Major perturbations in the Earths forest ecosystems. Possible implications for global warming [303810]

Accepted Manuscript

Major perturbations in the Earth's forest ecosystems. Possible implications for global warming

Remus Prăvălie

PII: S0012-8252(17)30629-3

DOI: doi:10.1016/j.earscirev.2018.06.010

Reference: EARTH 2649

To appear in: Earth-Science Reviews

Received date: 11 December 2017

Revised date: 25 May 2018

Accepted date: 12 June 2018

Please cite this article as: [anonimizat]'s forest ecosystems. Possible implications for global warming. Earth (2017), doi:10.1016/ j.earscirev.2018.06.010

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. [anonimizat], and review of the resulting proof before it is published in its final form. [anonimizat].

Major perturbations in the Earth's forest ecosystems. Possible implications for global warming

Remus PRĂVĂLIE1

1[anonimizat], Center for Coastal Research and Environmental Protection, 1 Nicolae Bălcescu str., 010041, Bucharest, Romania, [anonimizat]

Abstract

Forests are among the most important terrestria l eco logical systems in terms of the multitude of ecosystem functions and services they provide. These biotic systems are v [anonimizat], decarbonizing the atmosphere via carbon sequestration (in bio mass or underlying soil carbon pools) and evaporative cooling processes that mitigate climate warming. However, forest ecosystems are currently being subjected to a wide range of natural and anthropic disturbances that pose a [anonimizat]. Th is paper is a revie w [anonimizat] a [anonimizat], both the obvious (e.g. deforestation) and discrete/silent ones (e.g. defaunation) that have generally not yet been tackled strictly as ecological forest issues in the international scientific literature. [anonimizat], through carbon flu xes and biogeophysical feedbacks between these terrestrial systems and the atmosphere. Upon analysis of a [anonimizat] 12 ma jor forest disturbances that can be grouped into three categories based on the prevalence of triggering cau ses, i.e . climat ic ([anonimizat], die-[anonimizat]), anthropic (deforestation, frag mentation, a ir pollution) and mixed (defaunation, fires, [anonimizat], biogeochemica l shifts) perturbations. [anonimizat], [anonimizat]’capacity to stabilize the climate system. All identified d [anonimizat]chanis ms in the case of climat ic perturbations. Finally, this revie w paper proposes five major anthropogenic strategies to fight this multidimensional fo rest crisis – mitigate, adapt, repair, protect and research actions, which, if imple mented rapidly, e ffic iently and on a large scale via international policies, can successfully stabilize these terrestrial ecosystems and, implic itly, the climate system in the 21st century.

Ke ywor ds: fo rest perturbations; ecosystem functions and services; carbon sequestration; climate warming; positive feedbacks.

Introduction

Forest ecosystems cover ~30% of the global land surface area (Bonan, 2008). Given the various ecosystem functions and services they provide, forest ecosystems are essential factors for ensuring the stability of ecological and anthropogenic systems. Ecologica lly, they play a vital ro le in preserving global biologica l diversity, as it is estimated that tropical forests alone account for ~50% of the Ea rth's 5 to 20 million plant and animal species (Lewis et a l., 2015). For mankind, these ecosystems are essential due to the large-scale provisioning (e.g. food, wood, fresh water), regulating (control of c limate, flood, d isease), supporting (soil formation, nutrient cycling, prima ry p roduction) and cultural (spiritual, recreational, educational benefits) services they provide (MEA, 2005).

Given their capacity to decarbonize the atmosphere, forests are also a regulatory mechanism for the climate system. It is estimated that forests store ~45% of the terrestrial systems' carbon (C), account for ~50% of te rrestria l net p rimary production and are able to store over 25% of annual carbon anthropogenic e missions (Bonan, 2008; Anderegg et al., 2012), currently estimated to ~10 petagra ms (Pg) of C per year (or 10 billion tons C yr−1), when considering fossil fue ls (coal, o il, gas) and industry (cement) sources, and land-use changes (Le Quéré et al.,2016). Forests also mitigate global wa rming by means of evaporative cooling (a ir cooling determined by high evapotranspiration rates), as is the case of tropical ecosystems (Bonan, 2008; Jackson et al., 2008), considered to be the biome with the highest water exchange rate (through respiration) with the atmosphere (Lewis et al., 2015).

It is highly ala rming that these biotic systems have experienced several ma jor perturbations in less than a century, considering the key-role they play in the survival of biodiversity and man kind, as we ll as the fact they are one of the few Ea rth systems that can regulate the climate system. There are two ma jor causes for forest disruptions, i.e. c limate change and anthropogenic pressures, although the former , a lbeit indirectly, is also a form of anthropogenic pressure, considering the anthropogenic emissions that cause global wa rming (Abra m et a l., 2016; Rogelj et a l., 2016; M illa r et a l., 2017). While in the case of anthropogenic pressures the types of perturbations can be identified relat ively easily (they main ly consist of deforestation and habitat fragmentation), c limate change perturbations are significantly more co mp le x. There fore, in the context of climate change, a series of ma jor reports tackled a part of the forest changes triggered/amplified especially over the past decades in various global regions (USDA, 2000; Silvistrat, 2005; M EA, 2005; IUFRO, 2008; FAO, 2010). These reports approached, either directly or indirect ly, several ma in types of modifications/perturbations, e.g. disruptions related to phenology, primary productivity, range shifts and fires regime.

The consequences of these transformations are mu ltid i mensional and have global-scale imp lications. A telling e xa mple consists of the total global forest area, ~50% of which has disappeared over the past 300 years as a result of changes in land use (MEA, 2005). Including the impacts of frag mentation, this massive destruction of forest habitats had serious effects on biodiversity (Pfe ife r et al., 2017) – over 300 species of terrestria l vertebrates have become e xtinct over the past centuries especially due to deforestation (Dirzo et a l., 2014). It is currently estimated that the conversion of natural ecosystems (like forests) into agriculture or artific ial a reas accounts, to a large e xtent, for the general e xtinction rate of

>100 species per million species per year, currently estimated to be 100 to 1000 times higher than what is considered to be a natural rate (Rockström et al., 2009).

Another instance consists of the decrease in the ecosystems' capacity to provide basic services to human society, in the context of their anthropogenic alterat ion, e ither d irectly (especially deforestation) or indirectly (climate change). It is estimated that over 60% of global ecosystem services (terrestrial, ma in ly provided by forests, and marine) have deteriorated, especially over the past 50 years (Mooney et al., 2009), in the peak intensity Anthropocene phase called the

Great Accele ration (Steffen et al., 2011). A ma jor concern these perturbations raise with regard to the climate system consists of the decline in carbon assimilation cap acity. In the past decade (2006– 2015), it is estimated that global deforestations account for almost one fifth of carbon emissions (up to 1.5 Pg C yr−1, or ~5.5 Pg CO2 yr−1) (Le Quéré et al., 2016). Th is endangers the possibility to limit c limate warming under the 2 °C target set for this century (equivalent to a CO2 concentration <450 pp m), which is dee med essential for stabilizing the p lanet's ecological systems (Le Qué ré et a l., 2009; Molina et al., 2009).

Considering the mult iple negative, often catastrophic, effects of forest perturbations, there currently are several international policy initiat ives that directly or indirectly a im to stabilize forest ecosystems in a changing world. A highly relevant e xa mp le is the programme for Reducing Emissions from De forestation and forest Degradation (REDD +) in developing countries, init iated in 2008 by the Un ited Nat ions (UN) Fra me work Convention on Climate Change – UNFCCC (FAO, 2014a ). Th is programme is a financial mechanism (ma inly financed by developed countries) imp le mented globally in order to reduce e missions in developing countries through sustainable forest management, forest protection measures, and/orlarge-scale carbon biosequestration (Phelps et al., 2010).

This paper is a revie w that prima rily a ims to analyse the main current perturbations of global forest ecosystems by briefly, yet holistically, tackling forest changes prompted by climate change and anthropogenic interventions. A secondary objective is to concisely assess the possible effects these perturbations generate in the process of global warming, through carbon flu xes and biogeophysical feedbacks. Based on recent relevant scientific sources, the paper aims to approach this particularly vast issue in an integrated manner (in terms o f the d isturbances ' causes – climat ic, anthropogenic or mixed), considering that the identified perturbations have not been analysed in a similarly broad context in specialized papers.

General considerations on forests on a planetary scale

Forests are unanimously defined as lands with a tree cover (at least 5 m high) of at least 0.5 ha, and a canopy cover of more than 10% (FA O, 2000). While there are several important definitions adopted by international environmental/forestry organizat ions , they generally delimit forests based on these tree characteristic thresholds (Chazdon et al., 2016). The definitions include natural and planted forest areas, as well as areas that are temporarily unstocked with trees due to human or natural causes, but which are expected to revert to forest (Chazdon et al., 2016).

Globally, these lands dominated by trees and other bio logical co mmunit ies are attributed, based on their latitudes, to three forest biomes, i.e. t ropical, te mperate and boreal forests. The three biomes total ~40 mil km2 globally (Keenan et al., 2015). The world's forests are ma inly located in Asia (31% of the global fo rest area, which a lso inc ludes Asian Russia), which is followed by South A merica (21%), North and Centra l A merica (17% ), Africa (17%), Europe (9% ) and Oceania (5%) (Pan et a l., 2013). It is interesting to note that 5 countries alone (Russian Federation, Bra zil, Canada, the United States and China) hold 53% of the globe’s total forest area (FAO, 2010). The Earth's forests account for 80% of the planet's total plant biomass , and the amount of carbon they store in biomass and soil is greater than the one currently present in the atmosphere (Pan et al., 2013). It is estimated that the total amount of carbon stored in these terrestrial systems is of 861 ± 66 Pg C, 44% of wh ich is stored in the soil, up to a 1 m-depth, 42% in live b io mass, 8% in deadwood, and 5% in litter (Pan et al., 2011).

Tropical forests total almost 20 mil km2, i.e . ~50% of the global forest area (Pan et a l., 2011). This forest bio me 's e xtent ranges from 30 °N to 30 °S, and it consists of tropical and subtropical mo ist broadleaf forests, tropical and subtropical dry broadleaf forests , and tropical and subtropical coniferous forests (Fig. 1a). Tropical and subtropical mo ist

broadleaf hold by far the largest share of tropical forests (Fig. 1a), and are present in areas with mean annual temperatures of at least 24 °C (in most cases) and precipitation that generally exceeds 2000 mm (Fig. 1b,c).

Special attention must be paid to tropical and subtropical mo ist broadleaf forests, the extent of which is la rgely influenced by climate conditions (humidity). Studies show these forest ecosystems can be bimodal in South Ame rica and Africa, in areas with a mean annual rainfa ll that ranges from 1000 to 2000 mm (or even up to 2500 mm in certain parts of South America), where forests and savannas can exist as alternative stable states (Staver et al., 2011a,b ) (Fig. 1d). It was found that at this intermediate ra infa ll level, fire beco mes a strong predictor of savannas/forests spatial distribution, which promotes savanna and open canopies, if p resent, and tree cover expansion and closed -canopy forests, if absent. In fact, it was suggested that fire ma inly conditions the forest or savanna state in these areas with intermed iate ra infa ll a mounts (Fig. 1d), where c limatic and edaphic conditions could generally support a c losed -canopy forest (Staver et a l., 2011a,b). As a result, in the case of forest-savanna bistability, fire is less important under the 1000 mm threshold, where lo w rainfa ll limits tree cover and savanna could dominate without fire, but also above 2000 mm, where fire is rare and tropica l fo rests strongly persist (Staver et al., 2011b ). Under conditions o f well-known fire-induced bistability (Bond, 2008; Hoffmann et al., 2012; Staver and Levin, 2012; van Nes et a l., 2014; Wuyts et al., 2017), in which both forests and savannas are present in distinct (bimodal) states, it is therefore difficult to delimit the e xact e xtent of tropical fo rests, which can be larger in the 1000– 2000 mm rainfall range (Fig. 1d) in the years with low fire activity.

Regardless of these spatial uncertainties , tropical forests are known for their ext re mely varied biodiversity, which enables them to provide a wide range of ecosystem services on which 1.5 billion people are d irect ly dependent (Le wis et al., 2015). The most representative exa mp le for illustrating the ecosystems' biodiversity consists of Amazonian tropica l mo ist (rain ) forests, which are the richest assemblage of forests on Earth (ter Steege et al., 2013). They bring together 16000 tree species, of which only 227 are the so-called "hyper-dominant" species (like Euterpe precatoria, Pseudolmedia laevis, Eperua falcata, Hevea brasiliensis, Attalea butyracea and many more) (ter Steege et al., 2013).

Fig. 1. Global spatial d istribution of forests (and of other terrestria l ecoregions) (a ) and of mean mult iannual values (1950– 2000) for te mperature (b) and precipitation (c), wh ich are the climate para meters that have the highest influence on the spatial distribution of forest biomes; spatial representation of 1000– 2000 mm intermediate ra infa ll range, in which tropica l forests (tropical and subtropical mo ist broadleaf forests) and savannas can exist as alternative stable states (bimodal states) (d). Note: the e xtre me precip itation amount (over 11000 mm) is specific to a very limited area in Southeastern Asia (Northeastern India), while this para meter’s peak global values typically reach ~4000 mm; the sources of the mapped data are http://www.worldwildlife .org/publications/terrestrial-ecoregions-of-the-world for g lobal ecoregions, and Hijmans et al. (2005) for temperature and precipitation data.

A particularly impo rtant characteristic that influences the stability of the Earth's climate system consists of the fact that these forests store massive amounts of carbon, i.e. 471 ± 93 Pg C (ca rbon stored in the soil up to a depth of 1 m, in live biomass, deadwood and litter), or 55% of the total global forest stocks of carbon (Pan et al., 2011). At the same time, it is estimated that tropical intact forests (which account for two thirds of tropical forests ) assimilate through photosynthesis large a mounts of at mospheric carbon (1.2 ± 0.4 Pg C yr−1 for 1990 to 2007), equivalent to 50% of the global forests' annual assimilation. However, it must be noted that the net carbon balance in all tropical forests consists of emissions , as a result of higher carbon leakage generated by anthropogenic activities (deforestation) in non -intact forests (Pan et al., 2011).

Considering these forests' role in the mitigation of c limate warming by means of the evaporative process (more specifically through latent heat flu x that generates cooling, but also through the formation of clouds that reflect incoming radiat ion bac k to space) (Bonan, 2008; Jac kson et al., 2008), in addit ion to their carbon storage role, it can be argued that these ecosystems currently generate a strong climate control and therefore contribute to the stability of the global c limatic system to a considerable extent (Fig. 2).

Temperate forests total ~8 mil km2 (Pan et a l., 2011), and account for appro ximately 20% o f global fo rests. They are present between 25– 60 °N and 25– 55 °S and consist of temperate broadleaf and mixed forests , and temperate coniferous forests (Fig. 1a). Te mperate broadleaf and mixed forests are the largest (Fig. 1a), and they feature, for instance, numerous Fagus species in Europe and No rth A merica, and dominant Eucalyptus and Nothofagus species in the southern hemisphere (Kirschbaum and Fischlln, 1996). Although temperate forests are characterized by a much lowe r biodiversity co mpared to tropical forests, they hold notable records such as the tallest trees in the world, e.g. Sequoia sempervirens trees, which are found in the coastal areas of north-western North America (Pan et a l., 2013). Te mpe rate forests are found in regions with mean annual te mperatures of about 5–17 °C (genera lly for te mperate broadleaf and mixed fo rests, as, according to the figure 1b, te mperatures in temperate coniferous forest areas drop considerably below 5 °C in Western Canada or Southern Asian Russia), and with annual precipitation that generally e xceeds 500 mm (Fig. 1b,c). Historica lly, th is biome has been the most heavily modified by anthropogenic activities (M EA, 2005). They are particula rly impo rtant for the climate system, as they account for 119 ± 6 Pg C (14% of the total), while the net annual carbon assimilation rate is 0.7 ± 0.1 Pg C, a fter 1990 (Pan et a l. 2011). Ho wever, unlike tropica l forests, it can be argued that this bio me generates a moderate control over the climate system (Fig. 2), as a result of the moderate air cooling process determined by evapotranspiration rates, which are not particularly high in temperate regions.

Fig. 2. Representation of climat ic control mechanisms generated by global forest biomes. Note: the images of trees in this figure were sourced from Bonan (2008).

Borea l forests total over 11 mil km2 (Pan et a l., 2011), i.e. ~30% of the global forests . They range between 45–70 °N latitude (Fig. 1a) and are adapted to the severe climatic conditions of these high -latitude environments , with mean annual temperatures fro m about -10 to 5 °C (or even below -10 °C in Northeastern Asia) and annual precipitation a mounts generally below 500 mm (notable e xceptions are boreal forests in Eastern Canada and Northern Europe, that receive precipitation above this threshold) (Fig. 1b,c) . The diversity of tree species is low (especially Picea, Pinus, Abies and Larix species), and only few are widespread, e.g. Picea mariana, Picea glauca, Picea abies, Pinus ponderosa and Pinus sylvestris. The anthropogenic influence is also apparent in terms of fo rest management, mostly for industrial wood production, and it is estimated that two thirds of these forests are managed in regions such as Fennoscandia (90% ), Russia (58%) or Canada (40% ) (Gauthier et al., 2015). Borea l forests (or the taiga) hold la rge carbon pools (272 ± 23 Pg C, 32% of the total), e xt racting 0.5 ± 0.1 Pg C fro m the atmosphere between 1990 and 2007 (Pan et al., 2011). It was suggested that this biome generally has a less important ro le in c limate control, considering the moderate ca rbon storage rate via photosynthesis, weak evaporative cooling and strong albedo decrease (i.e. strong sun absorption) (Bonan, 2008) . It can however be stated that boreal forests have at least a moderate role in climat ic stability (Fig. 2) especially g iven the large underlying carbon reservoirs in soil, peat and permafrost deposits (Gauthier et al., 2015).

Types of major forest perturbations in the global context

Forests are currently subjected to the pressure of a wide range of disturbances all over the world. Based on a recent, comple x and credible conceptual approach (Tru mbore et al., 2015), forest perturbations can be concisely defined as all the disruptive factors that directly or indirectly threaten the ecosystems' health, functions and services, or the very e xistence of forests locally, regionally or globally. These recurrent stressors can act individually or synergistically, significantly endangering the forest systems' capacity to return to pre-disturbance levels, by prolongedly affecting e xtre me ly important attributes such as biomass, biodiversity or carbon storage functions, which need decades to centuries to return to their in itia l states (Trumbore et al., 2015). At the same t ime, against the background of growing natural and anthropogenic perturbation frequencies, intensities and variety, the loss of mult iple intrinsic characteristics of forests can neutralize the systems' ability to recover, wh ich will inc rease the likeliness of abrupt transitions from forests to new vegetation states (Brando et al., 2014; Millar and Stephenson, 2015; Trumbore et al., 2015).

There currently a re several h ighly impo rtant international sources that systematically describe various forest perturbations globally (USDA, 2000; Silv istrat, 2005; M EA, 2005; IUFRO, 2008), the most notable of which are FA O Global Forest Resource Assessment reports (FAO, 2005, 2010, 2015). Ho wever, unlike this paper's objectives, these sources generally don't focus e xclusively on forest perturbations, don't include all types of disturbances (e.g. defaunation or biogeochemical changes ), don't clearly classify changes based on their driving forces (climatic, anthropogenic, mixed ) and don't analyse them in re lation to the imp licat ions /feedback they generate in the global climate system. Therefore, based on the information e xtracted fro m nu me rous and generally recent scientific papers, this study focuses on 12 ma jor perturbations of forest ecosystems, grouped in three categories based on the dominance of the major t riggering factors, i.e. climatic (phenological shifts , range shifts, die-off events, insect infestations), anthropic (deforestation, fragmentation, air pollution) and mixed (defaunation, fires, co mposition shifts , net prima ry productivity shifts, biogeochemical shifts) perturbations (Fig. 3). Therefore, wh ile for mixed perturbations it was estimated that climatic and anthropogenic factors had approximately equal contributions (Fig. 3), in the other two cases it was found that forest changes were due, for the most part, to either climate-re lated or anthropogenic causes, even though the two factors may at times overlap. Fo r instance, while the main cause for deforestation is by far human activity (wh ich is why this change is classified as an anthropic perturbation), it is also known that e xtre me climat ic events (e.g. drough ts) can contribute, in certain cases, to this type of forest dynamics.

