than 5° (Figure 8). There was no significant difference in either annual precipitation or annual mean temperature between the driest 10% and the residual relevés (Figure 9 and Figure 10).

3.2. Geographical Distribution of the Relevés and Locations of Potentially Drought-Tolerant Provenances

Details of the geographical distribution of the relevés with interesting tree provenances are given in Appendix A. Figure A1 demonstrates the distribution of relevés comprising of four important Austrian Conifer trees, whereas Figure A2 and Figure A3 show the same for ten Angiosperm trees.

The selected tree species are economically and ecologically important for actual forest management in Austria. From these maps one can deduce in which parts of Austria source populations can be sampled for the proliferation of trees adapted to climate change in tree nurseries.

In case of Norway spruce the map of evaluable relevés (Figure A1, green symbols) does not reflect exactly the natural distribution within Austria. This map fails to illustrate the very hotspot of natural Spruce occurrences in the Central Alps. Unfortunately, 75% of all relevés with Spruce missed detailed geographical coordinates and cannot be shown in our map. Nevertheless, one can forage for the red dotted source populations of Spruce in the northern and eastern parts of the Alps and the Hercynian region. The lowland spruce forests cannot provide pre-adapted spruce, due to the fact that all these stands are planted.

The map for Scots pine (Figure A1) illustrates that its occurrences in Austria are scattered and bound to dolomitic or acidic substrates. Like with Spruce, Scots pine is mostly planted in the lowlands and suffers from the drought periods. Drought-adapted provenances can be expected to the west in the Tyrolean Inn valley, in some dolomitic outcrops of the northeastern and southeastern Alps, and in the easternmost part of the Hercynian region at the border to the Pannonian lowlands.

In case of Silver fir (Figure A1), there can be seen also a gap of relevés in the Central Alps that can be called ± natural. High rainfall areas in the Alps and the Hercynian region serve several relevés with fir, and the corresponding 10% of the driest sites are to be found throughout the whole distribution area. This allows a representative selection of drought-adapted ecotypes for Silver fir nurseries.

The Austrian pine is documented naturally only from few dolomitic sites towards the eastern edge of the Alps (Figure A1). There are more plantations on extreme soils such as in the plains south of Vienna and in some sand dune systems east of Vienna. Austrian pine is well adapted to soil generated drought and was, therefore, planted far from its natural area, such as the Vinschgau in South Tyrol or the dolomitic and sand hills in Slovakia and Hungary. Foresters experienced pest problems in plantations outside the natural area [54]. Therefore, planting of this species sounds very interesting due to its high drought resistance, but needs more empirical exploration and testing.

We found in the database also many relevés with Stone pine, which occur preferably in the subalpine zone of the central Alps and a few places in the Northern Calcareous Alps of Austria. There were not enough relevés from the 10th percentile of the lowest mM-value that were assigned coordinates, which would allow exact positioning in our map. In consequence, we cannot provide a separate map. The same problem of too few relevés for mapping holds true for European Larch, which is generally more widespread throughout the Alps. Detailed descriptions of both relevé pools are given in Appendix A.

From the deciduous native trees, some Oak species were specifically promoted and propagated by foresters in some regions. More widespread trees such as European beech are not so much kept in nurseries. Indeed, European beech is the most widespread deciduous tree in Austria (Figure 11), avoiding the central parts of the Alps and the very flat plains of the Pannonian lowlands because of rather continental climate [55]. This species grows on all kinds of basic to acidic soils and is currently expanding its local range towards the lowlands as well as upwards to the timberline. Beech is a tree dominating wherever it can grow undisturbed. Skeletal dolomitic soils throughout the Northern Calcareous Alps and sites at the dry edge of the Alps and the Hercynian region provoke the establishment of several populations showing awesome survival under dry conditions (Figure A2). The example of Beech illustrates the practical use of the results of this dataset and other such studies. We could select drought-tolerant populations in all forest ecoregions [26] where Beech represents a focus species in Austrian forestry. Specifically, there are several relevés with low mM-values filtered in the forest ecoregions 4.1 and 4.2 (Figure 12) that represent the center of beech forestry. The identification of potentially drought-adapted provenances for the different forest ecoregions will help identify and collect seeds from source populations for nurseries to help Austrian forestry adapting to the projected changes in climate.

Ash is very specific, as it is a formerly widespread but rarely dominating species, but it is suffering enormous from pest attacks at stand level [56] causing serious diebacks. The potential for the selection of drought-tolerant populations is obviously higher in the eastern part of Austria (Figure A2). In this species, the search for drought tolerance is probably less important than the search for pest resistant progeny [57].

From the oaks, Sessile oak is commonly used by lowland forestry in eastern Austria. Several relevés on dry sites are aggregated at the border between the Hercynian Region and the northern part of the Pannonian lowland (Figure A2). In case of Pedunculate oak (Figure A3), several ecotypes differing by their need of water are known and commercially used in forestry. Though there is a 10th percentile of relevés from dry habitats, it seems unclear if they would fit to other site parameters as well. The remaining two oak species are of low frequency at the moment, and their areas are restricted to the east (Turkey oak) and to the east and south of Austria (Downey Oak), respectively (see Figure A3). Both species are widespread towards the south of Europe, and one can expect that the 10th percentile of the driest sites may provide very drought- and heat-resistant provenances.

