Reforestation challenges in Southeast Europe facing climate change

Projections of the regional climate model for Southeast Europe generally predict an increasing of temperature and a decrease in precipitation, with some local variations. Higher frequency of extreme weather events and increased flooding can also be expected. This climate change will, among other things, result in changes in habitats and species distribution, and a decrease in biodiversity. In most cases, forest ecosystems will be unable to adapt fast enough to keep pace with changes in climate. Extreme weather events and low precipitation during the growing season will cause high mortality of seedlings after planting. New forests will face the whole range of these changes because of the long lifetime of trees. Reforestation programs must take projections of climate change into consideration. In the long term, new guidelines for site-species matching, provenance selection, and genetic diversity need to be adopted. In the short term, site preparation, planting techniques, and post planting protection need to be improved. In addition, seedling quality (morphological, physiological, and genetic) and planting time need to be specific for each site. New site preparation, planting, and post-planting protection methods are useful tools for short term success measured in seedling survival and initial growth. Seedling quality is essential for short and long term success. Different strategies, such as assisted migration and increased genetic diversity of planting material, can provide better chances for long term success measured in growth, fitness, and capability to produce the next, better adapted generation.


1
In trod u ct io n 17 9 1. 1 D e fin it ion s -Th e fr a m e 18 0 1. 2 R e fo re st ati on i n S ou th ea s t E u r op e -Th e sta g e 18 1 2 Cl i mat e ch an g e -Th e ch all en g e 18 2 2. 1 Ran ge 18 2 2. 2 Sp a ce an d t i m e -Th e r a c e 18 3 3 Lon g -t e r m str at eg ie s -T h eo re ti cal ac tio n s

Introduction
Reforestation challenges are many. Except in industrial plantations, we are witnessing a global trend of reduced investments in reforestation. Overexploitation, land-use politics, forest fragmentation, and urbanization are threats to existing forests. In addition to the human factor, we are witnessing more frequent extreme weather events, the conquest of habitats by invasive tree species, migration and areal spread of pests and diseases, and the maladaptation of tree populations transferred to new sites.
Afforestation, reforestation, and deforestation avoidance are three types of climate change mitigation projects in the forestry sector (Reyer et al. 2009). Facing climate change, future management plans need to integrate adaptation and mitigation approaches (Millar et al. 2007). Climate change mitigation is not a part of this review and we will focus on management approaches available in changed climate. In changed climate, a number of adaptation strategies are available, and need to be involved in current and future afforestation/reforestation projects. Forest tree populations can extirpate in a rapidly changing environment or can persist thorough migration or adaptation (Aitken et al. 2008). "Migration and adaptation are considered as alternative responses to environment changes because evolution allows populations to adapt to novel conditions without migrating, whereas migration lets populations track favorable conditions without evolving" (Kremer et al. 2012). Reforestation and restoration programs facing climate change need to consider and to promote both strategies.
Looking into the future, climate change and the uncertainty of the environment following this change will be the most important challenges to reforestation success. It is the regeneration phase that will initially be susceptible to the changed climate (Spittlehouse and Stewart 2003) and additional efforts need to be invested in artificial regeneration of forests. To improve reforestation success, selection of species and genetic material as well as nursery cultural practices must focus on overcoming planting stress on harsh restoration sites. This can be done by enhancing the ability of seedlings to withstand drought, frost, vegetative competition, nutrient deficits, and animal damage (Jacobs et al. 2015). Additionally, planting techniques (pre planting, planting, and post planting operations) need to be improved in order to provide better chances for seedlings to survive the critical establishment phase.
The Southeast Europe (SEE) region ranges from sea level at coastal areas and plains, to river valleys, karstic area, and high mountains, resulting in rich biodiversity and natural resources. Many SEE regions belong to the areas of Europe most vulnerable to climate change, where we can expect the highest future impacts of climate change to ecosystems, with associated impacts on biodiversity (Laušević et al. 2008).
In this paper, we review the scale of reforestation in SEE, discuss theoretical predictions of climate change and its potential influence on reforestation success, and finally review the literature on seedling production and planting strategies and actions as a response to challenges presented by climate change. Seedling survival after field planting is a keystone to any reforestation project. This short-term success can be diminished in subsequent decades due to changing environment, especially when mistakes occur in the planning phase. This is why, following our review on challenges, we first pay attention to long-term actions that promote short-term success during the seedling establishing phase. http://www.efi.int/portal/members/membership_service/benefits/forest_map_of_europe Except Serbia and Greece, the area of planted forests in all SEE countries has decreased during the last 25 years (Tab. 2). These data should, however, be used cautiously. For example, in Serbia, comparing the FAO assessment with the last national forest inventory (Banković et al. 2009) showed a similar total area of planted forests; but positive trend with an inconceivable increase in the area of planted forests during the last 10 years. Similar caution should be used when examining trends on planted forests in Bosnia and Herzegovina and FYR Macedonia. 2 Climate change -The challenge

Range
The climate of SEE ranges from subtropical to continental, with a temperature range from 2 to 25°C and a precipitation range from 19 to 7,000 mm (Tab. 3). Precipitation levels follow a clear annual pattern in Albania, with the maximum in winter and the minimum in summer (Diku 2011). Other SEE countries followed that pattern, except Bulgaria and Serbia, where the maximum precipitation level is in spring and the minimum occurs in summer (Önol and Semazzi 2009).
Projections of the regional climate model for SEE predict an increasing of temperature and a decreasing of precipitation, with some local variations. According to the Α2 scenario, the mean maximum temperature across SEE is projected to increase between 2.4 and 7°C. Except for a projected increase in some regions of Serbia of plus 5%, a general decrease of precipitation is predicted for other SEE countries; Croatia may see 54% less precipitation (Tab. 3). According to projected simulations, the maximum decrease of precipitation (in percent) in all SEE countries will be in summer (Önol and Semazzi 2009), which will significantly increase the fire risk in SEE and Mediterranean areas (Moriondo et al. 2006).  (Önol and Semazzi 2009) Abbreviations: S -subtropical, C -continental, M -Mediterranean, A -alpine, T -temperature (°C), P -precipitation (mm), Tp -predicted increase of temperature (°C), P -predicted change of precipitation (%).
In addition to the significant increase of air temperature and decrease in precipitation, the biggest challenge to reforestation success will be the frequency, duration, and severity of extreme weather events. Further decreases of precipitation during summer, which already sees the minimum precipitation levels, will further increase chances for drought. We may already be witnessing some of these extreme events. From years 1958 to 2000 in the Eastern Mediterranean, the most significant temperature trends were revealed for summer, where minimum and maximum temperature extremes show statistically significant warming trends, together with increasing trends for an index of heat wave duration (Kostopoulou and Jones 2005). Out of seven events of the most relevant heat waves that hit the Carpathian Region (including Serbia and Croatia) from years 1961 to 2010, four occurred from 2000 to 2010 (Spinoni et al. 2015). Summer 2007 was abnormally warm for many areas of SEE with deviations from the seasonal means exceeding 4°C in some areas but also distinct periods of extremely hot weather (Founda and Giannakopoulos 2009). During summer 2007 the maximum temperature reached 44.9°C, the record value ever recorded in Serbia (Hydrometeorological Service of Serbia 2015) since regular measurements started in 1848. In year 2011, only 65% of normal, average precipitation for the period 1971-2000 was recorded in Serbia (Ivetić 2015). The summer of 2012 was very hot and dry in SEE (Sippel and Otto 2014). At 19 main meteorological stations in Serbia (of 29) the summer of 2012 was the hottest since measurements began (Hydrometeorological Service of Serbia 2013). In spring 2014, heavy rains caused floods and landslides in Serbia and Bosnia and Herzegovina, resulting in loss of human lives and severe damages to infrastructure and habitats. Hydrometeorological hazard has changed in SEE including a distinct increase in the frequency of summer heat waves (Sippel and Otto 2014).