Fig. 3. Sche mat ic approach on forest perturbation types (identified based on the dominant influence of triggering factors) currently occurring g lobally. Note: the yellow-red grad ient e mpirica lly highlights the approximately equal influence of climatic and anthropic factors on triggering mixed perturbations.

At the same time , these changes were called perturbations due to the fact that, in all considered instances, they can result, directly or indirectly, in the large-scale decline of the forests' functioning state. Although in the aforementioned list the net primary p roductivity can partially be considered to be an e xception fro m th is general conte xt (as higher at mospheric CO2 concentrations may stimu late ecosystem productivity), this change in forest functionality is c lassified as a perturbation because there is no current generally agreed-upon stance on the increase or decrease of this ecological para meter. For instance, a highly relevant study on global net primary productivity dynamics found that, over the past decade, the Earth's

forest area recorded a net biomass carbon flu x decrease (Zhao and Running, 2010), which had negative ecological and climatic effects (therefore, this change can be considered a disturbance).

Also, certain forest perturbations such as windstorms were not included in this revie w because, while they do have the potential to negatively affect forest ecosystems, their effects cannot be compared to the ones the 12 perturbations generate, in terms of e xtent and severity. This was confirmed, for instance, by a recent study that tackled global forest cover changes between 2000 and 2012, which restated that forest losses caused by windstorm da mages are specific to a fe w a reas world wide (Hansen et al., 2013). The 12 perturbations were therefore selected in order to highlight the most significant global changes (in terms of frequency, intensity and especially e xtent) in forests, which a re currently able to greatly disrupt or even completely eliminate forest ecosystemservices and functions .

A brief cla rification is a lso necessary for these perturbations' imp licat ions for global warming, wh ich are a secondary objective for this analysis. They ess entially consist of the changes' long-term e ffects on atmospheric (troposphere) temperature dynamics . They were on ly analysed as consequences for global wa rming if perturbations were triggered directly by anthropogenic activities , such as deforestation and fragmentation. In contrast, for forest perturbations driven directly by climate wa rming itself, their e ffects were ca lled positive (wa rming) or negative (cooling) c limate system feedbacks, based on the positive or negative influence of forest changes on temperature dynamics. In other words , only climatic forest disturbances were analysed considering the two types of feedback, as carbon and radiative flu xes triggered directly by land use changes or other anthropogenic activities (indiv idually or in synergy with climate change) cannot be considered feedbacks in climate warming (or cooling) (Schime l et al., 2015). In the end, both types of effects (consequences and feedbacks) were approached as "implications" for climate warming.

Climatic perturbations

Phenological shifts

It is believed that plant phenology (the period of recurring life h istory events) is the most sensitive and directly noticeable change in vegetation in a reaction to climate change (Linderholm, 2006; Sherry et a l., 2007; Bandoc et al., 2017) . The mean g lobal te mperature increase of 0.2 °C per decade in the past ~30 years (Hansen et al., 2006) caused significant shifts in phenological events, such as an increase in the growing season length – GSL (the timing between vegetation growth onset and leaf fall), one of the most important indicators of phenological changes (Cle land et al., 2007). A lthough GSL changes in ecosystems (both forest and other types) are not uniform g lobally, it is unanimously accepted that this ecological indicator has grown steadily after 1970, with a general magnitude of 3 to 4 days per decade (Peñuelas et a l., 2009). While this is ma inly due to spring phenological events, which have advanced by 2– 3 days per decade, phenological disturbances of the past decades are also due, in part, to the delayed onset of autumnal phenological activity, wh ich was delayed by 0.3– 1.6 days per decade (Sherry et a l., 2007). According to a reliable revie w paper, the total increase of the GSL in the past half-century can roughly be defined in a 10–20-day interval for most global regions (Linderholm, 2006).

Phenological changes are currently a world wide ecolog ical perturbation. A recent study signalled that forest vegetation phenology (and of other types of vegetation) changed dramatica lly between 1981 and 2012 on 54% of the Ea rth's land surface (Buitenwe rf et al., 2015). The study found significant changes in phenological activ ity especially in the boreal and northern te mperate regions (Buitenwerf et a l., 2015), which are already being affected by adverse consequences in terms of terrestrial-climate feedbacks and ecosystem functionality (Post et al., 2009; Callaghan et al., 2011; Bokhorst et al., 2016). At the same time, it was found that phenological changes in forest ecosystems in te mperate areas and at high

latitudes are ma inly regulated by te mperature and, in nu merous cases, by the photoperiod (length of vegetation e xposure to sunlight) (Linderholm, 2006). In contrast, tropical ecosystems are less sensitive to these climatic variables , but are also more heavily influenced by precipitation variability (Cle land et al., 2007). The influence of El Niño-Southern Oscillation on Amazonian forest phenological shifts is a remarkable example in this respect (Asner et al., 2000).

There are several majo r implications of phenological changes for climate change, determined by the vegetation- atmosphere interaction (Fig. 4). One o f the first consequences of a longer growing season of forest ecosystems consists of a more intense process of CO2 storage in bio mass (together with the higher carbon uptake period pro mpted by the photosynthesis process ), which results in one of the most important negative feedbacks for c limate warming (Peñuelas et al., 2009). For instance, it was found that a 5–10-day increase in the GSL can a mp lify net primary productivity of forests by

~30% (Cleland et al., 2007). The growing season lengthening mechanism is currently active especially in northern high latitudes, where climate warming accounts for the higher primary productivity and promotes plant growth (Zhao and Running, 2010).

On the other hand, a relat ively recent study signals that, in the past two decades, terrestrial ecosystems at northern latitudes (which ma inly consist of boreal forests ) have released large carbon amounts as the growing season lengthened during autumn, against the background of the vegetation respiration increase e xceeding that of photosynthesis in this season (Piao et al., 2008). The research showed that, although spring climate wa rming accounts for a more notable increase in photosynthesis compared to the respirat ion process in this season , autumn warming can result in carbon losses due to respiration of up to 90% of the carbon sequestration increased during spring warming, by means of photosynthetic activity (Piao et al., 2008). Th is can be e xpla ined by the fact that the net productivity of boreal conife rous forests is light-limited during autumn, considering that, during this season, at high latitudes, warming is associated with more cloud cover and less solar radiation (Richardson et al., 2010). However, in contrast with these findings, a study showed that in certain forest areas in the temperate zone carbon storage through photosynthesis exceeds carbon release through respiration for both earlier spring and later autumn phenology (Keenan et al., 2014).

Another important climat ic effect of e xtended phenological activity is the higher emission of biogenic volatile organic compounds into the atmosphere, which can generate organic ae rosols on a large scale. This negative feedback mechanis m can generate climatic cooling through the format ion of cloud condensation nuclei (and, therefore, through the interception of solar radiation in the atmosphere), as well as through the diffusion of light intercepted by tree canopy, thus amp lifying CO2 storage in forest biomass (Claeys et al., 2004; Peñuelas et al., 2009). While evapotranspiration increase (associated with growing season lengthening) is another mechanism that can attenuate climate wa rming by means of evaporative cooling, there are however climate warming intensification e ffects as we ll, associated with longer phenological activity. For instance, an increase in biogenic volatile organic co mpounds can generate an enhanced greenhouse effect by stimulat ing atmospheric ozone production and methane lifetime (Peñuelas and Llusià, 2003). Moreover, a lowe r albedo in snow-covered areas, especially in borea l forests (where taller and darke r canopies, due to increased phenological activity (Fig. 5), a lready absorb more solar radiation), is one of the most important instances of positive feedbacks of green-cover lengthening in climate warming (Bonan, 2008; Peñuelas et al., 2009).

Range shifts

It is known that forests and plant species in general have a certain physiological tolerance to changes in climat ic parameters, especially te mperature. As climate warming intensifies globally, the response of forest species consists of

spatial redistribution by generally shifting towards northern latitudes and upward to higher elevations, where the lowe r temperatures allow the m to re main in the ideal range of environ mental conditions (MacDonald et a l., 2008; Lenoir et a l., 2008; Ju mp et a l., 2009; Chen et al., 2011; IPCC, 2014). It was therefore signalled that, in borea l and te mperate ecosystems, forest species change their habitat both latitudinally and altitudinally, while in the tropical region climate warming ma inly leads to elevational shifts in the distribution of species (Colwell et al., 2008).

A recent meta-analysis showed that living species (especially fauna) ac ross extensive regions, such as Europe, North America and South-eastern Asia (Malaysia), had recently migrated with a med ian rate of 1.7 km per year to higher latitudes, and with 1.1 m per year to h igher e levations (Chen et al., 2011). Un like anima l species, trees face a ma jor difficulty in that their dispersion potential is notably lower (IPCC, 2014). More specifica lly, their migrat ion rate can generally reach several hundreds of meters per year latitudinally (e.g. over 200 m/year for map le and beech trees in North America, during the postglacial wa rming period) (Mc Lachlan et al., 2005), or a fe w meters per year a ltitudinally (e.g. ~3 m/year, on average, over the past decades in temperate forests in Western Europe) (Lenoir et a l., 2008). Another meta- analysis conducted globally, based on recordings of treeline dynamics in 166 sites (most of wh ich were located in the arct ic region), showed that, since 1900, tree lines advanced (at higher a ltitudes/latitudes) in 52% of cases, re mained stable in 47% , and retreated in 1% of cases (Harsch et al., 2009). While this general migrat ion process is main ly associated with increasing temperatures in the summe r and especially winter seasons, it was noticed that treeline advance is not universal against the background of an overall warming recorded in the two seasons, considering temperature variab ility (e.g. a single year with very low te mperatures in the arctic region can nullify the tree line advance generated by several previous warme r winters) and especially the influence of other factors on their dynamics such as precipitation, photoinhibition, topography, local climate or species diversity (Harsch et al., 2009).

At present, probably the most important consequences of vegetation migration are occurring at h igh lat itudes (Fig. 4). Range e xpansions of boreal trees and shrubs into tundra biome are responsible for various ma jor transformations this biome is e xperiencing, such as herbaceous species richness decline (Myers -Smith et a l., 2011), an ima l invasions (Post et al., 2009), changes in the nutrient circuit (Bucke ridge et al., 2010) and in surface energy balance and hydrology (IPCC, 2014; Myers-Smith et a l., 2015). Even though climate warming is currently not strong enough to trigger a large -scale e xpansion of boreal forests and shrubs in tundra (Post et al., 2009), the temperature increase expected in the ne xt decades can significantly accelerate these transformations. A relevant exa mp le in this respect consists of the amplification of ecological perturbations such as caribou and reindeer herd losses, which are e xtre mely important in the arctic region in terms of ecologica l interactions, as well as for ensuring the survival and the cultural integrity of loca l human co mmunit ies (Post et al., 2009). In this instance, a significant decline in these anima l populations was already noticed, la rgely due to the lichens' decline (primary winter forage for these animals), ma inly triggered by shrubs and trees cover increase in tundra ecosystems (Joly et al., 2009; Vors and Boyce, 2009).

Certa in important changes also occur at low lat itudes. Considering that, in the tropics, the elevational te mperature gradient is far higher than the latitudinal one (alt itudinal decrease of over 6 °C/1000 m, a change rate far higher than the temperature decrease rate latitudinally) (Bendix et a l., 2008; Co lwe ll et a l., 2008), it is to be e xpected that forest species (and vegetation in general) will be subjected to upslope range shifts in order to maintain their therma l niches under the pressure of c limate warming. Under these conditions, tropical mountains are vita lly important for the survival of tree species that are already at the limit of their therma l optimu m. Ho wever, many forest commun ities located near the top of

elevational gradients are particularly vulnerable to climate change, also due to the fact that many tropical mountain ranges (especially the ones in Africa and Asia) and currently insufficiently well protected (Elsen et al., 2018).

In the 21st century, forest range shifts will la rgely depend on the velocity of te mperature change, wh ich regulates the species' migration direction and speed across local terrestrial areas in order for them to ma intain constant temperature conditions (Loarie et al., 2009; Burro ws et al., 2014). Climat ic mode ls (previous A1B e mission scenario) showed that, in this century, almost 30% of the world's land a rea will e xperience te mperature change velocit ies of over 1 km per year (Fig. 6). It was also suggested that the changes will be faster than tree migration rates, which generally fa ll under this threshold during the Holocene (Loarie et al., 2009). It seems however that, in certain limited areas in tropical (e.g. A ma zonia, in Western Bra zil, Eastern Co lo mbia and Peru), te mperate (parts of Centra l-eastern US) and boreal (south of Hudson Gulf in Canada and Western Siberia in Russia) forests, these velocities could reach and even exceed 5 km per year (Loarie et al., 2009). The e mergence of a h igh number of mig ration corridors (areas with rap id thermal shifts and high proportion of species moving convergent trajectories ) for living systems is therefore e xpected, mostly in borea l forest regions (Burro ws et al., 2014). To keep pace with these temperature changes, forest vegetation and plant species in general will not only have to use their dispersal abilities, but also other survival mechanisms such as evolutionary adaptation or new ecological interactions (Diffenbaugh and Field, 2013). Most probably, human interventions will also be necessary for large-scale forest displacement, e.g. for tree species with high economic value (IPCC, 2014).

The ma jor consequences of forest vegetation migration for the global climat ic system are in this instance as well representative for boreal fo rests (Figs. 5, 6). It is known that the currently incip ient arct ic tundra co lonizat ion by trees (and shrubs) has amplified snow cover decrease and has considerably influenced albedo decrease, as tree canopy absorption of incoming rad iation increases, and less solar rad iation is reflected into the atmosphere (Myers-Smith et al., 2011). This snow cover decrease, in synergy with marine a lbedo decrease effects, due to the ma jor decline of sea ice over the last decades (when the seasonal min ima l sea ice e xtent retreated at an alarming rate of 45000 km2/yr) (Post et al., 2009), is la rgely responsible for the so-called "Arctic a mp lification" – ma jor arct ic warming feedback characterized by higher te mperature increase rates compared to lower latitudes (Macias-Fauria et al., 2012; Bokhorst et al., 2016).

Furthermore, the albedo decrease – climate warming interaction can a lso result in another ma jor threat to the climate system, name ly perma frost degradation. The permafrost covers an immense share of the terrestrial northern hemisphere (~22.8 mil km2, o f wh ich 37% in Canada and Alaska, and 63% in Eurasia), and it can cause disastrous climate wa rming – related effects due to the fact that it stores large amounts of CO2 and especially CH4, wh ich has a greenhouse effect that is 25 times more aggressive than that of CO2 (Zhang et a l., 2008). However, it is interesting that certain scientific findings signalled that, in the Siberian tundra, shrub expansion (Betula nana) may reduce perma frost thaw during summe r, due to the shading effect of shrub canopies, which significantly lowe rs the ground heat flux and soil te mperature (Blo k et al., 2010). Nonetheless, even considering the triggering of th is local soil cooling effect, it is highly like ly that the net climate feedback of large-scale e xpansion of shrubs and trees will be wa rming, due to the greater impact generated by a lower albedo, which will determine a higher permafrost vulnerability to thawing (Bonfils et al., 2012).

Die-off events

Another global forest perturbation consists of tree morta lity rise (o r of forest die-off/dieback), caused by increasingly dry or hot conditions , as the Earth's climate continues to warm (Anderegg et al., 2012). Exc luding other factors that result in tree mortality, such as fires and insect outbreak, c limate-triggered forest mortality events are defined as a loss of at least

10% of dominant canopy trees (over an area of at least 250 km2), triggered directly by persistent drought and heat stress (Anderegg et al., 2012).

There are nu merous studies that reported in the past decades recent die-off events of forests in various continental regions, such as in North Ame rica (Van Mantgem et a l., 2009; M ichaelian et al., 2011; Peng et al., 2011), South America (Phillips et al., 2009; Xu et a l., 2011; Hilker et al., 2014) and Europe (Land mann and Dreyer, 2006; Lindner et al., 2010; Prăvălie et al., 2014a; Prăvălie et al., 2014b). In these instances, as well as in many other reports, it was found that the common cause of la rge-scale tree morta lity was ma inly connected to high temperatures and/or water stress (Allen et a l., 2010, 2015; Bennett et al., 2015; Seid l et a l., 2017). For instance, water stress subjects global forest ecosystems to an immense pressure, considering that drylands, crit ical terrestria l systems of the Globe due to lo w water ava ilab ility (and a lso the Earth's largest bio me) (Sch ime l, 2010), currently cover a significant share of the Earth's terrestria l area (~45% or ~67 mil km2) (Prăvălie, 2016).

Considering the well-documented tree morta lity in No rth America, a continent that has vast dryland areas (Prăvălie , 2016), it is estimated that the increasing water deficit and thermal stress on the trees recorded in the past decades have devastated ~20 mil ha of forests from Alaska to Me xico (a lso including the indirect impact of e xtensive insect outbreaks), especially in the continent's western region (Allen et al., 2010). Representative exa mp les of affected species include Pinus contorta in Brit ish Co lu mbia (>10 mil ha in this western area of Canada), Populus tremuloides in Saskatchewan and Alberta (Canada, ~1 mil ha) and Pinus edulis in the southwestern U.S. (>1 mil ha) (Allen et a l., 2010). Relat ively recent studies show that, in the past decades, tree morta lity rates increased by 4.9% yr–1 in Western Canada (Peng et al., 2011), and by 4.2% yr–1 in North-western US (Van Mantgem et al., 2009). It was also signalled that direct regional warming and the resulting water stress were most likely the drivers of widespread increases in boreal forest mortality in North America.

It is however interesting that tree dieback/mortality is a forest perturbation that occurs almost all over the world, even in environments that are not a ffected by a rid ity or that typically are not subjected to humidity stress. The A mazon region is a re markable e xa mp le in this respect. Ext re me droug hts in 2005 and 2010, classified as one-in-100 years events, affected 37% (~1.9 mil km2) and 57% (~3 mil km2) of the region (which totals 5.3 mil km2 ), considering standardized anomalies of dry-season rainfa ll fro m long-term mean (Le wis et al., 2011). The two major droughts, associated with high North Atlantic sea surface te mperatures (not with El Niño events, as is often the case of the Ama zon region) (Ph illips et al., 2009; Lewis et al., 2011), caused a massive greenness decline over the Amazon ra inforests, especially in the identified drought epicentres – one in 2005 (in southwestern Amazonia ) and at least three in 2010 (in southwestern Amazonia , Mato Grosso state in Bra zil, and in north-central Bo livia) (Le wis et al., 2011; Xu et al., 2011). Moreover, at least one ma jor drought epicenter was recently identified in 2015 (this t ime fo llo wing the e xtre me conditions of El Niño in 2015– 2016), in the region's northeastern sector (Jiménez-Muñoz et al., 2016).

In the first decade of the 21st century (2000–2012), other studies signalled that the large-scale decrease in precipitation recorded after the year 2000, during El Niño events, determined significant perturbations in the state of equatorial forests due to a notable vegetation greenness decline across more than 1.5 million km2 (~30% ) of the A mazon basin, compared with average conditions (Hilker et a l., 2014). Co mple mentary with decreasing atmospheric humid ity, this greenness decline is associated to an increase in the stress higher temperatures generate, considering that the Ama zon forest e xperienced in the past decades a cloud cover decrease and, therefore, longer sunny periods and greater canopy heating (Doughty and Goulden, 2008). A lthough the longer e xposure of forests to solar radiat ion can also be associated to higher ecosystem productivity in t ropical regions (Ne man i et a l., 2003), it seems that there is an increasingly high risk of losing

large a mounts of carbon in the A ma zon fo rest (wh ich is estimated to hold >100 Pg C in aboveground biomass ), against the background of climate warming and collateral die-off effects (Doughty and Goulden, 2008; Hilker et al., 2014).