Hornbeam is known in forestry, i.e., for its high ecological input by stimulating nice humus conditions. It is widespread throughout Austria’s lowlands and peripheral parts of the Alps. Consequently, also the 10th percentile of driest relevés is to be found outside or towards the edge of the Alps (Figure A2).

Scyamore grows from the lowlands up to the subalpine zone, preferring nutrient-rich and relatively moist sites (Figure A2 and Figure 13). The driest 10% of relevés cover quite nicely the whole distribution area.

From the maple species, Field maple is the second most important in Austrian forestry. Its natural area covers the Pannonian lowland, where it is widespread (Figure A3). There are several populations that could serve as drought-resistant source populations for propagation of seedlings well adapted to climate warming.

Finally, the Small-leaved lime is valuable also for increasing the quality of humus even on dry sites. Source populations for drought-tolerant provenances are to be found mostly in the lowlands, but few are also available in the continental parts of the alpine Inn Valley in Tyrol (Figure A3).

4. Discussion

According to the Austrian Forest Act [22], provenances from the same forest ecoregion represent those best adapted to the respective site conditions, i.e., soil and climate conditions. The Forest Reproductive Material Act [25,58], subsequently, states that these provenances represent the best sources of reproductive material to maintain the optimal performance of Austria’s forests, i.e., the optimal provision of production, welfare and protective functions. However, with the ongoing climate change it seems questionable whether these provenances will be the best adapted under future site conditions.

Currently, there are relatively few data sources that forest managers can consult when selecting the most appropriate reproductive material for a given site. Regarding performance in terms of timber production (i.e., yield potential), extensive data on yield and increment from several thousand plots in Austria, covering the main production forest species, are made available by the national forest inventory [59]. However, these data only provide an indication of production potential, and, importantly, production potential under current site conditions; the dataset gives no indication of performance under future site conditions. Austrian forestry is therefore in need of additional data so as to select the provenances best suited to the likely changes in climate, i.e., increases in mean temperatures and drought incidence [60].

This study demonstrates the potential of using the Austrian Vegetation Database [61] in identifying the provenances likely to be best adapted to future climate conditions. Like the Austrian forest inventory, this database is very extensive in its spatial coverage, with over 36,000 vegetation relevés taken from all over Austria. By deriving ecological indicator values of site moisture, our study demonstrates the relative range in site moisture conditions where 16 Austrian tree species occur. Furthermore, the driest sites identified here (i.e., the driest 10%) potentially represent the provenances of each species that are most adapted to drought conditions (Figure 13). While inter-population variations in phenotype may result from acclimation mechanisms [62], site-specific selection pressures (natural or even artificial, e.g., negative selection thinning based on vitality), if allowed to proceed over several generations, are likely to have facilitated some adaptation to drought due to the high within-population genetic variation typical for trees [63]. While the derived indirect ecological indicator values give no indication of yield potential, the data may help managers promote stability by identifying drought-tolerant provenances. This represents a key objective within the emerging sustainable forest management paradigm, where the optimal provision of functions and products (e.g., promoting biodiversity and protective functions) is delivered by maximizing ecosystem function rather than maximizing yield [30,33,64] and promoting biodiversity in forests even under climate change prospects [65,66].

Modeling can help to understand the general future processes of stand dynamics in forests under climate change [67,68], but distribution models are commonly based on the actual or formerly used and widespread provenances of trees [69]. The potential of pre-adapted ecotypes of native tree species is not included in any model. According to these models, European beech might disappear from the Middle European lowlands and withdraw to higher mountain regions [69]. This scenario could possibly be avoided if drought-tolerant provenances were used for reforestation.

As expected, many of the identified potentially drought-adapted populations identified (i.e., the driest 10%) were, as expected, located in the warm and dry regions of Austria (Figure 1). Many were found, for instance, in the Pannonian region of the Vienna Basin. As this area is subject to rather intensive agriculture, most of the dry sites were located on the margins of this area, e.g., along the Manhartsberglinie, the area of Wachau or the Thermenlinie, where due to the higher reliefs, these areas are less suitable for agriculture purposes than say the foot of the Leiser and Hainburg Mountains or the Leithagebirge. The dry sites of Turkey and Downy oak, Field maple and Austrian pine were found limited to these particular areas. In addition to the Pannonian areas, other dry areas such as the inner Alps, the Styrian Fringe Alps and the South-eastern foothills were also identified as potential locations for drought-adapted provenances of Pendunculate oak and Hornbeam. Furthermore, the dry sites of Scots pine were also restricted to the drier eastern parts of the country.

The two-sample t-tests indicated that reduced precipitation (thirty-year mean annual precipitation) was a decisive factor distinguishing between the dry and residual sites of Norway spruce, Scots pine, European beech, Hornbeam, Sycamore, Ash, and Large-leaved lime, as well as Turkey-, Sessile- and Downy oak (Figure 5 and Figure 9). Furthermore, due to increasing evaporative demand with increasing temperature, it was also shown that the dry sites of European beech, Norway spruce, Ash, Sessile oak and Large-leaved lime were on average significantly warmer than the corresponding residual sites (Figure 5 and Figure 10).