Space and time -The race
Following climate change, the current environmental conditions on which forest populations are adapted will move northward and upward (Ledig and Kitzmiller 1992;Loarie et al. 2009;Gray and Hamann 2013). These changes are expected to be faster than ever before (Kirschbaum and Fischlln 1996). The main concern is that migration of forest populations and genes cannot keep the pace with these changes.
The SEE region lies between latitudes 34° and 47° N, with a distance from south to north of approximately 1,100 km. Elevation ranges from 0 m by the coastline to 2,925 m at Musala peak in Bulgaria. Large mountain ranges running from north-west to southeast cover most of the region. In the southern part of the region, evergreen vegetation is predominant, whereas in the north, oak and beech forests are dominant at lower elevations with spruce, fir, and pine in the mountains. Tree line rises from 1,800 m in the north to 2,300 m in the south.
With projected warming of between 0.1 and 0.35°C per decade, tree species would have to migrate 1.5-5.5 km northward per year or to increase elevation by 1.5-5.5 m per year in order to remain within similar climatic conditions (Kirschbaum and Fischlln 1996). Many studies of past changes have estimated natural rates of migrations of trees ranging from 1.26 to 1,000 m per year (Tab. 4). Forest inventory Woodall et al. 2009 The maximum migration distance during a given time period depends on: 1) number of dispersal events in that time period and 2) distance covered by each event (Corlett and Westcott 2013). On one hand, the rate of present-day tree migrations is frequently slower than expected (Renwick and Rocca 2015). Bodin et al. (2013) found that only 2 out of 31 tree species studied were migrating fast enough to keep pace with temperature changes. On the other hand, estimates on long distance gene flow suggest that genes can move over spatial scales larger than habitat shifts predicted under climate change within one generation (Kremer et al. 2012).
Estimations on speed of climate changes and migration rate of tree species vary depending on input data and methodology. From the literature, we conclude that most tree species in SEE will not be able to keep pace with climate change under current scenarios. The region's geography is one obstacle to migration of tree species. Even if the Carpathian Mountains provide a natural corridor for migration to the north-east, the Pannonian Basin is an obstacle for migration of mountain tree species northward. Additional obstacles will include habitat fragmentation, unsuitable soil conditions (both depth and nutrients), and absence of beneficial microorganisms, such as mycorrhizae.

Goals and objectives: Site specific or species specific projects?
Setting the right goals and choosing an appropriate strategy to achieve them is essential for long-term reforestation success. In this sense, we need to distinguish site specific from species specific projects. Preserving a forest at a specific site will differ from translocation of specific species for conservation purposes. The main difference is that in site specific projects we need to match appropriate species and provenance with the focal site, whereas in species specific projects we need to find the best site for focal species. Although both types of projects rely on site-species matching, tools and available materials differ.

Site-specific projects
Site specific projects generally occur when the goal is afforestation of bare land or reforestation following harvesting or disturbances such as forest fire. Artificial or assisted natural regeneration also can be considered as site specific projects. In these types of projects, selection of the most appropriate species is essential for success and the only way this can be properly done is by investigating all environmental conditions on the site at the very beginning.
Despite a wide range of forest functions, the main goal of a site specific project can be prevailing productive, conservative or ameliorative. However, the uncertainty of future conditions will bring these goals closer together. The main goal determines the selection of appropriate planting material (species, provenance, level of genetic diversity).
In the case of a productive goal, the new genes (species, provenances) can be introduced with the aim to exploit the focal site production capacity to the highest amount. Many examples show higher productivity of non-local provenances (Schmidtling and Myszewski 2003;Ivetić et al. 2005;Krakowski and Stoehr 2009) and non-native species (Heryati et al. 2011;Kawaletz et al. 2013;Guo and Ren 2014;Kjaer et al. 2014) compared to local populations at a specific site. Non-native species could play an important role in cases where they provide short-term benefits to ecosystem function and promote the potential for longer-term succession to native species, but this practice is controversial and debated vigorously (Thomas et al. 2014;Jacobs et al. 2015).
In the case of a conservative goal, the goal, priority is to maintain the existence and sustainability of the forest ecosystem. The usual practice is to reintroduce the same species using local forest reproductive material because that material is best adapted to the local conditions. Maintaining the current state or restoring a previous forest composition will be difficult in a changing environment. This is why the primary objective of forest restoration should be focused on functional ecosystems at the landscape level (Stanturf et al. 2014).
In case of an ameliorative goal in SEE, degraded forest is usually removed by a clear-cut harvesting and traditionally planted with fast-growing pioneer species (usually Pinus nigra Arnold) with a goal of facilitating the introduction of late-successional hardwoods (Stilinović 1991). Plantations established on sites more productive than that actually needed to support Pinus nigra are characterized by rapid succession dynamics of mainly broadleaved species (Zlatanov et al. 2010).

Species selection
Most forest sites can support more than one tree species, which allows selection of the best species to meet project goals. Species selection in a reforestation project is a complex decision that should be based on site conditions, environmental issues, economic criteria, and climate change predictions. "Species choice in large-scale reforestation programs should be determined by the maintenance of the realized niche under most climate change scenarios, avoiding potential exposure of forest trees to pests and diseases under a continued warming trend" .
Species biological characteristics should be considered as well. For example, the water-acquisition strategy of tree root systems can determine the survival capacity under severe drought. Deep-rooted species are highly recommended for reforestation in dry conditions, even under low soil water availability (Ovalle et al. 2015). The hydric model of plant water use behavior is potentially useful in selecting appropriate species in arid to semi-arid conditions (Kjelgren 2010). Quercus ilex trees, which showed a more anisohydric behavior, replace Pinus sylvestris L., which showed a typical isohydric behavior in a montane Mediterranean forest in association with recent episodes of drought-induced mortality (Aguade et al. 2015). Table 5. Heat degrees (coordinate E) as combination of slope and aspect (Lujić 1960).