The most important positive feedback of forest dieback for g lobal warming is connected to the release of carbon into the atmosphere (Figs. 5, 6). There are however other notable indirect pathways through which the carbon cycle can be disrupted due to forest canopy cover loss, such as more intense solar radiat ion, soil mo isture and infiltration decrease, or higher surface runoff (Anderegg et al., 2012). Regarding carbon loss fro m forest dieback, the A ma zon fo rest is once more a notable exa mp le, considering that, in this region, the 2005 and 2010 d roughts alone generated carbon net losses of 1.6 Pg C, and 2.2 Pg C, respectively (Ph illips et al., 2009; Lewis et al., 2011). These losses account for 16% and 22% of g lobal annual carbon emissions – currently estimated at ~10 Pg C (Le Quéré et al., 2016), wh ich therefore proves that as few as two such extre me c limatic events occurring in a decade may greatly offset the net gains of ~0.4 Pg C y r−1 in A mazon forests in non-drought years (Lewis et a l., 2011). In this conte xt, there is a h igh risk that, in this century, the region will become a major "tipping element" for global warming (Lenton et al., 2008).

Although in the past decade there have been other global cases of large-scale droughts impact on net primary productivity and carbon storage decline in forest ecosystems (representative exa mp les consist of the droughts in 2000 in North A merica and China, 2002 in North A merica and Australia, 2003 in Europe, and 2007 in Australia ), dieback events in tropical forests are perhaps the greatest problem in terms of terrestrial ca rbon sink degradation (Zhao and Running, 2010). However, there a re optimistic scenarios on this issue. Although tropical forests could presently be close to a high therma l threshold, above which carbon assimilation can reach a rapid decline (Doughty and Goulden, 2008), certain studies suggested that increased photosynthesis (in the context of the e xpected increase in at mospheric CO2) in the ne xt decades should more than offset the decrease of photosynthetic activity generated by therma l stress and forest die-off events (Lloyd and Farquhar, 2008).

Insect infestations

Even though forests are currently being affected by strong biotic stressors, such as pathogens (viruses, bacteria, oomycetes or fungi), insect pests are probably the greatest biological threat to fo rest ecosystem health, as they cause the most co mple x and d irect da mages to trees (Boyd et a l., 2013). While in the case of managed forests (tree fa rms ), insect pests are a consequence of globalization (e .g. via t imber trade), with significant negative effects especially in tropical and subtropical regions (Wingfield et al., 2015), for natural ecosystems this biological perturbation is ma inly due to climate warming (Bentz et al., 2010; Jacte l et a l., 2012), and is representative especially for borea l forests and, to a lesser extent, for temperate forests. This is due to the fact that high latitudes were a ffected in the past decades by a climate warming that wa s twice as h igh as the global average (IPCC, 2007, 2013), which e xtended insect life cycles and, imp lic itly, a mp lified tree damage in the growing season (Tru mbore et a l., 2015). This increase in insect outbreaks , ma inly caused by temperature increase during winter (pest insects are controlled especially by low winter temperatures , which kill eggs) (IUFRO, 2008), can even determine a higher sensitivity of forests to pathogen attacks (Dwyer et al., 2000).

There are various cases of devastating insect infestations in temperate and boreal natural forests. Representative e xa mples include the mountain pine beetle (Dendroctonus ponderosae) – native to western North America (Kurz et al., 2008a), spruce beetle (Dendroctonus rufipennis) – native to northern No rth A merica (Hansen et al., 2016), Asian longhorn beetle (Anoplophora glabripennis) – native species to China and Korea that spread to the US, Canada and in several Western and Central European countries (Hu et a l., 2009), the gypsy moth (Ly mantria dispar) – native Eurasian species

that spread to North Ame rica (Logan et a l., 2003), and the Siberian silk moth (Dendrolimus superans sibiricus Tschetw.) – native to Northern Asia (Kharuka et a l., 2007). Considering the la rge-scale in festations of the past decades , the direct connection between these major infestations and climate wa rming, as well as the special attention the species was given in numerous specialized studies, the mountain pine beetle is of para mount impo rtance for forest disturbance assessment, being probably the most intensely studied insect species in this respect (Bro wn et a l., 2010; Collins et al., 2011; Edburg et a l., 2011; Schoennagel et al., 2012; Landry et al., 2016).

The mountain pine beetle (present from No rthern Mexico to British Co lu mbia in Canada) is a special kind of threat as, unlike defoliating insects (e.g. gypsy moth), which generally da mage trees by reducing growth, this species kills trees (healthy and unstressed by other factors ) by feeding on phloem t issue and therefore interrupting the tree's nutrient supply (Brown et a l., 2010; Hicke et a l., 2012). It was suggested that insect outbreaks, coupled with wildfires, caused after 1980 the highest forest morta lity fro m documented records of the last century (Willia ms et al., 2010). The ma in causes are closely connected to more intense droughts and higher temperatures, which a llo wed insect outbreaks to occur in habitats where te mperatures were in itia lly too lo w for the survival o f the beetles (Clow et al., 2011). Ho wever, even though this change in insect outbreaks in the past decades was also associated with the intensification of other perturbations, such as fires, new findings suggest that forest withering following pine beetle infes tations does not necessarily mean higher fire severity in forests in Western North America (Carswell, 2014).

It is estimated that bark beetles have affected 17 million ha of forest in Western US regions since 1996, and more than 5 million hectares are attributed to the mountain p ine beetle (Carswe ll, 2014). The most heavily affected reg ions were the Rocky Mountains and Colorado Plateau, and the key affected forest species were the lodgepole pine (Pinus contorta), yellow pine (Pinus ponderosa) and piñon pine (Pinus edulis) (Boyd et al., 2013). The state of affa irs is even more serious in western Canadian forests (especially British Co lu mbia) (Fig. 4), where it is estimated that Dendroctonus ponderosae severely affected more than 10 million ha of forest (main ly lodgepole pine), between 2002 and 2006 a lone (Kurz et a l., 2008b). In the following years, an intensification of the consequences generated by these large-scale outbreaks is expected

– between 2000 and 2020, over 37 million ha of forest will be a ffected in British Colu mb ia (Fig. 6), with disastrous effects on regional ecosystem services (Boyd et a l., 2013). In the 21st century, it is high ly like ly that other insect outbreaks intensify in various regions in North America and Eurasia, considering that temperature increases, and implicit ly the intensification of climat ic a rid ity (Huang et al., 2016), are e xpected to occur especially in the borea l b io me (Gauthier et a l., 2015).

Major insect outbreak feedbacks for climate wa rming are contrary, i.e. a positive effect o f carbon e missions and a negative one of albedo increase. Considering yet again the exa mple of the we ll -documented North Ame rican forests, models showed a cumulative impact of carbon releases (ma inly due to the mountain pine beetles) equivalent to 270 Tg (teragra ms) carbon (o r 0.27 Pg C) during 2000–2020 in British Colu mbia (Kurz et al., 2008a ), 580 Tg C during 1999– 2050 in the same Canadian region (Arora et al., 2016), or 35 Tg C fro m 2000 to 2009 in the western United States (Ghimire et al., 2015). These ecosystem carbon storage decreases are substantial considering, for instance, that in the most severe beetle outbreak years in Brit ish Colu mbia, e missions reached ~75% of average emissions resulting from a ll Canadian forest fires between 1959 and 1999 (Kurz et a l., 2008a). As a result, during and post-outbreak state of the vegetation, these forest ecosystems (mostly temperate coniferous forests) turn fro m s mall net carbon sinks into significant net carbon sources (Kurz et al., 2008a). At the same time, although it was suggested that albedo-induced cooling (through forest mortality) is a

negative feedback for climate wa rming, it seems the net climate impact resulting fro m the contrary effects of carbon emissions (warming)/surface albedo (cooling) is warming in this vast western region of Canada (Landry et al., 2016).

Fig. 4 Global instances of occurrence for the 12 forest perturbations and their implications for the Ea rth's climate system dynamics. Note: biogeochem. – b iogeochemica l; the superscript numbers are citations that support the information presented in the concrete e xa mp les in the bo xes (1 – Jeong et al., 2011; 2 – Mac Donald et a l., 2008; 3 – Le wis et a l., 2011; 4 – Kurz et al., 2008b; 5 – Hansen et al., 2013; 6 – Haddad et al., 2015; 7 – Sitch et a l., 2007; 8 – Maisels et a l., 2013; 9 – Page et al., 2002; 10 – Mc Intyre et al., 2015; 11 – Zhao and Running, 2010; 12 – Penuelas et al., 2013); a lthough in the case of some mixed perturbations (e.g. fires) the climat ic in fluence seems to be the main triggering factor, the anthropic factor also plays an important role in the occurrence of forest disturbances (in this instance, by increasing forest susceptibility to fires, as a result of certain persistent environmental changes like fo rest clearing and drainage, according to source 9 cited in the box); in the case of the implications for warming bo x section, only climate warming effects were mentioned, although in some cases there are also climate cooling effects, as mentioned throughout this paper.

Anthropic perturbations

Deforestation

Probably the most easily and direct ly noticeable forest ecosystem issue (and thus the most serious perceived forest perturbation) is deforestation, which is why this type of forest change is currently getting heightened attention from the public (Mac Dic ken, 2015). This disturbance is defined as forest clearance and subsequently conversion to another land use, which means the permanent loss of forest cover (FAO, 2013). Although the forests/non-forests conversion can also be the result of natural causes (e.g. fires or droughts), it is much more appropriate to consider this d isturbance to be anthropic, considering that human activity (e.g. c learance for the e xpansion of farm lands or of urban, industrial o r transport infrastructure) is by far the most important in forest ecosystem dynamics (FAO, 2010). Fo rest net cover change must be analysed both in terms of decreases/losses (deforestation and far less important natural disasters) and of forest area increases/gains from afforestation (planting of t rees on lands that were not previously occupied by forests) or natural forests expansion, e.g. in abandoned agricultural areas (FAO, 2010).

As deforestation is the main vector of net loss in forest ecosystems, this environmental issue has been given special attention in recent years, as attested by the numerous scientific studies conducted globally . These global analyses were generally based on processing national data from satellite-sourced databases provided by governments. One of the most e xtensive and credible analyses of global forest area dynamics was recently conducted by Keenan et al. (2015), based on national data e xt racted fro m the 2015 FA O Globa l Forest Resources Assessment (FAO, 2015). According to the study, the global forest area decreased by ~1.3 mil km2 (3%) in the past two and a half decades, from 41.3 mil km2 in 1990 to 40 mil km2 in 2015. It was found that these overall dynamics were due to the decrease in natural forest areas from 39.6 mil km2 (1990) to 37.2 mil km2 (2015), and to the e xpansion of planted forests (including rubber p lantations, but exclud ing oil palm or other agricultural p lantations ) from 1.7 mil km2 in 1990 to 2.8 mil km2 in 2015 (Keenan et al., 2015). If the change balance surprisingly had a slight surplus in the case of boreal (fro m ~12.20 mil km2 in 1990 to 12.24 mil km2 in 2015) and temperate (6.2 mil km2 in 1990, 6.8 mil km2 in 2015) forests, tropical and subtropical ecosystems e xperienced a massive decline (2 mil km2), fro m 22.9 mil km2 in 1990 to 20.9 mil km2 in 2015 (Keenan et al., 2015). The study's results confirmed that countries in the tropical forest region (e.g. Brazil, Bo livia, Nigeria, Tan zania, Indonesia, Myanmar or Ma laysia) held

the heaviest forest losses globally, as a result o f anthropic de forestation that aimed to e xpand agricu ltural lands, e.g. for oil palm plantations in Indonesia and Malaysia (MEA, 2005).

Regarding the results obtained by means of teledetection, a representative study showed that, between 2000 and 2012, global forests had losses of 2.3 mil km2,concomitantly with gains of 0.8 mil km2, i.e. in little over a decade there was a net decline o f 1.5 mil km2 (Hansen et al., 2013). The same study signalled that the tropics currently have the most apparent forest area decrease trend (Fig. 4), which is also the only statistically significant globally. The source states that, while the heaviest tropica l forest losses occurred in countries such as Bra zil, Bolivia , Nigeria, Angola, Tan zania, Indonesia and Malaysia, a ma jor decline is also noticeable in boreal regions in A laska, central-northern Canada and especially eastern Siberia. Results based on Earth observation satellite data showed that, interestingly, in Bra zil, although there still are e xtre me ly alarming los ses in forest cover due to deforestation (e.g. fro m over 40000 km2/yr in 2003– 2004 period to under 20000 km2/yr in 2010–2011), Indonesia seems to currently be new global deforestation leader, considering the fact it had the largest global increase in fo rest loss in this period (e.g. fro m under 10000 km2/yr in 2000–2003 to over 20000 km2/yr in 2011–2012) (Hansen et al., 2013). This a larming overall dyna mics (decrease) of t ropical forests was confirmed by other similar studies, based on satellite data analyses (Achard et al., 2014; Kim et al., 2015).

Deforestation has various environmental imp licat ions. For instance, extensive forest clea rance in tropica l Africa is the ma in cause for the intensificat ion of the water e rosion process, which is currently the ma in form of land degradation in the region, as well as in other parts of the continent (Prăvălie , 2016). The impact on biodiversity is another relevant e xa mple , especially for tropical forests, which a re the most heavily affected by the conversion of forests into agricultural lands (Lewis et al., 2015). It was however suggested that this ecological issue can also be noticed in vast managed boreal forests, where the explo itation of older forests caused the decline of valuable biodiversity elements, such as cavity shelters or coarse woody debris (Gauthier et al., 2015). In the next decades, large-scale deforestation is expected to continue especially in tropica l forests, considering the 6 b illion people (55% of g lobal population) who will live in the t ropics by 2100, and, implicitly, the increased demand for agricultural lands (Lewis et al., 2015).

The effects of global deforestation on the climatic system are connected to two mechanisms – carbon cycle and radiation balance (Figs. 5, 6). Fo r the former, it is estimated that deforestation is the second largest anthropic source of atmospheric CO2 (after fossil fue l burning), and contributes with up to one fifth of total e missions of anthropic activit ies (van der Werf et al., 2009). In the tropical region, where this process is the most intense and where it generates a net emission effect, unlike te mperate and boreal forests (net carbon accumulat ion), it is estimated that the total impact of deforestation is of 1.3 (±0.7) Pg C y r−1 (1990–2007), considering that gross deforestation emissions of 2.9 (±0.5) Pg C yr−1 are partially annulled by assimilation in forest regrowth, estimated at 1.6 (±0.5) Pg C yr−1 (Pan et al., 2011). Other studies confirmed ca rbon sources fro m tropica l deforestation at ~1 Pg C yr−1 between 2000– 2010 (Baccini et a l., 2012) or under this threshold between 2000 and 2005 (Ha rris et al., 2012). The sa me studies also note that Brazil and Indonesia, with forest carbon stocks of 53 and 19 Pg C (together they total 35% of tropica l fo rest carbon stocks), account for the most massive atmospheric CO2 e missions, through tropical land-use changes (Bacc ini et a l., 2012). While simulat ions show that the possible (yet unlikely ) co mplete future deforestation of tropical forest ecosystems could generate a mean global wa rming of almost 1 °C (Lewis et al., 2015), it is possible the value be largely underestimated, considering the forests' majo r role both in the carbon cycle and in stabilizing atmospheric temperatures through the evaporative cooling process.

Regarding the radiat ion balance, it is known that high-lat itude deforestation causes radiative cooling (Fig. 5) by increasing the surface albedo effect (Bala et a l., 2007; Davin and Noblet-Ducoudré, 2010; Lee et al., 2011; Lo ranty et al.,

2014). Paleoc limatic studies confirm this fact, as it was found that the loss of boreal forests in the past caused positive effects for ice ages, while their e xpansion 6000 years ago amp lified climate warming in the Ho locene (Bonan, 2008). Based on this reasoning, it can be concluded that promoting deforestation in the boreal zone is benefic ial for the climate system (Ba la et al., 2007; Anderson et al., 2011). This conclusion is however re lative, considering the decreasing importance of albedo, against the background of a warmer planet with a constantly shrinking snow cover (Bonan, 2008).

Fragmentation

Amid various types of land cover transformations, for the most part generated by human society, another fo rm of forest cover loss is frag mentation, which entails d ividing a forest in s ma lle r isolated sectors, and therefore enhancing its e xposure to anthropic influence along frag ment edges (Haddad et al., 2015). Th is e xposure accounts for the occurrence of edge effects, which are associated with various physical and biotic perturbations in the abrupt transition sectors that separate adjacent forest ecosystems (Laurance et al., 2007). While these perturbations can cause mu ltiple negative effects in the structure and functionality of frag mented forest ecosystems , forest fragmentation is a serious ecological threat also due to the fact that it interrupts the nutrient flow, organism dynamics and the connectivity of species between two or more ecosystems (Billings and Gaydess, 2008; Ibáñez et al., 2014; Auffret et al., 2015).

Some of the most important edge effects are the change in microc limate, shifts in composition and dynamics of biological co mmun ity, higher tree mortality, lower b io mass and changes in the biogeochemica l cycles (Laurance, 2002; Broadbent et al., 2008; Laurance et a l., 2011). Moreover, these perturbations can generate other negative effects, such as enhancing fires as a result of changing loca l c limatic conditions by intensifying incident solar radiation, increasing temperatures and reducing humidity (Dantasde Paula et al., 2016). In terms of spatial range, the most important transformations generated by edge effects generally occur up to 200 m into intact forest areas, but can frequently reach 1 km (Dantas de Paula et a l., 2016). In certain cases, the propagation of edge effects can e xtend as much as 5–10 km into natural forests (Broadbent et al., 2008). Te mpora lly, in these transition areas, certain changes, such as biomass content, can last at least 5 years, but can also reach one century (Dantas de Paula et al., 2015).

Re latively recent studies showed that all of the world's forests are generally undergoing a fragmentation process (temperate forests are the most heavily affected historically ) (Fig. 4), e xcept for the Ama zon and Congo River Basins, the largest forest regions on Earth that are still largely co mpact (Haddad et al., 2015). Even though extensive areas of tropical (and, in part, borea l) ecosystems are affected to a small e xtent by landscape fragmentation, the tropical bio me is however subjected to this global threat. It is estimated that only 24% of its forests rema in intact, while 46% a re already frag mented and 30% otherwise degraded (Le wis et al., 2015). Thus, even in the Bra zilian A ma zon it is estimated that frag mentation and deforestation result in up to 50000 km of new forest edges every year (Laurance et al., 2007). Given this context, negative implications are connected to edge effects, as well as to other ma jor impacts, e.g. disrupting metapopulation dynamics and species losses, especially in the smallest fragmented areas (Lewis et al., 2015).

One of the most disastrous effects of forest habitat frag mentation is biodiversity decline, as a result of the accelerated e xtinction of loca l species . Ho wever, this process depends on a variety of factors, such as forest frag ment size , surrounding habitat, characteristics of species (dispersion and reproduction capacity, diet, etc. ) and the interaction between them and the distance to anthropic disruption sources (Gibson et al., 2013; Estrada et al., 2017). For instance, considering tropical forest animal species, representative for the very high global biodiversity, it was found that in certain 1000 ha forest frag ments some ta xonomic groups (birds) in Eastern Africa can beco me 50% e xtinct in the first 50 years after isolation (Brooks et a l.,

1999), while in certain 100 ha (o r less) frag ments in Ama zonian forests half of the in itia l species can disappear in ~15 years (Ferra z et a l., 2003). Ho wever, the state of nume rous ma mma l groups (e.g. primates) is at least equally a larming throughout the tropical region, which, a mid long-term deforestation and forest fragmentation, has experienced dra matic changes, such as decreasing numbers, population restructuring or loss of genetic diversity (Estrada et a l., 2017). The decline of ma mma ls, especially herbivores , can have serious consequences by disrupting the functionality of tropica l forests, which will be addressed at length in the defaunation section.

A forest fragmentation increase is expected to occur worldwide in the coming decades . Considering the expected development of linear infrastructure, a ma jor anthropic vector of frag mentation alongside agricultura l e xpansion (Laurance et al., 2009; Sloan and Sayer, 2015), it is estimated that at least 25 million km of new roads will be built by 2050 globally, most of which will occur in the tropica l region (Laurance et al., 2014). Moreover, it is like ly that this massive infrastructure e xpansion (equivalent to ~600 Equator circu mferences ) will generally be chaotic or poorly planned in the next decades (Laurance et al., 2014). Considering the subsequent collateral e ffects such as possible secondary and tertiary roads, the establishment of new human settle ments and overexploitation of local resources , the new roadways will inevitably be a ma jor driver of forest frag mentation and environmental degradation . This will prima rily endanger tropica l environ ments, as they are characterized by e xceptionally h igh biodiversity and mult iple vital ecosystem services (Laurance et a l., 2014; Lewis et al., 2015).