Despite the often-significant difference in the mean climatic variables between the dry and residual sites, it was also evident that topographical features such as aspect are equally relevant in terms of site dryness. For example, southern-exposed (SW, SE, S) slopes where due to higher incident radiation, evaporative demand is typically higher, were particularly frequent amongst the dry sites of Silver fir, Norway spruce, Larch, Hornbeam, European beech, Ash and Small-leaved lime as well as Turkey-, Pendunculate and Sessile oak, (Figure 4 and Figure 8). The data also revealed a role played by elevation; the dry sites of Stone-, Austrian-, Scots pine, Field maple, Sycamore, European beech, Ash and Small-leaved lime as well as Sessile- and Pendunculate oak were, typically, more frequent at lower and, therefore, warmer elevations with higher evapotranspiration than the corresponding residual sites (Figure 3 and Figure 7). However, this may indirectly correspond to a precipitation effect due to increased orographic precipitation with increasing elevation. Finally, the data may also indicate a relationship between hydrological soil properties and the ecological indicator values derived. For example, many of the dry sites of European beech, Norway spruce, Silver fir and Sycamore were found within the Northern and Southern Calcareous Alps, where precipitation is relatively high. This observation may be due to the low water holding capacity of the thin rendzina soils typical of these regions [38]. Nevertheless, the dry sites of the Small-leaved lime and Ash were, despite their wide distribution, limited almost exclusively to areas of low rainfall, despite the wide distribution of these species.

These variations in dryness, due to factors such as topography and hydrologically-relevant soil properties, may be particularly significant when identifying drought-adapted provenances within areas with a rather homogenous climate, such as forest ecoregions [26]. Due to the wide distribution of Beech and, thus, the number of available relevés, it was possible to identify potentially drought-adapted populations for different forest ecoregions (Figure 12). Given the uncertainty posed by the introduction of new untested provenances [33], the identification of potentially drought-adapted provenances in the different ecoregions would at very least provide greater flexibility to forest managers who want to foster adaptation to climate change. However, these provenances must nonetheless be screened themselves via provenance trials/pot experiments to confirm whether or not they are superior with respect to drought resistance [66].

While our results illustrate the potential of the Austrian Vegetation Database to facilitate the adaptation of Austrian forestry to climate change, there are significant weaknesses in the database, particularly concerning the geographic location of the relevés. In some cases, precise geographic coordinates have been assigned to the relevés. However, in many cases, the relevés have only been assigned to quadrants of the floristic mapping according to [52], i.e., grid cells of 5 × 3 geographic minutes. A substantial number of relevés currently bear no geographic information at all, except for a verbal description of the sample region. In many cases, the spatial accuracy of the data could be improved by consulting the original sources, which is, however, a time-consuming task. Another disadvantage of the database is the age of some relevés, so that one has to check if the described stands even still exist.

If an identified autochthonous provenance still proves to be inadequate in terms of adapting to climate change, forest management may nonetheless look at provenances of the same species from abroad or new exotic species [33,70]. However, as the data of this study indicates, forest managers may not need to look abroad for alternative species more resistant to drought stress. As Figure 13 shows, rather drought-tolerant provenances for a number species already exist here in Austria, e.g., Austrian- and Scots pine and Downy-, Sessile- and Turkey oak. While their suitability to other site-specific factors must of course be considered (e.g., soil pH, light and nutrient requirements etc.), the data nevertheless illustrate that there is enough potential within Austria’s autochthonous tree species for them to be at least be considered alongside alternative adaptation strategies such as exotic provenances or species. First research projects are conducted, e.g., for Sessile oak in the national park Thayatal [71]. This area belongs to the Manhartsberglinie where several of the 10% driest relevés of this species are situated. Furthermore, stands of Pendunculate oak, which are identified as vulnerable to climate change, could be steadily adapted to climate change by steadily introducing more drought-tolerant Oak species such as Sessile oak or Downey oak. The well-known hybridization of those Oak species [72] may support the natural introgression of drought-tolerance traits into former less resistant Pedunculate oak populations [73].

5. Conclusions

While the pressure of climate change continues to push the case for foreign provenances or exotic species [19,70], their introduction, particularly that of exotic species, may have negative ecological effects [33,74,75]. Forestry should, therefore, not ignore the autochthonous options available to adapt to the projected changes in climate. Our study illustrates that a potentially significant amount of inter- and intraspecific variation in drought tolerance is available within Austrian tree populations. The next steps are to assess this potential and to design effective ways of incorporating it into the strategies aimed at helping the Austrian forestry to adapt to climate change. Seed material from drought-resistant populations designated by our selection strategy can be cultivated in tree nurseries to be planted on sites that are expected to be more prone to water deficiency under future climate scenarios. However, the potential of more sustainable water management in some tree provenances cannot guarantee the overcome of other hazards, such as unexpected pest waves [9,10,54].