Aspect
Slope ( In Serbian reforestation programs, species selection is usually based on: 1) potential vegetation, 2) ecological differentiation, and 3) local heat potential. Species selection based on potential vegetation is limited to a small area with uniform environmental conditions. This method requires recognition of natural vegetation in nearby surroundings and reconstruction of potential natural vegetation at the focal site (Tomić et al. 2011). Species selection based on ecological differentiation is multidisciplinary approach relying on basic forest types -ecological units (Jović et al. 1998) -defined by three coordinates: dominant species, current vegetation, and soil properties. The local heat potential combines heat coordinates of slopes/aspects and altitude (Lujić 1960) and is usually used for afforestation of bare lands. The first heat coordinate (E) covers all possible combinations of slope and aspect, grouped in nine heat degrees, according to total annual sum of solar radiation (Tab. 5). The second heat coordinate (V) depends on altitude, with coldest sites on tree line (Tab. 6). As the result, the coldest sites will have values of coordinates of 1,1; and the warmest site will have combination of 9,9. When local heat potential is defined for focal planting site, it is necessary to find sites with the same local heat potential and natural vegetation. It is assumed that the dominant tree species at site with similar value of local heat potential will be well adapted for focal planting site.
Species selection is even more complicated when establishing mixed forests. For species with similar light requirements, small climate-induced differences in sapling growth can lead to significant differences in future species composition, but in mixtures of species with different light requirements, shade-tolerance has a more decisive effect (Ameztegui et al. 2015). In order to offer some tool for species selection, a hybrid Delphi-AHP methodology is proposed for species selection for reforestation in the Mediterranean region of Spain, which provides an optimal percentage distribution of the appropriate species to be used in reforestation planning (Curiel-Esparza et al. 2015).

Provenance selection
Long-term reforestation success depends on appropriate seed source (provenance) selection. Variation between populations may be clinal or ecotypic and knowledge on pattern of variation is important for seed transfer. Forest tree species are highly heterozygous and the most of total genetic variation can be found within populations (provenances). Provenances can, however, respond differently to climate change. Pseudotsuga menziesii (Mirb.) Franco growth response to climate change was dependent on seed source climate, with the mean temperature of the coldest month as the most sensitive indicator (Leites et al. 2012). The same study showed that all populations had optimum height growth when transferred to climates with warmer winters. Population differentiation (with high levels of genetic variation within populations) along temperature gradients is generally stronger for cold adaptation traits than for other quantitative traits and allozymes, indicating that these traits appear to be under strong natural selection (Howe et al. 2003).
Current seed transfer guidelines are based on the assumption that local tree populations are optimally adapted to environments in which they occur; this does not consider climate. However, this key assumption may no longer be valid . Changing climate conditions will complicate efforts to match seed sources with the site environments to which they are best adapted (Potter and Hargrove 2012). Two approaches for matching provenances to planting site facing climate change are suggested: climate envelope models and empirical response functions.
Climate envelope models compare the climate of seed sources and potential planting sites. Beaulieu and Rainville (2005) proposed a methodology combining a biophysical site index model and a seed source transfer model based on temperature and precipitation for identifying the most productive seed source of Picea glauca (Moench) Voss.  used bioclimatic modeling and multivariate approaches to identify the best matching seed sources for current and projected climates in Alberta, Canada. Potter and Hargrove (2012) developed quantitative ecoregion outputs in the form of similarity maps that can be used in cases of: 1) unidirectional seed transfer from a given location toward colder climate conditions, and 2) composite provenancing, for finding sources ecogeographically matched to focal site. In their meta-analysis based on long-term growth data of 2,800 provenances transferred to 120 European test sites, Isaac-Renton et al. (2014) used bioclimate envelope models developed for North America to guide assisted migration under climate change to retrospectively predict the success of these provenance transfers to Europe, with partial success. The model was generally successful in predicting the best performing provenances along north-south gradients in Western Europe, but failed to predict provenances with superior performance in Eastern Europe (Isaac-Renton et al. 2014). "An important criticism of bioclimate envelope models is that many wide-ranging species consist of locally adapted populations that may all lag behind their optimal climate habitat under climate change, and thus should be modeled separately" (Gray and Hamann 2013). For 15 wide-ranging forest tree species in western North America, Gray and Hamann (2013) found that on average populations already lag behind their optimal climate niche by approximately 130 km in latitude, or 60 m in elevation. They suggest using a general formula where a 100 km shift northward is equivalent to an approximately 44 m shift upward in elevation to guide assisted migration of planting stock in reforestation programs. This emphasizes the need for knowledge on species genetic variability at the provenance level and the importance for further research on tree species of SEE.
Empirical response functions are based on correlations of quantitative traits with climate variables using climate-response functions. This approach combines genetic and geographic information by analyzing geographic patterns of adaptive and neutral genetic variation. Wang et al. (2010) presented a single "universal response function" (URF), which integrates genetic and environmental effects, to predict the performance of any population of Pinus contorta Douglas growing in any climate. They concluded that URF can be used as a mechanistic model to predict population and species ranges for the future and to guide assisted migration of seed for reforestation, restoration, or afforestation and genetic conservation in a changing climate.  used multivariate regression tree analysis in two case studies for Populus tremuloides Michx. and Alnus rubra Bong. and concluded that this approach enabled a better-informed subjective decision on how seed transfer should be regulated. Finally, after comparison of the growth performance of the selected best performing populations with URF to populations identified through a climate envelope approach, Chakraborty et al. (2015) concluded that population recommendations based on empirical approaches were preferred.
In addition to matching the provenance to the focal site, many other provenancing strategies have been suggested as response to predicted climate change (Tab. 7). Breed et al. (2012) developed a decision tree for selection of a provenancing strategy (local, predictive, composite, and admixture) based on confidence surrounding climate change distribution modeling and data on population genetic and/or environmental differences between populations. No single strategy is likely to work universally, so selection of provenancing strategy should consider species genetic variation and local adaptation combined with climate projections for focal site. Where data on inter-population genetic variation are available. Prober et al. 2015 In SEE, additional research on genetic variation and local adaptation of targeted species is required, especially for those with small populations and limited or disjunctive ranges. Due to limited area and restricted south to north distances in SEE countries, foresters need to "look over the fence" in search of the best provenance to match with their focal sites. Seed transfer guidelines should allow and encourage movement of forest reproductive material across administrative and state borders.