While forest frag mentation can be considered less important in te rms of acce lerat ing climate change due to their lo w spatial range, recent studies conducted mainly on the tropical bio me showed that this perturba tion can in fact a mplify g lobal warming. Models showed that frag mented tropical bio mass emits up to 0.2 Pg C yr−1 (Püt z et al., 2014), as it is less developed in transition sectors (between forests and converted lands ) – on average, 25% lo wer in the first 500 m of forest edge, compared to forest internal a reas (Chaplin-Kra mer et al., 2015). As a result, significant carbon losses and degradation of ecosystem services occur in tropical forests (Numata et al., 2011; Dantasde Paula et al., 2015), but this issue is also noticeable in te mperate and borea l forests (Figs. 5, 6), as aboveground biomass is generally lowe r in all fragmented forests worldwide.

Air pollution

Another issue affecting forest health is air pollution, wh ich can generate significant pressure onto these terrestrial ecosystems due to the presence of various anthropic pollutants, the most important of which – and probably the most wide ly studied – is tropospheric ozone (O3) (representative for large scale vegetation injuries) (Ashmore, 2005; Bytnerowic z et a l., 2007; W ittig et al., 2009). Ho wever, nitrogen and sulphur are two other representative global pollutants that, through e xcessive acid deposition (NOx and SO2), have generated ma jor soil ac idificat ion and tree productivity decline in many te mperate and boreal forests (Lovett et al., 2009; Lu et a l., 2014) . At the same t ime, it appears terrestria l acidification is also a notable threat for the health of tropical forests (Azevedo et al., 2013; Lu et a l., 2014). A lthough sulphur deposition declined in the past decades in nu merous regions (as a result of SO2 e missions decreases), e.g. in Europe, eastern North A merica and China, the global increase of N deposition enhanced the pressure on terrestrial ecosystems. It is estimated that 11% of the worldwide natural vegetation is exposed to N deposition in e xcess of 10 kg N ha yr−1 critical leve l (Pa rdo et al., 2011; Lu et al., 2014). Thus, in many forest areas, the benefic ial effect of low rates of N deposition (wh ich is known to stimulate forest productivity) was annulled, as will be presented at length in the biogeochemical change section. Other exa mp les include mercury (Lovett et al., 2009) and certain dangerous radionuclides

(like rad iocesium) trans ferred to forest bio mass during nuclear tests (Prăvălie, 2014) or the ma jor nuclea r acc idents in Chernobyl and Fukushima (Thiry et a l., 2009; Koarashi et a l., 2016; Tera mage et al., 2016; Prăvă lie and Bandoc, 2018). However, the harmful e ffects are main ly regional and affect people and anima ls (through food webs and other pathways ), rather than forest ecosystems .

As previously mentioned, O3 is the most phytotoxic a ir pollutant (as well as an important greenhouse gas), which is why it has been carefully studied especially in No rth American and European forests (Holmes, 2014). Product of photochemica l reactions between nitrogen oxides (NOx) and volatile organic co mpounds , it can significantly decrease ecosystem productivity by reducing stomatal conductance, decoupling of carbon dioxide and water e xchange, or dimin ishing photosynthesis and biomass (Wittig et a l., 2007, 2009; Matyssek et al., 2010a; A insworth et al., 2012; Lo mbardozzi et a l., 2012, 2015). Ozone can therefore generate serious imp lications on forest health, by da maging stomatal functions, disrupting the carbon and water balance in forests and, therefore, by disrupting the gas e xchange between the atmosphere and biosphere (Matyssek et al., 2010b; Fares et al., 2013; Hoshika et al., 2015).

The intensity of forest perturbations depends however on the pollut ant's concentrations, which show large tempora l (seasons of the year) and spatial (d istance to anthropogenic precursor emissions) heterogeneity (Wittig et al., 2009). Historica lly, te mperate forests have been the most heavily affected by ozone stress, considering that the northern hemisphere has had the most massive e xposure to this type of pollution. These ecosystems are currently e xperiencing increases in tropospheric O3 concentrations that are 60–100% over industrial revolution levels (Hoshika et al., 2015). Vast forest areas in the northern (main ly the temperate forest bio me) and even southern (tropical bio me, ma inly in the A mazon region) he mispheres are currently affected by high levels of O3 concentrations (Figs. 4, 6), which, during summe r (when genesis conditions are favourable) e xceed the 40 ppb (parts per b illion) threshold, dee med suffic iently h igh to cause visible leaf in jury and cellula r da mage in trees (Sitch et al., 2007). However, mode lled ozone concentrations show that some forested lands in eastern US and China, central southeastern Europe, and Amazon and Congo Basins , are already being exposed to alarming concentrations that exceed 60 and even 75 ppb during summer (Sitch et al., 2007).

In the coming decades , a global increase in O3 concentrations is e xpected, under the assumption of the most pessimistic greenhouse gas emissions scenario (RCP8.5) (Young et a l., 2013). Spatially, certain models (based on previous pessimistic SRES A2 e missions scenario) fo resee increases of over 90 ppb by 2100 (during summer) (Sitch et al., 2007), including in e xtensive sectors of temperate (te mperate broadleaf and mixed forests in eastern US and Ch ina) and even tropical (t ropical and subtropical moist broadleaf forests in southern Amazonia, Congo, southeastern China and, partia lly, in northern Ind ia, where it seems even certa in tropica l and subtropical dry broadlea f forests will be a ffected) forests. Given this context, it is like ly that the risk of ecosystem service degradation will rise , considering that forest photosynthetic rates are expected to slow down in these regions that will be heavily affected by air quality changes.

Consequently, carbon sequestration decrease, coupled with b io mass losses, is one of the most like ly negative e ffects of higher concentrations of O3 in the at mosphere (Fig. 5). It was found that the mean concentration of tropospheric O3 at temperate latitudes of the northern he misphere (estimat ed at 40 ppb during summe r daytime ) is a lready reducing tree biomass by 7% in certain parts of the region (Wittig et al., 2009). Also, e xpe rimental studies showed this value could reach 17% by the end of the century, at least in areas that are heavily polluted with O3 (with concentrations close to the 100 ppb threshold) in the te mperate zone (Wittig et a l., 2009). Although these results (as well as others) can confirm the imp lic it presence of a warming e ffect in the climat ic system, due to decreased photosynthesis as a result of bio mass loss, recent studies show that the presence of forests in a world that is heavily polluted with tropospheric O3 does not necessarily imply

reduced productivity, if considering tree species that are tolerant to ozone and the competitive interactions between them (Wang et al., 2016). While it is difficult to foresee the exact response of forest ecosystems to higher O3 concentrations, considering the interference of various other factors (Ainsworth et al., 2012), a future possible acceleration of global warming can be e xpected, g iven the perturbations linked to decreases in photosynthesis , tree growth, bio mass accumulation and land carbon sink.

Fig. 5. Forest perturbations and exa mp les of imp licat ions (biogeophysical mechanisms) on c limate system dynamics (warming/cooling), determined by the three global forest bio mes. Note: NPP – net prima ry productivity; strong/moderate/weak feature of CSI/SAI/ CSD/ ECD/ SAD we re estimated based on the climatic control mechanis ms generated by forest biomes, presented in figure 2 (for instance, it was considered that, in the event of forest perturbations triggered in tropica l forests, CSD and ECD a re strong, similar ly to the magn itude of the carbon storage and evaporative cooling processes that characterize these ecosystems; another e xa mple – in the case of boreal forest perturbations that involve partia l or total tree mo rtality, it was estimated that SAI is strong, considering the high importance of snow in reflecting solar radiation at h igh latitudes); CSI and CSD processes were only considered for forest bio mass, without taking carbon dynamics in soil, peat and underlying permafrost deposits into account, triggered directly or ind irectly by forest perturbations; all imp licat ions for climate dyna mics are assessed based on the current forest perturbation context (as a result of the uncertainty with regard to concrete interactions between C, N and P in a ll g lobal fores ts, the implications of biogeochemical shifts in c limate dynamics were estimated only in terms of the intensification of the carbon cycle at high latitudes, which triggered CSI/ SAD through bio mass development, under the assumption of N availability for the stoichiometric demands of the forest vegetation).

Mixed (climatic and anthropic) perturbations

Defaunation

Indeed, life can be considered the most unique characteristic of Ea rth, while d iversity can be described as the most important feature of life (Card inale et al., 2012). Unfortunately, diversity is fac ing a massive threat in the Anthropocene age, as there are signs that say the sixth mass e xtinction could possibly become co mparable in the near future to the five previous global mass extinction events (Dirzo et al., 2014). Biodiversity loss is already alarming, considering that since 1500 over 900 species of vertebrates, non-vertebrates and plants have disappeared (Mooney et al., 2009). Of these, ~320 terrestrial vertebrate species have become e xt inct in the past 500 years (Dirzo et al., 2014). The most vulnerable vertebrate groups are birds (150 species extinct since 1500) and ma mmals (~80 species extinct in the past five centuries ) (Mooney et al., 2009). While in the past centuries e xtinctions were ma in ly caused by direct hu man p ressure (like land use change fro m forested lands to non-forest areas), the current extinction rate (>100 species per million species per year, up to 1000 times higher than the natural rate) is due to the synergistic impact of anthropic activ ities (which a re still probably the most important) and climate change (Rockström et al., 2009). Moreover, if the wide range of current global changes rema ins unchanged in the future, projections show that the planet will in fact undergo the sixth mass extinction in the ne xt 240 years (Barnosky et al., 2011; Hooper et al., 2012), which is extremely fast, even on a human time scale.

Defaunation or the loss of wildlife animal species (Dirzo et al., 2014) is a problem that is typical for forest biomes, more specifica lly of tropical ones, considering they are the major global terrestrial b iodiversity hotspot (Gia met al., 2012; Le wis et a l., 2015). The issue is in fact prehistoric (Corlett, 2013), as it is estimated that the init ial impact of hu man

colonizat ion of tropica l ecosystems (wh ich started tens of thousands of years ago) led to megafauna losses of 18% in Africa and 83% in South America (Le wis et al., 2015). Alongside climate change, anthropic pressures continue to this day via various pathways like fo rests fragmentation, habitat destruction or overhunting (Kurten, 2013; Poulsen et al., 2013), favouring the so-called "empty forest syndrome" (Redford, 1992). Thus, the highest global pattern defaunation intensity is currently found in tropica l forests of Latin A merica, Africa and Asia, which a re be ing affected by a rapid dec line of vertebrates, especially herbivore mammals (Kurten, 2013; Dirzo et al., 2014).

The loss of herbivore ma mma ls could account for the profound negative effects in these environments, considering their mu ltid imensional ecologica l importance for the tropical bio me . La rge-bodied herbivores (≥100 kg) are essential for ma intaining healthy tropical forests , as they disperse seeds, which is a vital ecolog ical process in plant germination , vegetation density growth and in the colonization of new territories by trees (Kurten, 2013; Vida l et al., 2013). Therefore , large ve rtebrates are an essential vector in the dyna mics o f forest structure and composition (Harrison et a l., 2013; Poulsen et al., 2013). These anima ls are also e xt re me ly important for the functioning of other p rocesses such as nutrient cycling, fires and carbon storage (Corlett, 2013; Ripple et al., 2015). For instance, herbivores can accelerate the nutrient circuit in forest ecosystems through plant consumption and subsequent defecation, thus returning nutrients into soil faster than normal decomposition processes (Pastor et al., 2006).

A special ro le in tropical ecosystem health is played by megaherbivores (≥1000 kg) in Africa and Asia, as they are highly effective seed dispersers, which strongly influence vegetation distribution and abundance. A special category in th is respect consists of elephants, which are capable of ingesting large seeds (≥1 c m) and dispersing them through defecation at great distances fro m their sources (Ca mpos-Arceiz and Blake , 2011). Given this conte xt, the case of African forest elephants (Loxodonta africana cyclotis) is particularly relevant as, in Congo alone, they disperse ~345 large seeds per km2 every day (from 96 species), generally more than 1 km a way fro m parent trees (Bla ke et al., 2009). However, the past decade's hunting pressure has considerably diminished this ecological attribute of elephants in Africa and, part ially, in South-eastern Asia. For instance, the number o f African forest elephants has decreased dra matically lately due to poaching,

i.e. by 62% between 2002 and 2011 a lone (Maisels et al., 2013). There a re however other endangered megafauna species (e.g. wild Africa 's rhinoceroses, which could become e xt inct in the next 20 years) or other smaller species of tapirs (which live in Central and South A merica, and in South-eastern Asia) or gorillas (Africa ), wh ich a lso generally play the role of forest gardeners, as they are highly effective seed dispersion agents (Ripple et al., 2015).

Therefore, a ma jor negative consequence of this e xtensive reduction in herbivore species lies in the decrease in biomass in the tropical forests of Africa and even Latin America (regions in which biotica lly-d ispersed tree species are dominant) (Figs. 4, 6), which is however less apparent in Asian forests, where the wind-dispersed mechanism is dominant (Lewis et al., 2015). The widespread decline of large herbivores can have multip le negative consequences for forest ecosystem functionality, especially since ~60% of the large herbivores of the world are currently endangered, and most of these vulnerable species are found in developing countries (Ripple et a l., 2015). These states are characterized by a critica l underfunding of conservation, which ma kes it impossible to control the key e le ments threatening this fauna and, imp licit ly , overall forest health.

While the current defaunation process is largely attributed to anthropic pressures, climate chang es will probably contribute over the next decades to an equal or even greater e xtent to the defaunation of tropica l forests and more (Dawson et al., 2011; Urban, 2015). In the near future, megaherbivores in part icular will likely become even more vulnerable , considering that climate changes will affect the actual few re ma ining megafauna, i.e. only 8 terrestrial species , compared to

~42 in late Ple istocene (Ripple et al., 2015). In fact, climate change has already affected fauna species in general, in numerous forest areas worldwide, but for the time be ing as a subtler form of e xt inction, i.e . the e xtinction of ce rtain ecological interactions (Estes et al., 2011).

Although defaunation is a silent threat co mpared to other forest perturbations (e.g. deforestation), it was found that it has the potential to accelerate climate warming by eroding the tropical forests' carbon storage capacity (Fig. 5). Recent studies showed that, in the Amazonian and Atlantic forests , the loss of large vertebrates has considerable effects in aboveground biomass decrease (and, imp licit ly, in carbon losses), considering the direct connection between large seeds dispersed by large-bodied anima ls and high wood density (Be llo et al., 2015). This reasoning was confirmed by other studies that analysed the relation between the defaunation process and carbon stock decrease in tropical forests (Osuri et a l., 2016; Peres et al., 2016). Moreover, simu lations show that tropical forests in Africa and Latin A me rica can e xpe rience carbon losses of up to 12% under the 100% e xtirpation scenario of large-seeded anima l-d ispersed species (Osuri et al., 2016). Southeastern Asian forests will however be less heavily affected, considering the impo rtance of wind as a do minant vector for seed dispersion and biomass control (Osuri et al., 2016).

Fires

It is known that fires are natural processes that contribute to the norma l functioning of terrestria l ecosystems via vegetation control mechanis ms such as biomass production, species succession, ecological heterogeneity, nutrient cycling and soil ecology (IUFRO, 2008; Bowman et al., 2009; Morris et al., 2015). For instance, considering bio mass control, it was suggested that forests would cover an area at least twice as la rge in the absence of fires, especially in high ly fla mmab le savannas (the most frequently burnt ecosystems on Earth), e.g. the ones in Africa and South A merica (Bond et al., 2005; Bowman et a l., 2009). However, large-scale tree morta lity due to increasing fire activity, against the background of climate change and anthropic influence, is currently an important forest disturbance in numerous regions worldwide.

Forest fires are an important consequence of mu ltiple drivers, such as dry environmental (at mospheric) conditions, available fuel (vegetation characteristics) and ignition patterns (main ly people) (Moritz et a l., 2005; Flannigan et a l., 2013), which means their occurrence depends on the interaction of these triggering factors. Dry atmospheric conditions constitute the largest driver of fires/wildfires, la rgely p resent in dry sub-humid and se mi-arid c limates worldwide, where there is a high humidity deficit, as we ll as suffic ient vegetation that can fuel fires . Thus, diverse vegetation (forests, shrubs and grasslands) fire events occur in these dryland systems in southwestern North America (US/Me xico), eastern South America (Bra zil), southern Europe (Mediterranean region), eastern Africa (Tan zania, Kenya or So ma lia) and eastern Australia (Prăvă lie, 2016). It was found that three of these regions , southwestern North America, eastern Bra zil and eastern Africa, recorded after 1979 a significant increase in the fire weather season length (number of days in a year during which fire danger is above half its value range), against the background of rising ma ximu m te mperatures, wind speed and rain -free days, and decreasing atmospheric humidity (Jo lly et a l., 2015). The sa me study signalled that between 1979 and 2013 the fire weather season length expanded (statistically significant) in total across 25% (~30 mil km2) of the global vegetated area, generally on all continents (Jolly et al., 2015).

Even though fires are not usually associated with humid areas, where it is considered that dry and flammable plant materia l are insufficient, in the past decades it was found that two of the most devastating global fires occurred in tropical rainforests, e.g. in southeastern Asia (Indonesia) in 1997– 1998 (Page et al., 2002) and A ma zonia in 2005 (Aragão et a l., 2008; Zeng et al., 2008). In these instances, while short-term dry weather conditions were the ma in cause in triggering these

environmental disturbances , the anthropic influence (in terms of ign ition source) was re ma rkab le in pro mot ing fires, due to woody debris and changes in the mic roclimate as a result of deforestation and forest frag mentation (Lewis et a l., 2015). These events' impact on tropical fo rest health is significant, and it is estimated, for instance, that fires in A ma zonia account for 12 to 30% decrease in aboveground live bio mass, 23 to 31% decrease in canopy cover, and 226 to 462% increase in tree morta lity (Brando et al., 2014). Economica lly, it is estimated, for instance, that the fires in Southeast Asia's tropical forests in 1997– 1998 a lone generated losses estimated at ~9 billions of dollars (Bowman et al., 2009). Under conditions of high- intensity, widespread and recurrent fires , in the near future a h igh risk of preventing forests from developing to maturity and invasion/maintenance of fire -resistant species like savanna (or shrubby) vegetation is e xpected, wh ich would imply the risk of a large-scale "savannization" of tropical forests (Lewis et al., 2015).

For the c limatic system, the threat primarily consists of carbon e missions. It is estimated that the fires in 1997– 1998 alone, during the El Niño-induced drought, devastated ~20 million ha of tropical forest, releasing net forest fire e missions equivalent to mo re than 40% of the global fossil fuel e missions recorded during this period (Cochrane, 2003). It appears that the fires in Indonesia alone released carbon amounts equivalent to 13–40% (0.8– 2.6 Pg C) of global fossil fue l emissions (Fig. 6), a lthough they only affected 1.4% of the global vegetated land area (Page et al., 2002). The cause of these massive emissions is connected to above-ground biomass and especially to the underlying peat fires (Fig. 4), which have become in the past decades more susceptible to burning due to anthropic interventions such as forest clearing and drainage (Page et al., 2002; Bowman et a l., 2011). Thus, given the context of fires, tropical peatlands are an important vector in carbon emissions, as it is estimated they hold ~50 Pg C (over 8% of ~600 Pg C stored in global peatlands), prima rily in Southeast Asia, where Indonesia holds by far the largest share of these terrestrial systems, especially in Kalimantan and Sumatra islands (Yu et al., 2010).

Carbon releases in the tropical region are a lso representative in the case of Amazon forest fires . They have become increasingly severe, against the background of ma jor droughts recorded in 1997, 1998, 2005, 2007, 2010 and 2015, coupled with the e xpansion of anthropic ignition sources such as human settlements, logging, forest fragmentation, and the spread of agricultura l areas, which resulted in more fire leaks into fla mmable forests (Aragão et al., 2007; Malh i et al., 2009; Brando et al., 2014). Considering the impact forest dieback and fire e missions during the 2005 and 2010 droughts alone, they are estimated to have reached 16% and 22% of current g lobal ca rbon e missions (Phillips et a l., 2009; Lewis et a l., 2011), as previously mentioned.

While ca rbon e missions can also be notable in the case of te mperate forest fires , e .g. in 2010 in western Russia, and in 2012 in the Un ited States and Spain (Jo lly et a l., 2015), the presence of these perturbations at higher latitudes, in the boreal region, is much mo re alarming. Th is is due to interconnected causes such as accelerated climate wa rming (with fire amp lification potential) – double at northern latitudes compared to the mean global state (IPCC, 2013), the presence of large a mounts of litter in these ecosystems – resistant to decomposition and highly fla mmable (Randerson et al., 2006) and the existence of massive carbon deposits (which can be released directly or indirect ly as a result of forest biomass combustion) in permafrost, peatlands and soils, estimated to total ~32% of global terrestrial ca rbon (Gauthieret al., 2015). Of these, total organic carbon deposits stored in northern permafrost areas (in Canada, Alaska, Rus sia and, significantly less, in Fenoscandia) a re by far the most important – a lmost 1700 Pg C, appro ximately double co mpared to the a mount currently present in the at mosphere (Schuur et al., 2015). These deposits are a majo r threat for climate wa rming in the near future, considering that projections show an increase in fire occurrence, area, and severity in special in Siberian Russia (Gauthier et al., 2015), wh ich holds most of these deposits (Schuur et al., 2015). In fact, considering the massive carbon

deposits in these frozen Russian soils (some with very high ca rbon content, e.g. in Yedo ma, located in northeastern Siberia that reaches Alaska and holds a total of ~450–500 Pg C) (Zimov et a l., 2006; Lenton, 2012; Strauss et al., 2017), it was suggested that carbon releases from perma frost thawing in this country alone could be, by the end of the 21st century, several times greater than current emissions of tropical deforestation (Gauthier et al., 2015).

Fires can influence at mospheric temperature through other biogeophysical mechanisms as well. For instance, the emission of black carbon aerosols by fires and their accumulation on ice and snow in the boreal region can reduce albedo, thus increasing air te mperature (Randerson et al., 2006). On the other hand, postfire changes in surface albedo, and subsequent albedo rises in the fo llo wing years as a result of an increased snow e xposure amid forest canopy loss after fires, can annul, to a significant e xtent, the wa rming effect generated by the initia l albedo decrease (Randerson et al., 2006; Chen et al., 2018). Other e xa mp les of mechanisms that influence a ir te mperature, in addition to d irect fire-e mitted greenhouse CO2, include changing solar radiation in multip le ways in the tropical region, due to fire -induced aerosols (Doughty et al., 2010), or indirect carbon emissions as a result of high soil erosion after fires, in the temperate region (Caon et al., 2014).

Composition shifts

The individual or synergistic action of certa in disrupting agents, such as climat ic changes, fires, insect pests, logging or gra zing, led to significant t ransformat ions in global forest ecosystems, including in terms of co mmunity co mposition. Losses and gains of tree species in the context of these disrupting factors have already affected forest co mposition in numerous regions worldwide (FAO, 2012; IPCC, 2014), alte ring their d iversity and, in many cases, making them mo re susceptible to the so-called "biotic homogenization" process (the increase in levels of similarity, ta xono mically and functionally) (Tabare lli et a l., 2004; Olden, 2006; Lôbo et a l., 2011; Ibarra and Martin, 2015; van der Plas et a l., 2016). However, in specialized studies it is generally difficult to identify a ll these changes on a planetary scale, in re lation to the various identified factors. Nonetheless, it is clear these perturbations have affected, in one way or another, the multip le ecosystemservices provided by these terrestrial ecological systems.

One of the most important factors of forest composition changes consists of anthropic deforestation, which over the past three centuries alone has eliminated ~50% of the planet's forest areas (temperate forests were the most heavily affected, generally due to the e xpansion of cultivated systems ) (M EA, 2005), thus completely annih ilat ing tree co mmunity composition in this instance. What is extre mely a larming is the fact that the global deforestation process is still ongoing to this day, at highly disturbing rates (Hansen et al., 2013; Keenan et a l., 2015). The process completely e liminates tree species in the case of arab le land e xpansion, or da mages forest species composition, structure and habitat, in the case of no net loss of forest areas. For instance, changing the composition of tropical forests by the widespread replacement of native tree species with e xotic tree species (e.g. Eucalyptus plantations that total ~20 million ha globally, main ly in t ropical and subtropical regions) (Wingfield et a l., 2015), p lanted for a higher t imber production, accounts for notable ecological perturbations, e.g. biodiversity loss (Keenan et al., 2015).

The human intentional or unintentional promotion of bio logical invasions (non-native pests, pathogens or plants) in numerous forests is also a global vector for changing the composition of forest vegetation. One of the most widely -known e xa mples of forest species that were replaced over e xtensive areas is the American chestnut (Castanea dentata), which e xperienced la rge-scale devastation in the Eastern United States starting with the first half of the 20th century, when a fungal pathogen (Cryphonectria parasitica) from Asia was introduced (Burke , 2011; Van Drunen et al., 2017). Another representative exa mple is the pinewood nematode, which has eliminated pine trees over the past century in large forest

areas in Eastern Asia, especially in Japan and, more recently, in China and South Korea (wood trade being a ma jor cause for accele rating this bio logical invasion) (Ma lle z et al., 2015). Bio logical invasions are still a major driver for changing tree composition around the world, especially in the temperate region, which is deemed the most vulnerable to various non – native species, generally as a result of high levels of general trade (Koch et al., 2011; Early et al., 2016).

The current reorganization of forest species is however a more like ly consequence of the interaction between climate change and anthropic interventions, via combined effects of fires, logging or farming. For instance, deforestation and fires in eastern US te mperate forests led in many areas the rep lace ment of p ine with oa k o r of oak with maple species (Fisichelli et al., 2014a ). At higher latitudes , in northeastern US (boundary with boreal forests), the same causes led to shifts of tree species such as Abies balsamea (balsam fir) to Betula papyrifera (paper birch) and Populus tremuloides (quaking aspen) (Fisichelli et a l., 2014a). Notable transformations also occurred in boreal forests, where, for instance, recurrent fires cause in northeastern Asia significant replace ment of larch forests (Larix gmelinii) with secondary Betula (Betula platyphylla) forests (Zyryanova et al., 2007). Fires are also responsible for mu ltiple transformations tropical forests species, e.g. in Ama zonia, where it was suggested that total tree co mpositional recovery fro m this disturbance can take centuries or even millennia (Barlo w and Peres, 2008). In these ecosystems , even a rise in at mospheric CO2 concentrations seems to have a ma jor influence in destabilizing vegetation composition by stimu lating faster-growing trees that, by becoming dominant and growing denser, cause the decline of slower-growing trees (Laurance et al., 2004).

At the same t ime, forest commun ity shifts can be a direct cause of climate change. For instance, the past century's rising te mperatures and humidity deficits have resulted in significant changes in forest composition in California by favouring the dominance of oak (Quercus) over pine (Pinus) species (McIntyre et al., 2015). Another representative e xa mple of warming effects in the te mperate region over the past decades could be the gradual replace ment of the beech (Fagus sylvatica) with oak (Quercus ilex) in the Py renees (Lindner et a l., 2010) or the e me rgence of a mixed native and alien broad-leaved species in temperate forests in the Alps (Walther, 2010). The direct effect of climate change is also felt in other types of ecosystems, such as tropical forests (in western Africa), where an increase in the do minance of drier-forest species over wetter-forest species was noticed, against the background of drought intensificat ion in the past decades in th is region of the continent (Fauset et al., 2012). As e xpected, climate warming has already generated apparent changes in high- latitude ecosystems as well, due to the expansion of temperate broadleaf species in boreal forests, across the boreal- temperate transition zone (Fisichelli et a l., 2014b). At the same time, the influence of climate changes, by favouring alien plant species in a warme r c limate, can be a significant threat to native species and to forest ecosystem functionality a ll over the world (Walther et al., 2009;Peltzer et al., 2010).

Throughout this century, important spatial changes are e xpected to occur in tree species composition. While in the boreal region these transformations will be driven by forests poleward expansion (certain boreal species will invade the tundra, and will simu ltaneously be replaced in many areas by temperate forest trees) as a result of climate warming, in the tropical reg ion it is possible that the composition of Ama zon and African forests will be influenced by precipitation variability and at mospheric CO2 fert ilization (IPCC, 2014). Moreover, important changes in tropica l species will a lso occur due to tree plantations , considering that the global shift of industrial roundwood production from natural forests to forest plantations is projected to increase by 75% by 2050 (Kirilen ko and Sedjo, 2007). In te mperate forests, considering for instance European ecosystems, warming will determine co mposition shifts especially by the translocation of broadlea f forests (e.g. Fagus sylvatica) to the north and by their large -scale rep lace ment with oak forests (e.g. Quercus cerris), in the south and central regions of the continent (Hanewin kel et a l., 2013). Negative imp lications of forest distribution changes

will be both ecological (Thuille r et al., 2011) and economic . For the latter, it is estimated that the vast extent of Mediterranean oak fo rests with low economic va lue in Centra l Eu rope could generate total losses of several hundred billion Euros by 2100 (Hanewinkel et al., 2013).

The effects of tree species shifts in the climate system dynamics can vary based on the different types of forest biomes. In the tropical region, the replace ment of native forests with large-scale plantations (e.g. Eucalyptus) is a climat ic warming vector (Fig. 5), considering the natural tropical ecosystems' capacity to store huge carbon amounts or to cool the atmosphere through high evapotranspiration rates (Jackson et a l., 2008; Ban-We iss et al., 2011). In te mperate forests, considering a longer European timescale, it appears that the large-scale replace ment after 1850 of broadleaved forests with conifers (due to the high commerc ial success of trees like the Scots pine or Norway spruce) has contributed to climat ic warming in Centra l and Eastern Europe, due to changes in albedo and evapotranspiration regimes (Naudts et al., 2016). Moreover, the large-scale transformation of natural fo rests into managed forests may have determined the release of massive amounts of carbon into the atmosphere, as it is believed that carbon store in coarse woody debris, living bio mass, litter, and soil is 43% , 24% , 8%, and 6% sma ller in managed forests than in natural forests (Naudts et al., 2016). This is due to the fact that natural forests assimilate CO2 fro m the atmosphere (in special old-growth forests) more effic iently not only in the temperate region, but also in general, worldwide (Luyssaert et al., 2008). In the boreal (tundra) region, although the e xpansion of woody plants is currently incip ient, it see ms the replace ment of herbaceous species with shrubs and trees in the next decades will result in a decrease in albedo that will be high enough to generate a net regional warming effect (Fig. 5), even though the new woody plant species will a lso be able to generate climat ic cooling by intensifying their photosynthesis activity (Pearson et al., 2013).

Net primary productivity shifts

It is generally known that the net primary productivity (NPP – net flu x of at mospheric carbon into green plants per unit area in unit time ) has globally increased in the past decades due to direct anthropic CO2 e missions and climate change (Ne man i et al., 2003; Boisvenue and Running, 2006; De l Grosso et al., 2008; Gang et a l., 2013). A representative study (Ne man i et a l., 2003) found that, between 1982 and 1999, NPP increased by ~6.2% (3.4 Pg C) g lobally, especially in tropical forest (the Ama zon region alone accounts for 40% of the total global increase in NPP) and boreal ecosystems, which together accounted for 80% o f the NPP increase over the two decades . It appears CO2 at mospheric fe rtilization is largely responsible fo r the carbon flu x increase in forest ecosystems , as experimental studies showed an average NPP increase of 0.2% per 1-pp m (parts per million) increase in CO2 in the atmosphere (global atmospheric CO2 levels grew by

~30 pp m between 1980–2000) (Ne mani et al., 2003). This effect of a ir CO2 enrich ment is double in natural ecosystems like forests, compared to other types of biotic systems, e.g. agricultura l crops (wheat, rice , and soybean), for which an average yield increase of ~0.1% was found for each ppm added to the atmosphere (Ainsworth et al., 2008; Lobell et a l., 2011; Prăvălie et al., 2017a ). Ho wever, the notable NPP increase in tropica l forests can also be lin ked to the direct changes of the climatic system occurring in the Equatorial region, e .g. decrease of the cloud cover and the resulting solar radiation increase (Nemani et al., 2003).

Over a longer period (1901– 2000), other studies highlighted the 13% NPP g lobal increase, as well as a greater NPP enhancement in boreal ecosystems, compared to tropica l ecosystems (De l Grosso et al., 2008). In terms of climat ic parameters, it was suggested that NPP changes in the tropical reg ion we re d riven especially by changes in the prec ipitation regime, wh ile the NPP increase towards high latitudes was driven by rising temperatures (Del Grosso et al., 2008).

However, other recent studies showed that, in the past decade (2000–2009), the wa rmest since the beginning of instrumental te mperature records, the global NPP decreased by 0.55 Pg C per decade, as a result of the strong negative trend of primary productivity in the Southern Hemisphere (–1.83 Pg C per decade), wh ich offset the weaker positive trend in the Northern He misphere (1.28 Pg C per decade) (Zhao and Running, 2010). The studies showed an increase in NPP over large parts of the Northern He misphere (i.e. across 65% of the vegetated land area) due to the positive response of ecosystems to higher te mperatures . The response was however counteracted by a significant decrease in the Southern He misphere (across 70% of vegetated land area), due to wa rming -associated drying trends (Zhao and Running, 2010). Thus, there is currently no globally agreed-upon consensus regarding spatiotemporal patterns of carbon fluxes in forest ecosystems, due to certain diffe rences in terms of type and period of data used or process-based models (Anav et al., 2015). There is however a general agree ment on the partially positive response of forest productivity (and of vegetation in general), ma inly due to the increase of at mospheric CO2 concentrations and air te mperature in the past decades (Ne mani et al., 2003; Bonan, 2008; Del Grosso et al., 2008; Anav et al., 2015; Schimel et al., 2015).

At present, there are several ma jor restrict ive factors for NPP enhancement in nu merous global reg ions. Forest interactions with water, te mperature or light-limited regions across the world, as we ll as with other non-climatic factors,

e.g. nitrogen limitation, are the main pathways of decrease in photosynthetic rates and carbon storage in forest biomass (Zaehle et a l., 2014; Schime l et a l., 2015; Seddon et al., 2016). It is estimated that water, te mperature and radiation-limited areas cover over 40%, 33%, and 27% of Earth's vegetated surface, and they significantly dimin ish forest productivity ma inly in drylands, boreal, and tropical regions , respectively (Ne man i et al., 2003). To a considerable e xtent, the ecosystems located in these limited regions have an amplified response to ecological disturbances and a high sensitivity to environmental perturbations, as they are characterized by low productivity conditions in terms of hu midity, te mperature and solar radiat ion (Seddon et al., 2016). Nitrogen limitation is another restrictive factor for forest productivity that, interestingly, in many cases occurs as a consequence of CO2 at mospheric fertilization. It is known that the increase of the carbon flu x into ecosystems generates an increase in the ecosystems' nitrogen require ments, in order to support stimu lated plant growth under higher CO2 concentrations in the atmosphere (Luo et al., 2004). This increased demand generates a decline in soil minera l N availability for forest vegetation growth (as nitrogen sequestration increases in long-lived plant biomass and soil organic matter), and the process is called progressive nitrogen limitation (Luo et a l., 2004; Zaehle et al., 2014).

Forests photosynthetic enhancement, the response of these ecosystems to higher atmospheric CO2 concentrations and temperatures, has certain important implications for climate dynamics. Considering the e xa mp le of the first case, e xperimental analyses showed that an increase in anthropic at mospheric CO2 concentrations up to 550 pp m in the following decades, almost 40% mo re co mpared to the current state of affairs (~400 pp m), would determine an enrich ment of NPP of over 20%, at least in te mperate forests (Norby et al., 2005). While this NPP inc rease is generally associated with climate cooling (due to h igher ca rbon absorption fro m the at mosp here), an e xtensive tree cover in the boreal region, against the background of more intense forest productivity at high northern latitudes (Figs. 5, 6), can generate climate warming at the same t ime by decreasing surface a lbedo (Bonan, 2008). Moreover, it is e xpected that the in itia l increase in g lobal photosynthetic activity will be succeeded by a decrease in NPP, as the stomatal conductance of forest vegetation dimin ishes as a result of the high atmospheric CO2 e xcess (Bonan, 2008). The consequences of partial stomatal closure (and, imp lic itly, of a decreasing gas e xchange between at mosphere-biosphere) are connected to global wa rming, for e xa mp le through the

decrease in evapotranspiration (a p rocess that consumes heat) especially in tropica l forests, in which the most intense water (and carbon) exchange with the atmosphere occurs (Lewis et al., 2015).

Biogeochemical shifts

This discreet and co mple x change in global forests is connected to the disruption of nutrient cycles as a result of climate change, anthropogenic emissions of carbon dioxide and the increasingly high presence of reactive nitrogen in the atmosphere. In this context, the analysis of certain biogeochemica l shifts is particularly re levant for certain chemica l ele ments that are essential for ecosystem processes such as carbon, nitrogen (N) and phosphorus (P). These nutrients were by far the most intensely studied in international lite rature in terms of dynamics and the changes that occur in their interactions, against the background of climate change and escalating air pollution (Langley and Megonigal, 2010; Berna l et al., 2012; Sardans et al., 2012; Peñuelas et al., 2013; Yang et al., 2014; Fazhu et al., 2015; Meyerholt and Zaehle, 2015).

As presented at length in this review, especially in the phenology and net primary productivity sections, the carbon cycle in forest biomass was generally changed globally, considering large-scale a ir CO2 enrich ment (against the background of direct anthropic e missions ) and growing season lengthening (as a result of climatic wa rming), which resulted in increasingly high amounts of carbon being stored in forest biomass. Also, the increasing deposition of reactive nitrogen (currently being produced up to five t imes more than one century ago ), as a consequence of the constant use of agricultural fe rtilizers and fossil fue ls (wh ich led to the increase of at mospheric N concentrations), led to a large increase of N availability for forest ecosystems, which resulted in significant changes in the global N cycle (Janssens et al., 2010; Meunier et al., 2016). However, these changes in the C and N cycles in forest systems do not act strictly independently. For instance, changes in carbon's biogeochemica l cyc le have significant repercussions for the ecological stoichiometry state – the balance of mult iple che mica l e le ments in ecologica l p rocesses (Elser and Ha milton, 2007), which causes profound changes in both N and P cycles (Peñuelas et al., 2013; Wieder et al., 2015).

It is known that forest NPP enhancement necessitates an increase in nitrogen require ments in order to meet the stoichiometric de mands of vegetation development (Wieder et al., 2015). Th is require ment was met in the past decades in many boreal and te mperate forests (where this was not the case, the process of progressive nitrogen limitation was triggered) (Fig. 4), which, a lthough generally limited by the low natural availability of N, e xperienced apparent rates of nitrogen deposition considering the vast industrialized reg ions (N sources) found in the temperate zone (Janssens et al., 2010). Even in the less industrialized tropica l regions , N additions enhanced carbon dioxide fe rtilizat ion effects on forests, which however benefit fro m a much greater natural N availability co mpared to temperate and boreal forests (Hiet z et a l., 2011). Nonetheless, there is a high degree of uncertainty with regard to the mechanism through which N inputs influence the C cycle in many forest sites of the world, as a result of the interference of other variables such as water availability or the role and interaction of other nutrients that can significantly influence the two biogeochemical cycles (Peñuelas et al., 2011, 2012). Overa ll, it was however signalled that low rates of N deposition on forestlands can be beneficia l for stimulat ing bio mass development, as the excessive N deposition on terrestrial and aquatic ecosystems is generally associated with diverse negative ecological e ffects such as soil acidification, eutrophication or biodiversity loss ( Janssens and Luyssaert, 2009; Janssens et al., 2010; Hietz et al., 2011).

At the same t ime, while the C–N balance re ma ined re lative ly stable in many forests across the globe, as a result of a simu ltaneous increase of CO2 and reactive nitrogen concentrations (as NOx and NHy, since in its N2 common form is not accessible to most plants) (Bala et al., 2013), an apparent imba lance was noticed for P, as it did not record similar increases.