Species-specific projects
Although species-specific projects are usually related to industrial tree plantations, we will focus on species translocation for conservation purposes. In these types of projects, the search is for the best site for focal species according to soil characteristics and projections of future climate. Facing climate change, the biggest challenge will be to match species to site, both spatially and temporally. For long-term success, future species and their sites must be selected based on predicted conditions that will support the species, realizing that conditions on those future sites may currently limit seedlings survival.
Species-specific projects often require colonization of new sites and areas where focal species are not currently found. This is possible thanks to fact that for many tree species the fundamental niche (conditions that a species can tolerate) is much larger than the realized niche (conditions where it is naturally established) . Realized niche of trees may be determined by the ability of seedlings to survive, but tree species can be successfully grown on new areas given to cultural treatments , which will be discussed in the section discussing short-term actions.
In SEE, long-lived, rare, geographically and genetically isolated species like Picea omorika (Pančić) Purk. and Pinus heldraichii H. Christ, are the most vulnerable to climate change. At the worst case scenario for them, the most appropriate sites should be found and colonized in order to prevent their expiration in the following decades.

Assisted migration
Assisted migration has been defined in the literature with number of terms (for detail review see Dumroese et al. 2015). For purpose of this paper, we will adopt definition of the assisted migration as intentional movement of forest reproductive material facing climate change. The practice of transferring forest reproductive material has a long tradition. Seed zones and regions of provenances have been established and seed transfer guidelines developed for most of the important tree species in SEE, but these will need to be adapted to meet projections of future climate change. Assisted migration can occur within a species current range, but species specific projects may require using assisted migration to rescue a species (Pedlar et al. 2012) or to transfer a species just beyond its current range (assisted range expansion) or quite distant from its current range (assisted species migration) as defined by Williams and Dumroese (2013).
Although assisted migration is controversial (Pedlar et al. 2012), its use has already been adopted by some forest administrations. The Canadian provinces of British Columbia, Alberta, and Quebec have altered seed transfer rules to allow assisted migration (Pedlar et al. 2011). Similar to any transfer of forest reproductive material, risks of maladaptation in assisted migration increase with movement distance. This distance can be geographic, climatic, and/or temporal and will depend on goals, the target species and populations, location, projected climatic conditions, and time (Williams and Dumroese 2013). In order to reduce risks,  propose that three conditions should be met before implementing assisted migration in reforestation programs: (1) evidence of a climate-related adaptational lag, (2) observed biological impacts, and (3) robust model projections to target assisted migration efforts.
Assisted migration is an appropriate strategy for facing climate change in the SEE, considering the differences in the pace of climate change and the natural migration of tree species. Large scale assisted migration is unlikely due to limitations in northwards and upwards distances but this strategy should be implemented in a small scale. Legislation changes will be needed to allow and define application of assisted migration at regional level, but these changes need to be supported by more research results for tree species growing in SEE.

Short-term tactics -Field actions
The long-term success of reforestation depends on field success of planting material. All long-term strategies and planning are wasted if seedlings fail to survive. Because of that, seedling survival rate is usually used as a measure of initial success of reforestation during the establishment phase (Ivetić 2015). This is why propagation and field establishment techniques must promote survival through seedling stress resistance and site preparation (Jacobs et al. 2015). In this section we offer a review of techniques that can promote seedling survival in SEE reforestation programs.

Planting material
High quality planting material (genetically, physiologically, and morphologically) improves the likelihood of reforestation success with the uncertainty of changing environments. Selection of planting material for any reforestation program depends on management goals, site conditions, and planting conditions, but unfortunately is too often solely based on the material available in nurseries. Between these selection criteria, site conditions will have a decisive effect on seedling survival. In afforestation projects, environmental conditions at the planting site can be diametrically opposed, e.g. rocky terrain and abandoned agricultural land. In reforestation projects, differences between site conditions are less pronounced, with the largest differences occurring between a mature forest site before and after clear-cut harvesting. Primary ecosystem functions such as energy, hydrologic, and nutrient cycles are altered in transition from climax forest to a clear-cut forest regeneration site (Grossnickle 2000), and this change can increase differences in environmental conditions between sites. These differences in site conditions should define the main goals of nursery production: vigorous, site adapted, and project specific seedlings.
Selection of the appropriate stocktype is important, especially on harsh site conditions. Stocktype basically describes a seedlings' age and production method, but generally production method has the stronger effect on seedlings quality and field performance. In their review on bareroot versus container stocktypes, Grossnickle and El-Kassaby (2016) found that container seedlings have a higher survival in a predominant number of trials (61%) than bareroot trials (15%). However, stocktype had a minimal effect under optimal site conditions (Sloan et al. 1987;South et al. 2005), while container seedlings generally have greater survival and growth on harsh sites (Barnett and McGilvray 1993;South et al. 2005).
The Target Plant Concept (Landis 2011;Dumroese et al. 2016) should be adopted and implemented in reforestation programs in SEE. This means that a project specific definition of seedling quality and nursery culture must be made to produce seedling with attributes that promote field survival and growth. Although physiological attributes can provide more accurate evaluation of seedling quality (Jacobs et al. 2004 a ;Davis and Jacobs 2005), in SEE nursery operation and reforestation programs, a seedling quality assessment is almost exclusively based on morphological attributes.

Morphological seedling quality attributes and field performance
Many studies show the relationship between seedlings morphological attributes at planting time and after planting success (Thompson 1985;Mexal and Landis 1990;Noland et al. 2001;Villar-Salvador et al. 2004 a ;Mexal et al. 2009;Oliet et al. 2009;Grossnickle 2012). Morphological attributes can forecast seedling field performance (Tab. 8) with different reliability and for different numbers of years after planting South et al. 2005;Tsakaldimi et al. 2012;Ivetić et al. 2016). Table 8. The synopsis of morphological seedling quality attributes effect on field performance.

Shoot height
Positive.

Nursery culture that promote target seedlings attributes
Better understanding and implementation of nursery cultural practices to improve seedling quality will enable better matching of seedlings to forest sites (Duryea 1984). In Table 9 we offer a synopsis of suitable nursery cultural practices for SEE nurseries that promote target seedling attributes and field performance. Some of these practices differ between species and stocktypes, e.g. between hardwoods and conifers and between bareroot and container seedlings. Table 9. The synopsis of the nursery cu ltu r e th at p ro m ote tar g et s e ed l in g s att r ib u t e s .

Pre-planting-site preparation
Site environmental conditions have a decisive influence on seedling field performance during the establishment phase (Hobbs 1992;Grossnickle 2000). The key to successful reforestation is quality site preparation followed by planting (Ehrentraut and Branter 1990). Pre-planting site preparation should provide advantage to the seedlings after planting, by controlling competitive vegetation, and by promoting access to water and root growth. A seedling's ability to overcome planting stress is affected by its root system size and distribution, root-soil contact, and root hydraulic conductivity (Grossnickle 2005 b ), which emphasizes the importance of proper soil preparation and water harvesting. It is a generally accepted principle in SEE that the objectives of site preparation should be achieved with minimal disturbance.