Recent estimations show that global at mospheric phosphorus deposition values currently reach ~3– 4 Tg P yr−1 (this lo w amount is linked to limited sources, such as dust and wildfire ash particles in the atmosphere), compared to reactive nitrogen deposition, which e xceeds 100 Tg N yr−1 (Peñuelas et al., 2012, 2013). Considering the low availability of this nutrient in the environment, it was suggested it could become more limiting if C and N addit ions continue to rise at high rates. In fact, instances in which the increase of N deposition caused a shift fro m N to P limitation we re a lready documented (Peñuelas et al., 2012). This overall state with higher increases of N inputs compared to P inputs (and therefore with a rise in the N:P rat io) was highlighted in forest bio mass and soils in te mperate (Peñuelas et al., 2013; Jonard et a l., 2015) and tropical (Huang et al., 2012; Yang et al., 2014) regions, while in certain boreal regions (Fennoscandian forests) the state appears to be relatively stable or even reversed (Högberg et al., 2017). Interestingly, in d ryland ecosystems (including forestlands) there are signs of contrary trends, in which nutrient decoupling occurs through the decrease of soil organic C and total N concentrations, due to vegetation degradation trends, and through the rise in inorganic P concentration, in the context of the dominance of mechanical rock weathering process (Delgado -Baquerizo et al., 2013).

Biogeochemica l changes can have profound effects for the dynamics of the climate system. It was suggested that one kg of deposited N can determine an increase of 15–60 kg of C uptake in trees (Meunier et al., 2016), although certain studies signalled a much higher nitrogen-carbon interaction – up to 400 kg of forest carbon storage per kg of N deposition (Magnani et al., 2007). However, regard less of the intensity of this relation, it can be stated that the climatic warming recorded in this century will be influenced by the availability of nutrients, which will shape the world's forests' large-scale carbon storage capacity (Arneth et al., 2010). It is however h ighly likely that N and P ava ilab ility will be insufficient for ma intaining the nutrient de mand for projected productivity increases of the ecosystems , which results in a warming e ffect – certain pro jections show that N limitation alone can reduce global terrestrial carbon storage by up to 149 Pg C by 2050 (Wang and Houlton, 2009). A lthough some models indicate a decrease in terrestria l C storage especially in mid-high latitude ecosystems, against the background of N dynamics (Zaeh le et al., 2010), and a decrease in tropica l ecosystem productivity, as a result of possible P trends (Wieder et a l., 2015), there is currently a high degree of uncertainty with regard to the extent to which nitrogen and phosphorus cycles will limit carbon sequestration in forests worldwide.

Fig. 6. Synthesis of diverse statistical data on the 12 g lobal forest perturbations. Note: GSL – gro wing season length, FWSL

– fire weather season length, NPP – net prima ry productivity, CS – ca rbon storage, CSI – carbon storage increase (in forest biomass), CSD – carbon storage decrease (in forest biomass), ECD – evaporative cooling decrease, SAI – snow albedo increase, SAD – snow a lbedo decrease; the information presented in the "climate impact" section is comprehensive for CSI and CSD, considering the abundance of quantitative data on carbon and climate forc ing in various specialized papers; the notified SAI and SAD processes are representative for boreal and temperate forests (e.g. in the case of biogeochemica l shifts, SAD is representative for te mperate and especially for boreal forests, considering the intensification of the carbon cycle and its imp licat ions for forest biomass increase at high latitudes; another exa mp le, in the case of composition shifts, SAD is representative for boreal forests due to the early replace ment of herbaceous species by shrubs and trees; moreover, this e xpansion of woody plants is the reason for including CSI, although CSD was simu ltaneously signalled in the sa me box, but which is however ma inly representative for tropical fo rests, marked by the large-scale replacement of natural tree species by plantations); CSI and CSD processes were only considered in terms of forest bio mass, as carbon dynamics in soil, peat and underly ing permafrost deposits, triggered direct ly or indirectly by forest perturbations, were not taken into account.

Major strate gies for combating the world forest perturbations

Fighting the multid imensional proble ms of g lobal forest ecosystems are currently fac ing requires mult iple and comple x strategies, which can be grouped under five major actions – mitigate, adapt, repair, protect and research (Fig. 7). These main pathways were proposed as strategies in solving other global environmental issues, e.g. ocean acidification (Billé et a l., 2013; Gattuso et al., 2015), but they were adapted for the part icularities of this specific proble m of the oceanic environment. The basic principles of these strategies can also be applied for the terrestrial environmental d isturbance analysed in this paper, if accordingly adapted to the concrete context of global forest perturbations.

The most important action for improving the state of forests can be considered the mitigation of climate change and anthropic pressures (Fig. 7), considering it addresses the fundamental disruption causes of these terrestrial ecosystems. In terms of climate change mitigation, notable international efforts are already being made in an attempt to drastically reduce carbon emissions. The Paris Agreement is the most important such global in itiative (under the aegis of UNFCCC), the ma in objective of which is stabilizing global warming under 2 °C by 2100, relat ive to preindustrial levels (Fawcett et al., 2015). However, in order to eliminate the pressure of climat ic warming on forest ecosystems, a rap id decarbonizat ion of the world 's economies is necessary, considering that a >66% probability of meeting the 2°C Paris target imp lies a reduction in global carbon e missions from the current ~37 Pg CO2 (or ~10 Pg C) (Le Quéré et a l., 2016) to ~5 Pg CO2 by 2050 (Rockström et al., 2017).

At the same time, the re moval of this greenhouse gas from the at mosphere can be another solution for fighting climate change (and, imp licit ly, for other forest-re lated ecologica l issues), co mple mentary to global policies to reduce carbon emissions. In this situation, the options imp ly the use of " Carbon Dio xide Re mova l" (CDR) geo-engineering, i.e . techniques for permanently storing CO2 in various natural deposits such as sediments, crust, soils, ocean or terrestria l biosphere (Vaughan and Lenton, 2011). It was suggested that CDR techniques imply less risks for the Earth's systems and can be a promising solution for limit ing global wa rming under 2 °C, if trad itional solutions for reducing carbon emissions are not enough for stabilizing the p lanet's climatic system over the ne xt decades (Swa rt and Ma rinova, 2010). In fact, these geo-engineering solutions are becoming increasingly important, as certain recent studies show that the 2 °C objective (or

1.5 °C, also considered by the Paris Agreement) is h ighly unlike ly to be met by 2100 with the current polic ies that are only targeting the reduction of carbon emissions (Raftery et al., 2017).

In this context, it appears that even the second proposed geo-engineering solution, "Solar Radiat ion Management" (SRM), could be a valid solution for stabilizing climate warming, even though it only addresses the global radiation balance (climate cooling), instead of removing carbon fro m the atmosphere (Vaughan and Lenton, 2011). Recently, an interesting study showed that the injection of ca lc ite aerosol partic les into the stratosphere could cool the planet surface while simu ltaneously reconstructing the ozone layer degraded by anthropogen ic activities (Ke ith et al., 2016). Therefore , it see ms these aerosols are an e xcellent alternative to sulphate aerosols, the most frequently previous option proposed of SRM, which does however imply major risks for the planet, such as the damage (loss) of the ozone layer (Keith et al., 2016).

The mitigation of anthropic pressures is also crucia l, and the reduction of defo restation is the most important step in this respect. One of the ma in init iatives is that of UNFCCC, wh ich, through the REDD+ progra mme , is trying to reduce deforestation in developing countries and, implicit ly, global carbon emissions (Romijn et al., 2012). A lthough the REDD+ fra me work ma inly aims to reduce deforestation in developing countries by providing financia l compensations from developed countries (based on a mon itoring, reporting and assessment system for the forest carbon stock and forest area changes in developing countries), it appears the mechanism's progress is re latively slow. In terms of financial support, only

~$9 billion were g ranted between 2006– 2014 via REDD+, which is we ll below the global require ments for stabilizing deforestation and carbon emissions . For instance, it was suggested that by 2030 between $17– 33 b illion per year could be necessary for reducing deforestation e missions to half their current value (Fischer et al., 2016). Therefore , a much faster (but carefully supervised) financing flow to developing countries is vital in order to urgently halt defo restation especially in deforestation hotspot countries, such as Indonesia, where the effects of the UNFCCC mechanis m have been min ima l (considering that the deforestation rate continued to rise, even after 2011) (Busch et al., 2015). However, in Bra zil, where annual deforestation rates dropped by 76% in the A ma zon between 2005–2012, it appears the REDD+ mechanism had a significant contribution for the preservation of forest areas, alongside other various causes that dramatica lly reduced deforestation, e.g. enforcement of forest la ws, establishment of new protected areas , and enhanced control of illega l logging by state and federal agencies (Nepstad et al., 2014; Azevedo et al., 2017).

In addition to this in itiative that also targets the fragmentation process (considering that frag mented forests have a lower ca rbon storage capacity) (Chaplin-Kra mer et al., 2015), the 2014 Ne w York Decla ration on Forests is another important vector of the UN that could generate actual effects in limit ing non-sustainable anthropic actions in forestlands. This international agreement is aimed at halving the rate of global deforestation by 2020 and completely e liminating deforestation of natural forests by 2030 (Houghton et al., 2015). In this conte xt, short-term perspectives on mit igating direct anthropic aggression on forests are promising.

However, fighting anthropic pressures on forest ecosystems also entails reducing air pollution. In a broader context, reducing air pollution was regulated a lmost four decades ago through the 1979 Convention on Long-Range Transboundary Air Pollution (CLRTAP), which re ma ins to this day the most important international air quality treaty (Rypdal et a l., 2005; Re is et al., 2012). The convention targets a wide range of atmospheric pollutants , including nitrogen and sulphur compounds (which can have a severe negative effect on forest ecosystems and mo re ), but also tropospheric ozone (the most important to xic gas for forests ), most notably regulated by the latest CLRTAP protocol – the 1999 Gothenburg Protocol (Re is et al., 2012). However, considering the geographical coverage of signatory countries – Canada, United States, European countries (inc luding Russia) and several states in western and central Asia –, an important shortcoming of th is international init iative consists in not inc luding highly industrialized states fro m ce rtain key-regions, e.g. eastern Asia (Ch ina, South Korea and Japan), known for obvious upward trends in O3 levels in the last decades (Cooper et al., 2014). Therefore, in order to mitigate air pollut ion and ensure the protection of global forests against this environmental threat, it is necessary to geographically e xpand the CLRTAP international fra me work or to establish similar conventions in regions that are experiencing a severe degradation of ecosystems due to atmospheric pollution.

Another crucial option for stabilizing forests in a future dominated by various environmental stressors is adaptation through different strategies such as assisted colonization, reversing defaunation, creating connectivity corridors or increasing species diversity (Fig. 7). Assisted colonization (also known as assisted migrat ion or managed re location), which entails the intentional movement of plant or an ima l species outside their historic range to avoid e xtinction due to climate change (Hoegh-Guldberg et a l., 2008), can also be a valid option for forest conservation , considering that tree species don't generally have a suffic ient dispersion capacity to keep up with climate change, wh ich is responsible for the rapid shift of suitable environmental conditions toward the poles or higher altitudes. However, imp le menting the assisted colonization concept depends on certain variables such as identifying a ne w habitat with appropriate biophysical conditions for meeting the ecological needs of tree species, as well as on considering the ecological risks (transfer of pests or diseases to the new

habitat, which can disturb the other native species and ecosystems ) associated with forest translocation procedure across the biogeographical boundaries (Hoegh-Guldberg et al., 2008; Minteer and Collins 2010; Iverson and McKenzie, 2013).

Reversing forest defaunation, or restoring anima l species in fo rest environments , is another essential adaptation strategy for ma intaining forest productivity in the age of environmental changes. It was signalled that translocating certain key animal species in defaunated forests (via various procedures, e.g. reintroduction or other more sophisticated options such as de-ext inction – the resurrection of e xt inct species ) can restore many ecosystem processes and services (including carbon sequestration) especially in e mpty forests of the world (Seddon et al., 2014; Galetti et al., 2017). Th is important action can be carried out under the aegis of key international treaties such as the Convention on Biological Diversity.

Another option for maintain ing forest health is creating connectivity corridors between forest habitat patches present all over the world, considering that they generally reduce the negative effects of fragmentation by promoting plant and animal move ment between patches (generally ~50% more co mpared to patches unconnected with corridors), preventing local e xt inctions and increasing biodiversity (Da mschen et al., 2006; Gilbert -Norton et a l., 2010). Such corridors will be vital especially in the ne xt decades, when major e xpansions of roads (Laurance et a l., 2014) and urban areas (Seto et al., 2012) a re e xpected in many reg ions of the world, wh ich will inevitably a mp lify the large-scale forest frag mentation process. At the same time, in addition to this strategy, another viable option for forest adaptation in a changing world is promoting tree species diversity by forest managers (as mixed species are generally mo re resistant to various environmental disturbances), which would a lso enhance forest genetic resource conservation – presently a topic of interest for certain important international init iatives , i.e. the European Forests Genetic Resources Progra mme (Fa res, 2015), and the Global Plan of Action for the Conservation, Sustainable Use and Development of Forest Genetic Resources (FAO, 2014b).

Fig. 7. Sche matic approach on the five majo r actions that can be taken by governments worldwide in order to co mbat global forest perturbations. Note: UNFCCC – United Nat ions Fra mework Convention on Climate Change, REDD – Reducing Emissions from Deforestation and forest Degradation, CLRTAP – Convention on Long-Range Transboundary Air Pollution, EUFORGEN – European Forest Genetic Resources Programme, FAO – Food and Agriculture Organization, GPA CSDFGR – Global Plan of Action for the Conservation, Sustainable Use and Development of Forest Genetic Resources, UNCCD – United Nat ions Convention to Co mbat Desertification; the list of politica l pathways through which each ma jor strategy can be imple mented is non -exhaustive; national government init iatives can be diverse and specific to each individual state (e.g. the Three-North Shelterbelt Forest Program of China, a pathway of repair strategy).

There are also other pathways for fighting forest perturbations, which can be catalogued as repair management options (Fig. 7). An important such option is afforestation – planting of trees on lands that were not previously covered by forests, which is different fro m reforestation, wh ich entails rep lanting trees after a forest is cut down, i.e . p lanting in an e xisting fo rested area. There are currently several ma jor actions in this respect in ce rtain trop ical and te mperate regions , particularly in Ch ina (IPCC, 2014). The Green Great Wall Progra m or the Three-North Shelterbelt Forest Progra m (launched in 1978 and foreseen to be finalized in 2050) is one of the most amb itious afforestation projects in the world, which so far, in accordance with its purpose of imp roving degraded lands, has been at least partially successful in mitigating desertificat ion and dust storm issues in the northern drylands of the country (Wang et al., 2010; Tan and Li, 2015). In addition to Ch ina, there are other countries (located in the tropica l region) with large annual rates of a fforestation (>100000 ha in 2010), i.e . Bra zil, Tanzania , India, Indonesia and Vietnam (FAO, 2015). At the same time, reforestation is a

ma jor conservation policy (ma inly in boreal forests) in Canada, United States , Sweden, Finland, Russia, Ch ina, Indonesia and Malaysia, considering the same high rate in 2010 (FAO, 2015).

Restoring degraded forestlands (which inc ludes, in addition to tree planting, other anthropic interventions that aim to restore forest ecosystem processes, functions and services) is another dimension of the repair options that could be e xplored. In fact, this strategy is already underway g lobally, as part of international polic ies such as the Bonn Challenge (2011) or the New Yo rk Decla ration on Fo rests, one of the objectives of which is to restore 150 million ha of degraded forestlands and landscapes by 2020 and an additional 200 million ha by 2030 (Climate Focus, 2015; Suding et al., 2015). The Atlantic Forest Restoration Pact, the largest forest restoration program in South America, is a relevant e xa mple of the current imple mentation of such environmental polic ies in the world (Pinto et al., 2014). These political efforts are comple mentary to other recent international initiatives that aim to stabilize degraded lands (e.g. Land Degradation Neutrality of the UN Convention to Combat Desertification) (Prăvă lie et al., 2017b) and represent a major pathway to meet ing the UN Sustainable Development Goal 15, i.e . ensuring sustainably managed forests, combating desertification, halting and reversing land degradation, and halting biodiversity loss by 2030 (United Nations, 2016).

An effective solution for fighting forest disturbances is also protecting them (Fig. 7), first and fore most by expanding protected areas worldwide, considering that many forest areas are still insufficiently we ll p rotected globally (Sch mitt et a l., 2009). The more these protected areas will integrate vast and numerous forest areas, the less vulnerable forests will be to various climatic and anthropic stressors. A major opportunity for such a forest ecosystem (and other types of terrestrial ecosystems) protection strategy is the Nagoya protocol (2010) of the Convention on Biologica l Diversity, which a ims to establish the forma l protection of at least 17% of the world terrestria l area by 2020 (Joppa et al., 2013). This international fra me work can also be a viable support for the protection of ecological refugia areas (locations where tree/plant/anima l species may retreat o r migrate due to more favourable environmental conditions ), less e xposed to the negative effects of climate change (Keppel et a l., 2015). Moreover, continuing interdisciplinary research on both forest disturbance patterns (causes, progress mechanisms, effects) and control solutions (Fig. 7) is an effective strategy for reaching an in-depth understanding fighting multidimensional forest perturbations across the globe.

Conclusions

Forests are probably the most important type of terrestria l ecosystem for ensuring the p lanet's health in an age of ma jor environmental changes . They are among the Earth's few b iophysical systems that are able to substantially influence the climat ic system on a g lobal scale, through energy exchanges with the at mosphere, by changing the terrestrial surface radiation balance, and especially by controlling and storing immense carbon amounts in live bio mass, deadwood, litter or in the superficia l soil layer. A lso, forest ecosystems are home to many millions of plant and anima l species, and therefore are a vital support for the global terrestria l biodiversity. They a re at the sa me t ime an essential p illar fo r the very e xistence of humanity, due to the continuous and large-scale delivery of several categories of ecosystem services, wh ich are classified in the Millennium Ecosystem Assessment framework as provisioning, regulating, supporting and cultural services .

These biotic systems are also a major natural pathway through which g lobal warming can be controlled, and is therefore a key lin k in ma intaining the stability of the other planetary systems and, implicit ly, of Earth as a whole. But th e influence of these terrestrial ecosystems in ensuring the Ea rth system's functioning is growing increasingly weak, as they are subjected to various fast and extensive transformations, which have already s ignificantly eroded the forests' capacity to mitigate c limate change through carbon sequestration and the evaporative cooling process . For the most part, human society

is directly or indirect ly responsible for the gradual or rapid triggering of a wide range of forest perturbations (phenological shifts, range shifts, die-off events, insect infestations , deforestation, frag mentation, air pollution, defaunation, fires, composition shifts, net prima ry productivity shifts and biogeochemica l shifts), the consequences of which a re in most cases disastrous for both humans and other species. These catastrophic effects are mult idimensional, and inc lude, a mong others, the intensification of global warming, as a result of the direct impact or of positive feedback mechanisms of forest disturbances in temperature increase in the atmosphere.

While some models indicate that a forestless world could generate a net cooling in fluence on global c limate, associated with a greater importance of surface albedo increase compared to warming carbon -cycle effects, humanity should be more cautious with regard to this finding, considering the constantly decreasing importance of the albedo in an increasingly wa rme r world, especially at h igh lat itudes. At the same time , hu man society should be aware that, in a broader context of co mple x interactions between planetary systems, a hypothetical disappearance of forests would most like ly accentuate global warming due to massive increases in CO2 concentrations in the atmosphere (considering that these ecosystems store roughly one quarter of annual carbon anthropogenic emissions) and, subsequently, in the ocean, thus causing an accelerated acidificat ion and affecting its role as th e planet's largest carbon reservoir. Therefore , regard ing the same e xa mp le, mankind should not rely on the climatic importance of the albedo in a hypothetical free -fro m-forests world or in the context of megadisturbed global forest ecosystems, as the worsening of ocean acidification wou ld most like ly result in additional warming due to carbon flux changes in the ocean-atmosphere boundary.