REMOVING OBSTACLES AND HARVEST RESIDUES:
Removing obstacles (i.e. stones) and residue (i.e. stumps) enables planting the total area with a regular layout and spacing, thus allowing easy movement on the site. Removing obstacles and residue can be done mechanically or manually; organic residuals after harvesting are usually burned either on-site or chipped and removed for energy production, which can reduce site preparation cost and generate extra revenue (Yoshida et al. 2015). As an alternative to their removal, rearrangement of obstacles and harvest residuals can promote seedling survival by creating the favorable micro-site conditions. Seedling shaded by a stump, log, or large rock tend to grow better on dry sites than those not shaded (Landis et al. 2010). The survival of Pinus sylvestris and P. nigra was remarkably higher when planted on the north side of spiny shrubs (Castro et al. 2002).
VEGETATION CONTROL: Competitive vegetation can be controlled chemically, mechanically, and manually. Chemical vegetation control in forest is limited or restricted in many regions. Vegetation management by non-chemical means is most critical and effective during site preparation as well as after crop establishment (Ehrentraut and Branter 1990). Mechanical site preparation often results in improved seedling survival and growth, but only intensive methods with much soil disturbance are an effective tool for controlling competing vegetation (Löf et al. 2012). Mechanical and chemical site preparation combined provides the best results. After five years of weed control treatment, site preparation by plowing and harrowing did not result in increased growth of Quercus macrocarpa Michx., but growth was superior when this mechanical site preparation was combined with a simazine herbicide application (Cogliastro et al. 1997).
SOIL PREPARATION: The usual methods of soil preparation in SEE countries are subsoiling, mechanical terracing, and mechanical or manual preparation of planting holes. On karst terrains in Croatia, afforestation by Pinus nigra and P. pinaster Aiton after subsoiling resulted in higher survival and growth rates compared to planting in pickaxe, and mechanically drilled holes (Tomašević 1994). Subsoiling with a bulldozer promoted higher survival of Pinus halepensis Miller (Barberá et al. 2005;Jelić 2012), Quercus ilex (Palacios et al. 2009, Jelić 2012, Cupressus sempervirens var. pyramidalis L., Pinus pinaster, and Pinus pinea L. (Jelić 2012) than holes made with an excavator in reforestation of a semiarid Mediterranean ecosystem. On some extreme sites with a shallow soil layer, explosives can be used for breaking the bed rock and creating an initial opening for a planting hole. At planting time, these holes are filled with soil from nearby forests. Used for afforestation of container Pinus nigra seedlings in Serbia, reported first year survival rate was 100% (Škorić et al. 1997) and 15 years after planting, this afforestation is considered successful (Fig. 2). WATER HARVESTING: Many techniques for harvesting water are described (Critchley and Siegert 1991;Prinz 1996;Mishra et al. 2011). In most of them, the basic principle is to make a reverse slope, usually done by different versions of bench terracing. Reverse slopes made during soil and planting spot (or hole) preparation can intercept runoff and redirect water to the seedling, and reduce runoff erosion by creating sinks along the slope (Fig. 3). Furthermore, the bench terraces significantly increase soil carbon stock compared to the soils between the bench terraces and soils planted by manual hole preparation (Lukić et al. 2015). Gradoni, a version of bench terracing, is a technique of soil preparation for afforestation used since the end of the 19 th century in France (Andrejević 1959) and Italy (Mercurio and Schirone 2015), and it was widely practiced in Serbia and FYR Macedonia. This method was used in Grdelica gorge in Serbia and provided more efficient soil erosion control compared to afforestation using planting holes without terracing (Lukić 2013). In addition, some different techniques, like diamond shaped negarim microcatchments (Critchley andSiegert 1991, Prinz 1996), characterized by small earth mounds, with an infiltration pit at the lowest corner, should be tested in SEE conditions.

FACILITATION BY NURSE PLANTS:
Some plants benefit from closely associated neighbors (Tab. 10), a phenomenon known as facilitation (Padilla and Pugnaire 2006). Nurse plants have been mainly used to restore vegetation in arid and sub-arid zones in recent years (Ren et al. 2008). Table 10. Reported positive effects of nurse plants to seedlings survival and growth.

Effect
Source Protect seedlings from frost. Stilinović 1991; LePage and Coates 1994 Reduce soil water evaporation, lower soil and air temperature, and decrease the amount of radiation reaching the plants by shading. Padilla and Pugnaire 2006;Endo et al. 2008 Improve the availability of water through the process known as "hydraulic lift". Padilla and Pugnaire 2006 Improve nutrient availability by nitrogen transfer between legumes and non-leguminous plants. Franco and Nobel 1989;Padilla and Pugnaire 2006;Rodríguez-Echeverría et al. 2016 Promote the development of differentiated soil microbial communities. Duponnois et al. 2011;Rodríguez-Echeverría et al. 2016 Promote survival and growth by mycorrhization. Bai et al. 2009;Bauman et al. 2012 Soil stabilization. Endo et al. 2008;Rodríguez-Echeverría et al. 2016 In reforestation of Mediterranean mountains, Pinus sylvestris and P. nigra survival was remarkably higher and growth unaffected when planted under individuals of the shrub Salvia lavandulifolia Vahl (Castro et al. 2002). Three years after planting, association between Cupressus atlantica Gaussen and Lavandula stoechas L. lead to a higher growth of C. atlantica and better soil microbial characteristics compared to the control treatment (Duponnois et al. 2011). Gómez-Aparicio et al. (2004) conducted a meta-analysis with seedling survival and growth data for the first year after planting at experimental reforestations with more than 18,000 seedlings of 11 woody species planted under 16 different nurse shrubs throughout a broad geographical area in southeast Spain. They concluded that facilitative effect was consistent in all environmental situations explored; but with differences in the magnitude of the interaction, depending on the seedling species planted as well as the nurse shrub species involved. Additionally, they found that nurse shrubs had a stronger facilitative effect on seedling survival and growth at low altitudes and sunny, drier slopes than at high altitudes or shady, wetter slopes. At dry sites with full sunlight, creating shadow by nurse plants can promote survival but reduce photosynthetic rate. Although shade enhances the probability of Pinus pinea survival, carbon assimilation reaches maximum values on more open sites (Calama et al. 2015).
On sites where no suitable nurse plants are already present, the simultaneous planting of targeted tree species and nurse-planting seedlings is recommended (Blanco-García et al. 2011). However, the selection of nurse plant species is not a simple task. The use of nurse plants is a dynamic system and what could start out as a beneficial strategy may turn detrimental with nurse plant strategy could become a vegetation management issue with unforeseen consequences and costs (Grossnickle, personal communication). Thus, nurse plant species should be featured by limited resources demands and growth. This indicates the need for research of nurse plants effect on tree species used in the range of environmental conditions specific for SEE.

Planting
All efforts invested in the selection and production of planting material and in site preparation can be rewarded or depreciated during planting. The full effect of the planting method may not be visible until the first stress event, even 10 or more years after planting (Stilinović 1991). Improper planting is one of the main reasons of reforestation failures in Serbia (Ivetić 2015 ; Fig. 4). A number of planting methods can be used (Stilinović 1991;Kloetzel 2004;Landis et al. 2010;Ivetić 2013), depending on tool and seedling stocktype. Assuming that appropriate planting is applied (e.g. putting seedling into the hole in an appropriate manner), several techniques are available to promote seedling survival and growth. Figure 4. Improper planting was observed with more than 20% of the total number of dead seedlings at 10 sites (Ivetić 2015). These seedlings are found dead after one growing season because of shallow planting and poor root contact with soil.

INCREASE OF PLANTING DEPTH:
Dimensions of the planting hole and seedling positioning depend on site conditions, species, and stocktype. Planting depth (i.e., distance between the root-collar and the groundline as defined by South 2005) can effect seedling survival. Heat, often confused with drought, and shallow planting are probably the most over looked causes of death of planted bareroot seedlings (Stroempl 1990) and shallow planting, regardless of taproot form, can promote seedling mortality (South 2005). Without rain, deeply planted Pinus echinata Mill. seedlings (11 cm below ground line) survived significantly better than seedlings planted with the root-collar slightly below the groundline (South et al. 2012). Deep-planting techniques promote seedling establishment by immediate exploitation of capillary fringe moisture (Dresen and Fenchel 2010). Effect of planting depth is species specific and the native habitat of a species should be considered (Bryan et al. 2010). Additional research on effect of planting hole dimensions and planting depth are needed for species and site conditions specific to SEE.

INCREASE OF NUTRIENT AVAILABILITY:
Field fertilization can improve performance of planted seedlings. An organic amendment significantly improved seedlings growth (Grossnickle and Reid 1983;Roldan et al. 1996;Barberá et al. 2005) with no apparent negative influence on seedling mycorrhization (Querejeta et al. 1998). Effect of organic amendment can be species specific. Grossnickle and Reid (1982) found that root system of 5-year-old Pinus flexilis James was not affected by the sewage sludge and wood-chips fertilization, but root system development of Pinus concorta and Picea engelmannii Parry. ex Engelm. was dramatically reduced. This reduction, however, was result of water stress rather than nutrient stress; because this ersatz soil created low soil bulk density conditions and reduced soil water movement to newly planted seedlings root systems (Grossnickle and Reid 1984). There is a concern that fertilization will stimulate competitive vegetation. Because of that, fertilization should not be broadcast over the site, but rather only applied close to the seedling root system. Field fertilization using controlled-release fertilizer has emerged as an effective means of promoting early growth of planted seedlings (Jacobs 2014). In order to use controlled-release fertilizers successfully, their formulation, release behavior, and environmental interactions must be understood (Rose et al. 2004). At dry sites, fertilization can result with root dehydration and limited water uptake, because of increased fertilizer salts in the soil solution (Jacobs et al. 2004 b ). Decision about field fertilization in any reforestation project should consider many variables, including soil conditions, competitive vegetation, site drought level, and seedling species demands. These variables, as well as ephemeral nature of field site fertilization (Grossnickle 2000), further emphasize that the decision making process must by project specific.

INOCULATION WITH MYCORRHIZAL FUNGI AND PLANT GROWTH PROMOTING RHIZOBACTERIA (PGPR):
Inoculation of seedlings with mycorrhizal fungi and PGPR can improve seedling survival and growth on sites with shallow or no soil profile. A combination of soil terracing, urban solid refuse amendment, and mycorrhization could be successfully applied in afforestation programs in semiarid and degraded sites (Roldan et al. 1996). Seedling mycorrhization at planting correlated positively with growth (Duñabeitia 2004) and survival (Óskarsson 2010). A simple and cost efficient technique for promoting mycorrhization is addition of forest soil to the planting hole. Although nursery inoculation with mycorrhizal fungi resulted in a faster growth rate, addition of forest soil to the planting holes to promote mycorrhization improved field growth of Pinus halepensis (Querejeta et al. 1998). Seedling performance can be significantly enhanced through PGPR inoculation of root systems (Chanway 1997), especially when moving from favorable to harsh sites (Chanway and Holl 1993;Chanway and Holl 1994). Additionally, seedling inoculation with PGPR is beneficial on contaminated soils (Babu et al. 2014;Karličić et al. 2016). Inoculation of Robinia pseudoacacia L. and Pinus sylvestris seedlings with PGPR increased seedling growth in coal mine overburden (Karličić et al. 2015). PGPR have a short-term, site specific effectiveness for reforestation of conifer seedlings, which necessitate matching PGPR strains to outplanting sites for effective growth promotion (Chanway et al. 2000). Although use of mycorrhizal fungi and PGPR has the potential to benefit seedling during establishment phase, the seedlingmicroorganism relationship does not always yield improved field performance (Grossnickle 2000).
AMENDMENT OF SUPERABSORBENT POLYMERS: Evidence suggests that application of superabsorbent polymers (hydrogels) during planting promotes seedling field performance. Polymer increased survival and leaf water potential of Pinus pinea seedlings in field conditions (Pery et al. 1995) and hydrogel combined with mycorrhizal inoculation promoted Picea abies (L.) Karst. resistance to drought stress (Višnjić et al. 2004). An innovative soil conditioner, comprising 23 ingredients including a new complex of hydro absorbent polymers, significantly enhanced tree growth in conditions of poor precipitation and/or thick textured soil with poor water holding capacity (Coello-Gomez et al. 2015). Šijačić-Nikolić et al. (2011) found that application of 5 g of polymer per planting hole increased two-year survival and growth of Pinus nigra and P. sylvestris seedlings. However, an excessive application of hydrogel increased mortality of Pinus sylvestris seedlings (Sarvaš et al 2007). Because contradictory results are reported, it is recommended that growers or planters conduct small trials to determine whether there are benefits to using hydrogels under their specific conditions (Landis and Haase 2012).
PLANTING LAYOUT AND DENSITY: Irregular planting pattern can complicate silviculture operations, but it can promote seedling survival and provide a close surrogate to successful natural regeneration that provides a more natural appearance of a landscape. Choosing the best planting spot is critical and more important than exact spacing (Landis et al. 2010), because initial seedling performance is related to microsite performance. On sites with harsh conditions, planting in groups on deeper soil is recommended, regardless of adopted regular planting pattern (Stilinović 1991). The risk of seedling mortality is not constant, but varies with tree species, planting density, tree age, and site conditions (Gadow and Kotze 2014). In most SEE reforestation programs, the prescribed planting density is 2,500 seedlings per hectare regardless of species characteristics and site conditions. The decision on planting layout and density should, however, be project specific with aim to promote survival and reforestation success.