In order to address the comple x threat of forest perturbations – climate warming and to mit igate/fight these ma jor environmental issues , it is imperat ive that all of the world's governments work together closely and effic iently. In th is respect, there is a wide range of strategies currently available and discussed in this paper, such as mit igate, adapt, repair, protect and research options. If applied rap idly, on a large scale and in a mixed manner, these anthropic solutions should b e successful in mitigating forest ecological issues and climate change by the end of this century.

References

Abram, N.J., Mc Gregor, H.V., Tierney, J.E., Evans, M.N., McKay, N.P., Kaufman, D.S., Thiru mala i, K., Martrat, B., et al., 2016. Early onset of industrial-era warming across the oceans and continents. Nature 536, 411–418.

Achard, F., Beuchle, R., Mayau x, P., St ibig, H.J., Bodart, C., Brink, A., Ca rboni, S., Desclée, B., et a l., 2014. Determination of tropica l deforestation rates and related carbon losses from 1990 to 2010. Global Change Bio logy 20, 2540–2554.

Ainsworth, E.A., Leakey, A.B.D., Ort, D.R., Long, S.P., 2008. FACE-ing the facts: inconsistencies and interdependence among field, cha mber and modeling studies of elevated [CO2] impacts on crop yie ld and food supply. Ne w Phytol. 179, 5–9.

Ainsworth, E.A., Yendrek, C.R., Sitch, S., Collins, W.J., Embe rson, L.D., 2012. The effects of tropospheric ozone on net primary productivity and implications for climate change. Annu. Rev. Plant Biol. 63, 637–661.

Allen, C.D., Macalady, A.K., Chenchouni, H., Bachelet, D., McDo well, N., Vennetier, M ., Kitzberger, T., Rig ling, A., et a l., 2010. A global overview of drought and heat -induced tree mortality revea ls e merging c limate change risks for forests. Forest Ecology and Management 259, 660–684.

Allen, C.D., Breshears, D.D., Mc Dowe ll, N.G., 2015. On underestimation of global vulnerability to tree mo rtality and forest die-off from hotter drought in the Anthropocene. Ecosphere 6, 1–55.

Anav, A., Friedlingstein, P., Beer, C., Cia is, P., Harper, A., Jones, C., Murray-Tortarolo, G.,Papale, D., et al., 2015.

Spatio-temporal patterns of terrestrial gross primary production: A review. Rev. Geophys. 53, 785–818.

Anderegg, W.R.L., Kane, J.M., Anderegg, L.D.L., 2012. Co nsequences of widespread tree mortality triggered by drought and temperature stress. Nature Climate Change, doi: 10.1038/nclimate1635.

Anderson, R.G., Canadell, J.G., Randerson, J.T., Jackson, R.B., Hungate, B.A., Baldocchi, D.D., Ban -Weiss, G.A., Bonan, G.B., et al., 2011. Biophysical considerations in forestry for climate protection. Frontiers in Ecology and the Environment 9, 174–182.

Aragão, L.E.O., Malhi, Y., Ro man-Cuesta, R.M., Saatchi, S., Anderson, L.O., Sh imabukuro, Y.E., 2007. Spatial patterns and fire response of recent Amazonian droughts. Geophysical Research Letters 34, doi:10.1029/ 2006GL028946.

Aragão, L.E.O., Malh i, Y., Barb ier, N., Lima, A., Shimabuku ro, Y., Anderson, L., Saatchi, S., 2008. Interactions between rainfall, deforestation and fires during recent years in the Brazilian A ma zonia. Ph il. Trans. R. Soc. B 363, 1779– 1785.

Arneth, A., Harrison, S.P., Zaehle , S., Tsigaridis, K., Menon, S., Bart lein, P.J., Feichter, J., Korhola, A., et a l., 2010.

Terrestrial biogeochemical feedbacks in the climate system. Nature Geoscience 3, 525–532.

Arora, V.K., Peng, Y., Kurz, W.A., Fyfe, J.C., Hawkins, B., Werner, A.T., 2016. Potential near -future carbon uptake overcomes losses from a large insect outbreak in British Columbia, Canada. Geophysical Re search Letters 43, 2590–2598.

Ashmore, M.R., 2005. Assessing the future global impacts of ozone on vegetation. Plant, Cell and Environment 28, 949–964.

Asner, G.P., Townsend, A.R., Braswe ll, B.H., 2000. Satellite observation of El Niño e ffects on A ma zon fo rest phenology and productivity. Geophysical Research Letters 27, 981–984.

Auffret, A.G., Plue, J., Cousins, S.A.O., 2015. The spatial and te mpora l co mponents of functional connectivity in fragmented landscapes. Ambio 44, 51–59.

Azevedo, L.B., van Zelm, R., Hendriks, A.J., Bobbink, R., Hu ijbregts, M.A.J., 2013. Global assessment of the effects of terrestrial acidification on plant species richness. Environmental Pollution 174, 10–15.

Azevedo, A.A., Ra jão, R., Costa, M.A., Stabile , M.C.C., Macedo, M.N., dos Re is, T.N.P., Alencar, A., Soares -Filho, B.S., et al., 2017. Limits of Brazil's Forest Code as a means to end illegal deforestation. PNAS 114, 7653–7658.

Baccin i, A., Goetz, S.J., Wa lke r, W.S., Laporte, N.T., Sun, M., Sulla-Menashe, D., Hackler, J., Beck, P.S.A., et a l., 2012. Estimated carbon dio xide e missions from tropica l de forestation improved by carbon -density maps. Nature Climate Change 2, 182–185.

Ba la, G., Caldeira, K., Wic kett, M., Phillips, T.J., Lobell, D.B., Delire , C., Mirin, A., 2007. Co mb ined climate and carbon-cycle effects of large-scale deforestation. PNAS 104, 6550–6555.

Ba la, G., Devara ju, N., Chaturvedi, R.K., Ca ldeira, K., Ne mani, R., 2013. Nitrogen deposition: how important is it for global terrestrial carbon uptake?. Biogeosciences 10, 7147–7160.

Ban-Weiss, G.A., Ba la, G., Cao, L., Pongratz, J., Calde ira, K., 2011. Climate forcing and response to idealized changes in surface latent and sensible heat. Environ. Res. Lett. 6, doi:10.1088/1748-9326/6/ 3/034032.

Bandoc, G., Prăvălie, R., Patriche, C., Drago mir, E., To mescu, M., 2018. Response of phenological events to climate warming in the southern and south-eastern regions of Romania. Stoch. Environ. Res. Risk Assess. 32, 1113–1129.

Barlow, J., Peres, C.A., 2008. Fire-mediated dieback and compositional cascade in an A ma zonian forest. Phil. Trans.

R. Soc. B 363, 1787–1794.

Barnosky, A.D., Mat zke , N., To miya, S., Wogan, G.O.U., Swart z, B., Quental, T.B., Ma rshall, C., Mc Gu ire , J. L., et al., 2011. Has the Earth's sixth mass extinction already arrived? Nature 471, 51–57.

Be llo, C., Ga letti, M., Pizo, M.A., Magnago, L.F.S., Rocha, M .F., Lima, R.A.F., Pe res, C.A., Ovaskainen, O., et a l., 2015. Defaunation affects carbon storage in tropical forests. Science Advances 1, e1501105.

Bendix, J., Ro llenbeck, R., Richter, M., Fab ian, P., Emc k, P., 2008. Climate. Grad ients in a tropical mountain ecosystemof Ecuador, 63–73, Berlin, Germany.

Bennett, A.C., Mc Dowe ll, N.G., Allen, C.D., Anderson -Teixe ira, K.J., 2015. Larger trees suffer most during drought in forests worldwide. Nature Plants 1, doi:10.1038/nplants.2015.139.

Bentz, B.J., Regnie re, J., Fettig, C.J., Hansen, E.M., Hayes, J.L., Hic ke, J.A., Kelsey, R.G., Negron, J.F., et al., 2010. Climate change and bark beetles of the western United States and Canada: Direct and indirect effects. Bioscience 60, 602– 613.

Berna l, S., Hedin, L.O., Likens, G.E., Ge rber, S., Buso, D.C., 2012. Co mple x response of the forest nitrogen cycle to climate change. PNAS 109, 3406–3411.

Billé, R., Ke lly, R., Biastoch, A., Harrould-Kolieb, E., Herr, D., Joos, F., Kroeker, K., Laffoley, D., et a l., 2013. Taking action against ocean acidification: A rev iew of manage ment and policy options. Environ mental Management 52, 761–779.

Billings, S.A., Gaydess, E.A., 2008. Soil nit rogen and carbon dynamics in a frag mented landscape experiencing forest succession. Landscape Ecology 23, 581–593.

Blake, S., Dee m, S.L., Mossimbo, E., Maisels, F, Wa lsh, P., 2009. Forest elephants: Tree planters of the Congo.

Biotropica 41, 459–468.

Blok, D., Heijmans, M.M.P., Schaepman-Strub, G., Kononov, A.V., Ma ximov, T.C., Be rendse, F., 2010. Shrub expansion may reduce summer permafrost thaw in Siberian tundra. Global Change Biology 16, 1296–1305.

Boisvenue, C., Running, S.W., 2006. Impacts of climate change on natural forest productivity – evidence since the middle of the 20th century. Global Change Biology 12, 862–882.

Bokhorst, S., Pedersen, S.H., Brucker, L., Anisimov, O., Bjerke, J.W., Brown, R.D., Ehrich, D., Essery, R.L.H., et al., 2016. Changing Arctic s now cover: A review of recent developments and assessment of future needs for observations, modelling, and impacts. Ambio 45, 516–537.

Bonan, G.B., 2008. Forests and climate change: Forc ings, feedbacks, and the climate benefits of forests. Science 320, 1444–1449.

Bond, W.J., Woodward, F.I., M idgley, G.F., 2005. The global distribution of ecosystems in a world without fire.

New Phytologist 165, 525–538.

Bond, W.J., 2008. What limits trees in C4 grasslands and savannas? Annual Review of Ecology, Evolution and Systematics 39, 641–659.

Bonfils, C.J.W., Ph illips, T.J., La wrence, D.M ., Ca meron -Smith, P., Riley, W.J., Subin, Z.M., 2012. On the influence of shrub height and expansion on northern high latitude climate. Env iron. Res. Lett. 7, doi:10.1088/1748 – 9326/7/1/015503.

Bowman , D.M .J., Ba lch, J.K., Arta xo , P., Bond, W.J., Carlson, J.M., Cochrane, M.A., D'Antonio, C.M ., De Fries, R.S., et al., 2009. Fire in the Earth system. Science 324, 481–484.

Bowman , D.M.J., Ba lch, J.K., Arta xo, P. , Bond, W.J., Cochrane, M.A., D'Antonio, C.M., De Fries, R.S., Johnston, F.H., et al., 2011. The human dimension of fire regimes on Earth. Journal of Biogeography 38, 2223–2236.

Boyd, I.L., Free r-Smith, P.H., Gilligan, C.A., God fray, H.C.J., 2013. The consequence of tree pests and diseases for ecosystemservices. Science 342, doi: 10.1126/science.1235773.

Brando, P.M., Balch, J.K., Nepstad, D.C., Morton, D.C., Putz, F.E., Coe, M .T., Silvério, D., Macedo, M.N., et a l., 2014. Abrupt increases in Amazonian tree mortality due to drought –fire interactions. PNAS 111, 6347–6352.

Broadbent, E.N., Asner, G.P., Keller, M., Knapp, D.E., Oliveira, P.J.C., Silva, J.N., 2008. Forest frag mentation and edge effects from deforestation and selective logging in the Brazilian Amazon. Biological Conservation141, 1745–1757.

Brooks, T.M., Pimm, S.L., Oyugi, J.O., 1999. Time lag between deforestation and bird ext inction in tropica l forest fragments. Conservation Biology 13, 1140–1150.

Bro wn, M., Black, T.A., Nesic, Z., Foord, V.N., Spittlehouse, D.L., Fredeen, A.L., Grant, N.J., Burton, P.J., et a l., 2010. Impact of mountain p ine beetle on the net ecosystem production of lodgepole pine stands in British Colu mbia . Agricultural and Forest Meteorology 150, 254–264.

Bucke ridge, K.M., Zufe lt, E., Chu, H., Grogan, P., 2010. Soil nit rogen cycling rates in low a rctic shrub tundra are enhanced by litter feedbacks. Plant Soil 330, 407–421.

Buitenwe rf, R., Rose, L., Higgins, S.I., 2015. Three decades of mult i-d imensional change in global leaf phenology.

Nature Climate Change 5, 364–368.

Burke, K.L., 2011. The effects of logging and disease on American chestnut. Forest Ecology and Management 261, 1027–1033.

Burro ws, M.T., Schoeman, D.S., Richardson, A.J., Molinos, J.G., Ho ffmann, A., Buckley, L.B., Moore, P.J., Brown, C.J., et al., 2014. Geographical limits to species -range shifts are suggested by climate velocity. Nature 507, 492–495.

Busch, J., Ferretti-Ga llon, K., Enge lmann, J., Wright, M., Austin, K.C., St olle , F., Turubanova, S., Potapov, P.V., et al., 2015. Reductions in e missions fro m deforestation fro m Indonesia's moratoriu m on new oil palm, timber, and logging concessions. PNAS 112, 1328–1333.

Bytnerowic z, A., Omasa, K., Paoletti, E., 2007. Integrated effects of air pollution and climate change on forests: A northern hemisphere perspective. Environmental Pollution 147, 438–445.

Ca llaghan, T.V., Johansson, M., Key, J., Prowse, T., Ananicheva, M., Klepikov, A., 2011. Feedbacks and interactions: From the Arctic cryosphere to the climate system. Ambio 40, 75–86.

Caon, L., Valle jo, V.R., Coen, R.J., Ge issen, V., 2014. Effects of wildfire on soil nutrients in Mediterranean ecosystems. Earth-Science Reviews 139, 47–58.

Ca mpos-Arceiz, A., Bla ke, S, 2011. Megagardeners of the forest – The role of elephants in seed dispersal. Acta Oecologica 37, 542–553.

Card inale, B.J., Duffy, J.E., Gon za lez, A., Hooper, D.U., Pe rrings, C., Venail, P., Na rwan i, A., Mace, G.M., et a l., 2012. Biodiversity loss and its impact on humanity. Nature 486, 59–67.

Carswe ll, C., 2014. Ba rk beetles have devastated western forests, but that may not mean mo re severe fires. Sc ience 346, 154–156.

Chaplin-Kra mer, R., Ra mler, I., Sharp, R., Haddad, N.M., Gerber, J.S., West, P.C., Mandle, L., Engstro m, P., et al., 2015. Degradation in carbon stocks near tropical forest edges. Nature Communications 6, doi: 10.1038/ncomms10158.

Chazdon, R.L., Branca lion, P.H.S., Laestadius, L., Bennett -Curry, A., Buckingham, K., Ku mar, C., Mo ll-Rocek, J., Vie ira , I.C.G., et al., 2016. When is a forest a forest? Forest concepts and definitions in the era of forest and landscape restoration. Ambio 45, 538–550.

Chen, I.C., Hill, J.K., Ohle mü ller, R., Roy, D.B., Tho mas, C.D., 2011. Rapid range shifts of species associated with high levels of climate warming. Science 333, 1024–1026.

Chen, D., Loboda, T.V., He, T., Zhang, Y., Liang, S., 2018. Strong cooling induced by stand -replacing fires through albedo in Siberian larch forests. Scientific Reports, doi: 10.1038/s41598-018-23253-1.

Claeys, M., Graha m, B., Vas, G., Wang, W., Vermeylen, R., Pashynska, V., Cafmeyer, J., Guyon, P., et a l., 2004.

Formation of secondary organic aerosols through photooxidation of isoprene. Science 303, 1173–1176.

Cleland, E.E., Chuine, I., Men zel, A., Mooney, H.A., Schwa rtz, M.D., 2007. Sh ift ing plant phenology in response to global change. Trends in Ecology and Evolution 22, 357–365.

Climate Focus, 2015. Progress on the New York Dec laration on Forests – An Assessment Fra me work and Init ial Report. Prepared by Climate Focus (in collaboration with Environ mental De fense Fund, Forest Trends, The Globa l Alliance for Clean Cookstoves, and The Global Canopy Program),Amsterdam, Netherlands.

Clow, D.W., Rhoades, C., Briggs, J., Ca ldwe ll, M., Lewis Jr., W.M., 2011. Responses of soil and water che mistry to mountain pine beetle induced tree mortality in Grand County, Colorado, USA. Applied Geochemistry 26, 174–178.

Cochrane, M.A., 2003. Fire science for rainforests. Nature 421, 913–919.

Collins, B.J., Rhoades, C.C., Hubbard, R.M., Battaglia, M .A., 2011. Tree regeneration and future stand development after bark beetle infestation and harvesting in Colorado lodgepole pine stands. Forest Ecology and Management 261, 2168– 2175.

Colwe ll, R.K., Breh m, G., Cardelús, C.L., Gilman, A.C., Longino, J.T., 2008. Global warming, e levational range shifts, and lowland biotic attrition in the wet tropics. Science 322, 258–261.

Cooper, O.R., Parrish, D.D., Zie mke , J., Ba lashov, N.V., Cupeiro, M., Ga lbally, I.E., Gi lge, S., Horo wit z, L., et a l., 2014. Global distribution and trends of tropospheric ozone: An observation -based review. Ele menta: Science of the Anthropocene, doi: 10.12952/ journal.e le menta.000029.

Corlett, R.T., 2013. The shifted baseline : Preh istoric defaunation in the tropics and its consequences for biodiversity conservation. Biological Conservation 163, 13–21.

Da mschen, E.I., Haddad, N.M., Orroc k, J.L., Te wksbury, J.J., Levey, D.J., 2006. Corridors increase plant species richness at large scales. Science 313, 1284–1286.

Dantas de Paula, M., Groeneveld, J., Huth, A., 2015. Tropica l forest degradation and recovery in frag mented landscapes – Simulating changes in tree co mmunity, forest hydrology and carbon balance. Global Ecology and Conservation 3, 664–677.

Dantas de Paula, M., Groeneveld, J., Huth, A., 2016. The e xtent of edge effects in frag mented landscapes: Insights from satellite measurements of tree cover. Ecological Indicators 69, 196–205.

Davin, E.L., Noblet-Ducoudré, N., 2010. Climat ic impact of g lobal-scale deforestation: Rad iative versus nonradiative processes. Journal of Climate 23, 97–112.

Dawson, T.P., Jackson, S.T., House, J.I., Prentice, I.C., Mace, G.M., 2011. Beyond predictions: Biodiversity conservation in a changing climate. Science 332, 53–58.

Del Grosso, S., Parton, W., Stohlgren, T., Zheng, D., Bachelet, D., Prince, S., Hibbard, K., Olson, R., 2008. Global potential net primary production predicted from vegetation class, precipitation, and temperature. Ecology 89, 2117–2126.

Delgado-Baquerizo, M., Maestre, F.T., Galla rdo, A., Bo wke r, M.A., Wallenstein, M.D., Quero, J.L., Ochoa, V., Gozalo, B., et al., 2013. Decoupling of soil nutrient cycles as a function of aridity in global drylands. Nature 502, 672–676.

Diffenbaugh, N.S., Fie ld, C.B., 2013. Changes in ecologically c rit ical te rrestria l climate conditions. Science 341, 486–492.

Dirzo, R., Young, H.S., Ga letti, M., Ceba llos, G., Isaac, N.J.B., Collen, B., 2014. Defaunation in the Anthropocene.

Science 345, 401–406.

Doughty, C.E., Gou lden, M.L., 2008. A re tropica l forests near a high temperature threshold?. Journal of Geophysical Research 113, doi:10.1029/2007JG000632.

Doughty, C.E., Flanner, M.G., Goulden, M.L., 2010. Effect of smo ke on subcanopy shaded light, canopy temperature, and carbon dio xide uptake in an A ma zon ra inforest. Globa l Biogeochemical Cycles 24, doi:10.1029/2009GB003670.

Dwyer, G., Dushoff, J., Elkinton, J.S., Levin, S.A., 2000. Pathogen -driven outbreaks in forest defoliators revisited: Building models from experimental data. The American Naturalist 156, 105–120.