Post-planting
The aim of post-planting silviculture during the establishment phase of plantation development is to promote success by improving microsite conditions and protecting planted seedlings. Post-planting treatments are essential for the success of seedlings establishing in droughty conditions (Klossas et al. 2012), but also on higher quality sites with risk of competitive vegetation and browsing. VEGETATION (WEEDS) CONTROL: At the planting site, weeds are those herbaceous and woody species that compete with planted seedlings for energy, water, and nutrients. However, not all existing vegetation at planting sites should be considered weeds (see Facilitation by Nurse Plants above), and not all of them should be controlled. At planting site, herbaceous vegetation control has more effect than woody vegetation control (Rose et al. 1999). Effect of existing vegetation on planting site can be twofold. Betula papyrifera Marshall reduces growth of shade-intolerant conifers but facilitates growth of shade-tolerant conifers (Simard and Vyse 2006).
Due to high costs and environmental issues (especially with herbicide use), decisions about vegetation control application should be based on a cost-benefit analysis. For example, growth and yield simulations using treatment-specific site index curves suggested that site preparation or post-planting vegetation control could reduce rotation length of Picea glauca by 12-16 years, but untreated areas were predicted to produce an equivalent volume if left to grow to mean annual increment culmination age (Boateng et al. 2009). Knowledge of the maximum (a level of vegetation cover where additional control will not increase tree performance) and minimum (a level of vegetation cover that must be reached before additional control will increase tree performance) response thresholds can be used to improve herbicide prescriptions (Wagner et al. 1989). A minimum response threshold level of 20% cover has been suggested (Wagner 2005). The response thresholds are species and site specific. Pinus contorta Dougl. var. latifolia Engelm. height and diameter growth and Picea glauca × Picea sitchensis diameter growth increased dramatically when cover of Rubus parviflorus Nutt. was below 5%, suggesting a response threshold (LePage and Coates 1994). Mortality of Betula pendula Roth and Pinus sylvestris seedlings increased significantly once the cover of competing vegetation reached 60% (Hytönen and Jylhä 2005;Jylhä and Hytönen 2006). Such thresholds should be considered during the decision making process on application of any method of vegetation control.
Another issue to consider is how many years after planting vegetation control is beneficial to seedlings performance. Intensive vegetation control in first couple of years after planting is critical for seedling field performance. The critical period in the establishing phase differs with species: 1-2 years after planting for shade-intolerant Pinus banksiana Lamb. and Pinus resinosa Sol. Ex Aiton, 1-4 years for more shadetolerant Picea mariana (Mill.) B.S.P. and Pinus strobus L. (Wagner et al. 1999). Rose and Ketchum (2003) found no observable effect of vegetation control in third year after planting on Pseudotsuga menziesii growth. Additional time aspect to consider is duration of vegetation control effect. Positive effect of weed control with herbicides last up to 11 (Hytönen and Jylhä 2005) and 30 (Wagner 2005) years after planting.
Vegetation control generally has a positive effect on seedling performance but efficiency depends on method and size of the control area. Competitive vegetation can be controlled physically by mulching, mechanically by cultivation, and chemically by herbicides (Tab. 11). Chemical vegetation control significantly improved seedling growth compared to the manual cutting (LePage and Coates 1994), tillage (Groninger et al. 2004), mulch and cover crop (Jylhä and Hytönen 2006). Combination of weed control methods can result in synergic effect. Rey Benayas et al. (2005) found a clear positive synergic effect of shading and weed mowing on seedling performance of three Quercus species (Q. coccifera L., Q. ilex, and Q. faginea Lam.).
Increasing in area of vegetation control has positive effect on seedling growth. Mean stem volume, basal diameter, and height of seedlings increased significantly with increasing area of vegetation control, and the magnitude of difference between treatments increased with time (Rose and Ketchum 2002).
The vegetation control efficiency depends on seedling stocktype as well: promoting survival of small and growth of large seedlings. Although the highest stand volumes of 15-years old Picea abies were obtained with the combination of large bareroot seedlings (4-year-old) and effective vegetation control, 2-year-old container seedlings because of their smaller size benefited more from vegetation control in terms of survival (Hytönen and Jylhä 2008). The volume return from increased weed control is maximized by planting the largest possible seedlings (Rosner and Rose 2006). On sites released from competitive vegetation larger seedlings grow quickly and occupy site resources during establishment due to greater level of incoming radiation and their greater photosynthetic capability (Grossnickle 2005 a ).
Chemical control of herbaceous competition in the first couple of years after planting promotes small seedlings survival and large seedlings growth. Decision on type, method, area, and number of applications should be made on species specific response thresholds and site specific conditions (e.g. vegetation cover, environmental restrictions, and erosion control).
MULCHING: Mulching is the spreading of material around planted seedlings to cover the soil with a goal of mitigating extreme temperatures, reducing evaporation and moisture loss, controlling weeds, and perhaps, enhancing soil structure and fertility. Some mulches are effective in some environments, some are not (McDonald and Helgerson 1990). Pseudotsuga menziesii seedlings grew best with treatments that promoted the most efficient use of available microsite water, either by reducing soil surface evaporation or vegetation competition (Flint and Childs 1987). The effect of mulching depends on material, size, cost, and longevity. Mulching positive effect on seedling growth increase with mulch size (McDonald et al. 1994;Harper et al. 1998). Although mulch mats controlled vegetation around trees as well as conventional herbicides, they were 7X more expensive (McCarthy et al. 2007). A variety of materials are used for mulching and they have a range of biological effects (Tab. 12). On one hand, the improvement of seedling survival and growth by mulching is stronger on high quality sites where seedlings can better utilized the rich soils, higher soil moisture contents, and reduced weed competition (Coello-Gomez et al. 2015). On the other hand, mulching can promote browsing. Quercus rubra L. seedlings grown with fabric mats had a greater frequency of deer browsing and a greater chance of dying than seedlings grown without mats, indicating that fabric mats should not be used in restoration projects with large deer populations (Stange and Shea 1998).
In addition to biological effect and price, the selection of appropriate mulching for reforestation programs in SEE should consider possible socio-economic effects. Some mulches, like those made of cardboard packaging (Fig. 5), can be easily produced in rural communities or by children in local schools helping their economics and raising awareness on environmental issues.

USE OF TREE SHELTERS:
Tree shelters provide physical protection and modify the environment conditions of planted seedlings. Tree shelters have proven to be a highly effective complement to vegetation control treatments (Navaro Cerrillo et al. 2005;Ceacero et al. 2014) (Fig. 6) and many studies report the positive effects of tree shelters on seedling performance: survival (Stange and Shea 1998;Oliet et al. 2005) and growth (Stange and Shea 1998;Chaar et al. 2008;McCreary and Tecklin 2001). Until the seedling grows above the shelter, height growth is favored compared to stem diameter (Oliet et al. 2005). Once this rapid, short term growth (Barberá et al. 2005) brings the terminal above the top of the shelter average height growth diminishes and diameter growth increases (McCreary and Tecklin 2001).
The effect of tree shelters depends on design, color, and size. Solid-wall shelters outperform screen cages (McCreary and Tecklin 1997), mesh or fabric sleeves (Ward et al. 2000), wire, and nylon mesh (Klossas et al. 2012). Shelter color can have a pronounced effect. For shade-tolerant Picea engelmannii planted at high elevation, the lighter-colored, solid-wall shelters yielded higher two-year survival (95 to 99%) than control seedlings (58%; shading seedlings using logging slash, stumps, and vegetation within the site) and the darkest, solid-wall shelter (5%); height and diameter growth followed the same pattern (Jacobs and Steinbeck 2001). Bellot et al. (2002) found that a 30-cm tall, brown plastic, solid-wall protector was most beneficial for total biomass growth of Quercus coccifera seedlings.
Various effects of tree shelters on seedlings physical environment are reported. Shelters increased maximum temperature by up to 10°C and decreased light intensity by 50% or more (Ward et al. 2000). In contrast, tree shelters color and venting did not influence air temperatures and only affected vapor pressure deficit late in the growing season (Devine and Harrington 2008). However, Laliberté et al. (2008) found that facilitation of hardwood growth was not caused by an improvement of tree water relations, but rather to an optimization of light levels inside the shelter: "low enough to lead to a photosynthetic system less costly to maintain due to a greater specific leaf area but high enough to have no adverse effects on photosynthetic rates". Environmental factors such as light availability need to be considered to optimize the effect of tree shelters. Tree shelters increased growth when shading by surrounding herbaceous vegetation was low, but reduced growth when surrounding vegetation blocked a substantial quantity of light (Laliberté et al. 2008). Shelters can increase water supply through dew harvesting. Both single and double wall design shelters had higher dew point temperature and lower minimum temperatures resulting in increased dew formation, but a significant increase in soil moisture was registered only with a singlewall shelter (del Campo et al. 2006). Even so, "seedling height growth in the single walled shelter was smaller indicating that only when soil moisture becomes over 6% does affect growth" of Pinus halepensis seedlings (del Campo et al. 2006).
Solid-wall, light-colored tree shelters can optimize light levels inside the shelter, initiate dew harvesting, and provide physical protection to seedlings during critical establishing phase of development. Given the wide variety of tree shelters on the market, shelters of different characteristics should be tested for specific combinations of species and site conditions.

BROWSING CONTROL:
Success or failure of reforestation often depends on control of animal damage. Compared to favorable sites where seedlings can resume normal growth more readily, the negative impact of animal damage persists longer on harsher sites because of slower growth. A variety of techniques have been used for browsing control, which generally can be divided into physical protection or chemical repellents. Although planted trees can be protected by fencing the planting site, individual seedling protection is prevailing in reforestation programs in SEE. Many tree shelters (plastic tubes) and guards (meshes, wires, etc.) are available on the market for individual seedling protection. Protectors were consistently effective in preventing browse damage without reduced survival or height growth (Schaap and DeYoe 1985). The use of plastic tree shelters prevented deer browsing and reduced the mortality rate of Quercus rubra seedlings from 35% to 3% (Stange and Shea 1998). "No Quercus pagoda Raf. seedlings in the tree shelters were browsed" (Dubois et al. 2000).
An additional, positive effect of tree shelters is promoting height growth (see Use of Tree Shelters above). Tree shelters can greatly reduce the time required by seedlings to grow above the browse line (McCreary and Tecklin 1997), after which the danger of animal damage is significantly decreased.
An alternative to physical protection is the use of chemical repellents that generally rely on fear, conditioned avoidance, pain, or taste Wagner 2000, Trent et al. 2001) (Tab. 13). New products are continually developed with variable efficiency depending on active ingredient, method of application, durability, and animal species. In general, repellents using fear as a mode of action were more effective than products using other modes of action (pain, taste, and aversive conditioning) and topical repellents were more effective than area repellents . Table 13. Delivery systems of chemical repellents (Adapted from Nolte and Wagner 2000).

Effect
Active ingredients

Fear of predator
Compounds that indicate predator activity emitting sulfurous odors (such as predator urine, meat proteins, or garlic).

Conditional avoidance
Compounds that cause nausea or gastrointestinal distress.

Pain
Compounds that cause pain or irritation, like: capsaicin, allyl isothiocyanate, and ammonia.
Although commercial repellents are effective for some species, they may not be cost-effective for most situations due to high cost . Low cost kitchen recipes can, however, be used as alternative (Fig. 7). These repellents are less effective and durable, but easy and inexpensive to produce. Figure 7. A homemade repellent from ingredients found in a local market. The basic ingredients are grained paprika containing capsaicin, and eggs. Milk, garlic, and other amendments can also be used. Hand sprayers with range of capacity can be used for application.

Conclusions
In the SEE, every reforestation project faces different challenges, due to uncertainty of climate change predictions and wide range of environment conditions. The biggest challenge to reforestation success will be frequency, duration, and severity of extreme weather events. Most of the tree species currently in SEE cannot keep pace with changes in environmental conditions on which forest populations are currently adapted. We reviewed a number of strategies and techniques that may promote reforestation success in changing environment. The toolbox is well-equipped and offers plenty of solutions.
Although we distinguish site specific from species specific projects, both rely on correctly matching site and species, and require different approaches to make that decision. Matching provenances to a planting site facing climate change requires additional research on genetic variation and local adaptation of targeted species, especially for those with small populations and limited or disjunctive ranges. Assisted migration is an appropriate strategy for facing climate change, considering the difference in pace of climate change and tree species natural migration rates, but large scale assisted migration is unlikely due to limitations in northwards and upwards distances available in SEE. This strategy should be implemented on a small scale. Seed transfer guidelines should allow and encourage movement of forest reproductive material across administrative and state borders.
A number of nursery cultural practices are available for SEE nurseries that promote target seedling attributes and field performance, depending on species and stocktypes, e.g. between hardwoods and conifers and between bareroot and container seedlings. Proper site preparation and planting techniques promotes reforestation success. Seedling benefits from microsite improvement by water harvesting, and facilitation by nurse plants. Inoculation of seedlings with mycorrhizal fungi and plant growth promoting rhizobacteria (PGPR), and application of superabsorbent polymers (hydrogels) during planting offer a variety of tools for promoting seedling's field performance; but require site/species specific testing. Vegetation (weed) control during first couple of years after planting is the postplanting treatment the most beneficial to seedling performance, but not all existing vegetation at the planting site should be considered competition. Complement techniques are mulching, tree sheltering and browsing control; with many available products available but still plenty of room for innovations and home-made solutions.
The need is constant for finding new solutions and testing existing strategies and techniques in conditions of SEE because no single solution or tool is best for all situations. Selection of appropriate strategies and techniques to achieve reforestation goals should combine knowledge on target species biological characteristics with focal site environment conditions and with climate change predictions.