Ea rly, R., Brad ley, B.A., Dukes, J.S., La wler, J.J., Olden, J.D., Blu menthal, D.M., Gonza le z, P., Grosholz, E.D., et al., 2016. Globa l threats from invasive alien species in the twenty -first century and national response capacities. Nature Communications 7, doi: 10.1038/ncomms12485.

Edburg, S.L., Hicke , J.A., Lawrence, D.M., Thornton, P.E., 2011. Simulat ing coupled carbon and nitrogen dynamics following mountain pine beetle outbreaks in the western United States. Journal of Geophysical Research 116, doi:10.1029/2011JG001786.

Elsen, P.R., Monahan, W.B., Meren lender, A.M., 2018. Global patterns of protection of elevational gradients in mountain ranges. PNAS, doi.org/10.1073/pnas.1720141115.

Elser, J.J., Ha milton, A, 2007. Stoich io metry and the new Biology: The future is now. PLo S. Biol. 5, e 181, doi:10.1371/ journal.pbio.0050181.

Estes, J.A., Te rborgh, J., Brashares, J.S., Po wer, M.E., Berger, J., Bond, W.J., Carpenter, S.R., Essington, T.E., et a l., 2011. Trophic downgrading of Planet Earth. Science 333, 301–306.

Estrada, A., Garber, P.A., Ry lands, A.B., Roos, C., Fernandez-Duque, E., Di Fiore, A., Ne karis, K.A.I., Nijman, V., et al., 2017. Impending extinction crisis of the world's primates: Why primates matter. Science Advances3, e1600946.

FAO, 2000. Co mparison of forest area and forest area change estimates derived fro m FRA 1990 and FRA 2000.

Forest Resources Assessment, UN Food and Agriculture Organization, Rome, Italy.

FAO, 2010. Global forest resources assessment 2010 – Main report. UN Food and Agriculture Organizat ion, Ro me,

Italy.

FAO, 2012. Forests and Climate Change Working Paper 10 – Fo rest Management and Climate Change: a literature

review. UN Food and Agriculture Organization, Rome, Italy.

FAO, 2013. Gu ide for Country Report ing for FRA 2015. FRA 2015 Working Paper. UN Food and Agriculture Organization, Rome,Italy.

FAO, 2014a. Eme rging approaches to forest reference emission levels and/or forest reference levels for REDD+. UN Food and Agriculture Organization, Rome, Italy.

FAO, 2014b. Globa l Plan of Action for the Conservation, Sustainable Use and Develop ment of Forest Genetic Resources. Co mmission on Genetic Resources for Food and Agriculture, UN Food and Agricu lture Organization, Ro me , Italy.

FAO, 2015. Global Forest Resources Assessment 2015. UN Food and Agriculture Organization, Rome, Italy.

Fares, S., Va rgas, R., Detto, M., Go ldstein, A.H., Ka rlik, J., Paoletti, E., Vita le, M., 2013. Tropospheric ozone reduces carbon assimilat ion in trees: estimates fro m ana lysis of continuous flu x measurements. Global Change Biology 19, 2427–2443.

Fares, S., 2015. Five steps for managing Europe's forests. Nature 519, 407–409.

Fauset, S., Ba ker, T.R., Lewis, S.L., Feldpausch, T.R., Affu m-Ba ffoe, K., Foli, E.G., Ha me r, K.C., Swaine, M.D., 2012. Drought-induced shifts in the floristic and functional co mposition of tropica l forests in Ghana. Ecology Letters 15, 1120–1129.

Fawcett, A.A., Iyer, G.C., Cla rke , L.E., Ed monds, J.A., Hult man, N.E., McJeon, H.C., Rogelj, J., Schule r, R., et al., 2015. Can Paris pledges avert severe climate change?. Science 350, 1168–1169.

Fazhu, Z., Jiao, S., Chengjie, R., Di, K., Jian, D., Xinhui, H., Ga ihe, Y., Yongzhong, F., et al., 2015. Land use change influences soil C, N, and P stoichio metry under 'Gra in-to-Green Progra m' in China. Sc ientific Reports, doi: 10.1038/srep10195.

Ferra z, G., Russell, G.J., Stouffe r, P.C., Bierregaard Jr., R.O., Pimm, S.L., Lovejoy, T.E., 2003. Rates of species loss from Amazonian forest fragments. PNAS 100, 14069–14073.

Fischer, R., Ha rgita, Y., Günter, S., 2016. Insights fro m the ground level? A content analysis review of mult i- national REDD+ studies since 2010. Forest Policy and Economics 66, 47–58.

Fisichelli, N.A., Abella, S.R., Peters, M., Krist Jr., F.J., 2014a. Climate, trees, pests, and weeds: Change, uncertainty, and biotic stressors in eastern U.S. national park forests. Forest Ecology and Management 327, 31–39.

Fisichelli, N.A., Fre lich, L.E., Re ich, P.B., 2014b. Te mperate t ree e xpansion into adjacent borea l forest patches facilitated by warmer temperatures. Ecography 37, 152–161.

Flannigan, M., Cantin, A.S., de Groot, W.J., Wotton, M., Newbery, A., Go wman, L.M., 2013. Globa l wildland fire season severity in the 21st century. Forest Ecology and Management 294, 54–61.

Galetti, M., Pires, A.S., Brancalion, P.H.S., Fernandez, F.A.S., 2017. Reversing defaunation by trophic rewild ing in empty forests. Biotropica 49, 5–8.

Gang, C., Zhou, W., Li, J., Chen, Y., Mu, S., Ren, J., Chen, J.,Gro isman, P.Y., 2013. Assessing the spatiotemporal variation in distribution, e xtent and NPP of terrestria l ecosystems in response to climate change fro m 1911 to 2000. PLo S ONE doi:10.1371/journal.pone.0080394.

Gattuso, J.P., Magnan, A., Billé, R., Cheung, W.W.L., Howes, E.L., Joos, F., Alle mand, D., Bopp, L., et al., 2015. Contrasting futures for ocean and society from d iffe rent anthropogenic CO2 e missions scenarios. Science, doi: 10.1126/science.aac4722.

Gauthier, S., Bernie r, P., Kuuluvainen, T., Shvidenko, A.Z., Schepaschenko, D.G., 2015. Borea l forest health and global change. Science 349, 819–822.

Gh imire, B., Willia ms, C.A., Collat z, G.J., Vanderhoof, M., Rogan, J., Ku lako wski, D., Masek, J.G., 2015. La rge carbon release legacy from bark beetle outbreaks across Western United States. Global Change Biology, doi: 10.1111/gcb.12933.

Gia m, X., Scheffe rs, B.R., Sodhi, N.S., W ilcove, D.S., Ceballos, G., Ehrlich, P.R., 2012. Reservoirs of richness: least disturbed tropical forests are centres of undescribed species diversity. Proc. R. Soc. B 279, 67–76.

Gibson, L., Lynam, A.J., Bradshaw, C.J.A., He, F., Bickford, D.P., Woodruff, D.S., Bu mrungsri, S., Laurance, W.F., 2013. Near-co mplete e xt inction of native sma ll ma mmal fauna 25 years after forest frag mentation. Science 341, 1508– 1510.

Gilbert-Norton, L., Wilson, R., Stevens, J.R., Beard, K.H., 2010. A meta-analytic revie w o f corridor e ffectiveness.

Conservation Biology 24, 660–668.

Haddad, N.M ., Brudvig, L.A., Clobert, J., Davies, K.F., Gon zale z, A., Holt, R.D., Lovejoy, T.E., Se xton, J.O., et a l., 2015. Habitat fragmentation and its lasting impact on Earth's ecosystems. Science Advances 1, e1500052.

Hanewin kel, M ., Cullmann, D.A., Schelhaas, M.J., Nabuurs, G.J., Zimmermann, N.E., 2013. Climate change may cause severe loss in the economic value of European forest land. Nature Climate Change 3, 203–207.

Hansen, J., Sato, M., Ruedy, R., Lo, K., Lea, D.W., Medina-Elizade, M., 2006. Global te mperature change. PNAS 103, 14288–14293.

Hansen, M.C., Potapov, P.V., Moore, R., Hancher, M., Turubanova, S.A., Tyukav ina, A., Thau, D., Stehman, S.V., et al., 2013. High-resolution global maps of 21st-century forest cover change. Science 342, 850–853.

Hansen, W.D., Chapin III, F.S., Naughton, H.T, Rupp, T.S., Verbyla , D., 2016. Forest -landscape structure mediates effects of a spruce bark beetle (Dendroctonus rufipennis) outbreak on subsequent like lihood of burning in Alaskan boreal forest. Forest Ecology and Management 369, 38–46.

Harris, N.L., Brown, S., Hagen, S.C., Saatchi, S.S., Petrova, S., Salas, W., Hansen, M.C., Potapov, P.V., et a l., 2012.

Baseline map of carbon emissions from deforestation in tropical regions. Science 336, 1573–1576.

Harrison, R.D., Tan, S., Plotkin, J.B., Slik, F., Detto, M., Brenes, T., Itoh, A., Dav ies, S.J., 2013. Consequences of defaunation for a tropical tree community. Ecology Letters 16, 687–694.

Harsch, M.A., Hulme, P.E., Mc Glone, M.S., Duncan, R.P., 2009. Are tree lines advancing? A global meta -analysis of treeline response to climate warming. Ecology Letters 12, 1040–1049.

Hic ke, J.A., Allen, C.D., Desai, A.R., Diet ze, M.C., Hall, R.J., Hogg, E.H., Kashian, D.M ., Moore, D., et a l., 2012.

Effects of biotic disturbances on forest carbon cycling in the United States and Canada. Global Change Biology 18, 7 –34.

Hiet z, P., Turne r, B.L., Wanek, W., Richter, A., Nock, C.A., Wright, J., 2011. Long -term change in the nitrogen cycle of tropical forests. Science 334, 664–666.

Hijmans, R.J., Ca meron, S.E., Pa rra , J.L., Jones, P.G., Jarvis, A., 2005. Very high resolution interpolated climate surfaces for global land areas. Int. J. Climatol. 25, 1965–1978.

Hilker, T., Lyapustin, A.I., Tuc ker, C.J., Ha ll, F.G., Myneni, R.B., Wang, Y., Bi, J., Mendes de Moura, Y., et a l., 2014. Vegetation dynamics and rainfall sensitivity of the Amazon. PNAS 111, 16041–16046.

Hoegh-Guldberg, O., Hughes, L., Mc Intyre, S., Linden mayer, D.B., Pa rmes an, C., Possingham, H.P., Tho mas, C.D., 2008. Assisted colonization and rapid climate change. Science 321, 345–346.

Hoffmann, W.A., Ge iger, E.L., Gotsch, S.G., Rossatto, D.R., Silva, L.C.R., Lau, O.L., Haridasan, M., Franco, A.C., 2012. Eco logical thresholds at the savanna-forest boundary: how plant traits, resources and fire govern the distribution of tropical biomes. Ecology Letters 15, 759–768.

Högberg, P., Näsholm, T., Fran klin, O., Högberg, M.N., 2017. Ta mm Revie w: On the nature of the nitroge n limitat ion to plant growth in Fennoscandian boreal forests. Forest Eco logy and Management, doi:10.1016/ j.foreco.2017.04.045.

Holmes, C.D., 2014. Air pollution and forest water use. Nature 507, doi:10.1038/nature13113.

Hooper, D.U., Adair, E.C., Card inale, B.J., Byrnes, J.E.K., Hungate, B.A., Matulich, K.L., Gonza le z, A., Duffy, J.E., et al., 2012. A global synthesis reveals biodiversity loss as a major driver of ecosystemchange. Nature 486, 105–108.

Hoshika, Y., Katata, G., Deushi, M., Watanabe, M., Koike, T., Paoletti, E., 2015. Ozone-induced stomatal sluggishness changes carbon and water balance of temperate deciduous forests. Scientific Reports, doi: 10.1038/srep09871.

Houghton, R.A., Byers, B., Nassikas, A.A., 2015. A role for tropica l forest s in stabilizing atmospheric CO2. Nature Climate Change 5, 1022–1023.

Hu, J., Angeli, S., Schuetz, S., Luo, Y., Ha je k, A.E., 2009. Ecology and manage ment of e xotic and endemic Asian longhorned beetle Anoplophora glabripennis. Agricultural and Forest Entomology 11, 359–375.

Huang, W.J., Zhou, G.Y., Liu, J.X., 2012. Nitrogen and phosphorus status and their influence on aboveground production under increasing nitrogen deposition in three successional forests. Acta Oecologica 44, 20–27.

Huang, J., Yu , H., Guan, X., Wang, G., Guo, R., 2016. Accelerated dry land e xpansion under climate change. Nature Climate Change 6, 166–172.

Ibáñez, I., Kat z, D.S.W., Pelt ier, D., Wolf, S.M ., Connor Barrie , B.T., 2014. Assessing the integrated effects of landscape fragmentation on plants and plant communities: the challenge of mu ltiprocess –multiresponse dynamics. Journal of Ecology 102, 882–895.

Ibarra, J.T., Martin, K., 2015. Biotic ho mogenization: Loss of avian functional richness and habitat specialists in disturbed Andean temperate forests. Biological Conservation 192, 418–427.

IPCC, 2007. Forestry. In Climate Change 2007: Mit igation. Contribution of Working Group III to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Ca mbridge, UK.

IPCC, 2013. Su mmary for Policy make rs. Climate Change 2013: The Physical Sc ience Basis Contribution of Working Group I to the Fifth Assessment Report of the Intergovern mental Pane l on Climate Change. Ca mbridge University Press, Cambridge, UK.

IPCC, 2014. Te rrestria l and inland water systems. In: Climate Change 2014: Impacts, Adaptation, and Vu lnerability. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, UK.

IUFRO, 2008. International Un ion of Forest Research Organizat ions – Adaptations of forests to climate change: A mu ltid isciplinary review. Occasional paper 21, ISSN 1024 -414X (accessed on http://www.iufro.org/publications/series/occasional-papers/).

Iverson, L.R., Mc Kenzie, D., 2013. Tree-species range shifts in a changing climate: detecting, modeling, assisting.

Landscape Ecology 28, 879–889.

Jackson, R.B., Randerson, J.T., Canadell, J.G., Anderson, R.G., Avissar, R., Ba ldocchi, D.D., Bonan, G.B., Calde ira, K., et al., 2008. Protecting climate with forests. Environ. Res. Lett. 3, doi:10.1088/1748-9326/3/ 4/044006.

Jactel, H., Pet it, J., Desprez-Loustau, M.L., Delzon, S., Piou, D., Battisti, A., Koricheva, J., 2012. Drought effects on damage by forest insects and pathogens: a meta-analysis. Global Change Biology 18, 267–276.

Janssens, I.A., Luyssaert, S., 2009. Nitrogen's carbon bonus. Nature Geoscience 2, 318–319.

Janssens, I.A., Die le man, W., Luyssaert, S., Subke, J.A., Re ichstein, M., Ceule mans, R., Ciais, P., Dolman, A.J., et al., 2010. Reduction of forest soil respiration in response to nitrogen deposition. Nature Geoscience 3, 315–322.

Jeong, S.J., Ho, C.H., Gim, H.J., Brown, M .E., 2011. Phenology shifts at start vs. end of growing season in temperate vegetation over the Northern Hemisphere for the period 1982–2008. Global Change Biology 17, 2385–2399.

Jiméne z-Muñoz, J.C., Mattar, C., Barichivich, J., Santa maría-Artigas, A., Ta kahashi, K., Ma lhi, Y., Sobrino, J.A., van der Schrier, G., 2016. Record-breaking warming and e xt re me drought in the A mazon ra inforest during the course of El Niño 2015–2016. Scientific Reports 6, doi: 10.1038/srep33130.

Joly, K., Jandt, R.R., Klein, D.R., 2009. Decrease of lichens in Arctic ecosystems: the role of wildfire , caribou, reindeer, competition and climate in north-western Alaska. Polar Research 28, 433–442.

Jolly, W.M., Cochrane, M .A., Freeborn, P.H., Ho lden, Z.A., Brown, T.J., W illia mson, G.J., Bowman, D.M .J., 2015. Climate-induced variations in global wildfire danger fro m 1979 to 2013. Nature Co mmun ications , doi.org/10.1038/ncomms8537.

Jonard, M., Fürst, A., Verstraeten, A., Thimonie r, A., Timme rmann, V., Potočić, N., Waldner, P ., Benha m, S., et a l., 2015. Tree mineral nutrition is deteriorating in Europe. Global Change Biology 21, 418–430.

Joppa, L.N., Visconti, P., Jen kins, C.N., Pimm, S.L., 2013. Achieving the Convention on Bio logical Diversity's goals for plant conservation. Science 341, 1100–1103.

Jump, A.S., Mátyás, C., Peñuelas, J., 2009. The alt itude-for-latitude disparity in the range retractions of woody species. Trends in Ecology & Evolution 24, 694–701.

Keith, D.W., Weisenstein, D.K., Dyke ma , J.A., Keutsch, F.N., 2016. Stratospheric solar geoengineering without ozone loss. PNAS 113, 14910–14914.

Keenan, T.F., Gray, J., Friedl, M.A., Toomey, M., Bohre r, G., Hollinger, D.Y., Munger, J.W., O'Kee fe, J., et al., 2014. Net ca rbon uptake has increased through warming -induced changes in tempe rate forest phenology. Nature Climate Change 4, 598–604.

Keenan, R.J., Rea ms, G.A., Achard, F., de Fre itas, J.V., Grainger, A., Lindquist, E., 2015. Dyna mics o f global forest area: Results from the FAO Global Forest Resources Assessment 2015. Forest Ecology and Management 352, 9–20.

Keppel, G., Mokany, K., Warde ll-Johnson, G.W., Phillips, B.L., Welbergen, J.A., Reside, A.E., 2015. The capacity of refugia for conservation planning under climate change. Frontiers in Ecology and the Environment 13, 106–112.

Kharuka, V.I., Ranson, K.J., Fedotova, E.V., 2007. Spatial pattern of Siberian silkmoth outbreak and taiga morta lity.

Scandinavian Journal of Forest Research 22, 531–536.

Kim, D.H., Se xton, J.O., To wnshend, J.R., 2015. Accelerated deforestation in the humid tropics fro m the 1990s to the 2000s. Geophysical Research Letters 42, 3495–3501.

Kirilenko, A.P., Sedjo, R.A., 2007. Climate change impacts on forestry. PNAS 104, 19697–19702.

Kirschbaum, M., Fischlin, A., 1996. Climate change impacts on forests. In: Watson, R., Zinyowera , M.C. & Moss,

R.H. (eds.), Climate change 1995 – Impacts, adaptations and mitigation of c limate change: scientific -technical ana lysis. Contribution of Working Group II to the Second Assessment Report of the Intergovernmental Panel of Climate Change. Cambridge University Press, Cambridge,UK.

Koarashi, J., Atarashi-Andoh, M., Matsunaga, T., Sanada, Y., 2016. Forest type effects on the retention of radiocesium in organic layers of forest ecosystems affected by the Fukushima nuclea r accident. Sc ientific Reports, doi: 10.1038/srep38591.

Koch, F.H., Ye mshanov, D., Co lunga-Garc ia, M., Magarey, R.D., Smith, W.D., 2011. Potential establishment of alien-invasive forest insect species in the United States: where and how many?. Biological Invasions 13, 969–985.

Kurten, E.L., 2013. Cascading effects of contemporaneous defaunation on tropical forest co mmunities. Bio logical Conservation 163, 22–32.

Kurz, W.A., Dy mond, C.C., Stinson, G., Ra mp ley, G.J., Neilson, E.T., Caroll, A.L., Ebata, T., Sa franyik, L., 2008a.

Mountain pine beetle and forest carbon feedback to climate change. Nature 452, 987–990.

Kurz, W.A., Stinson, G., Ra mp ley, G.J., Dy mond, C.C., Ne ilson, E.T., 2008b. Risk of natural disturbances makes future contribution of Canada's forests to the global carbon cycle highly uncertain. PNAS 105, 1551–1555.

Land mann, G., Dreyer, E. (Eds.), 2006. Impacts of drought and heat on forest. Synthesis of availab le kno wledge, with emphasis on the 2003 event in Europe. Annals of Forest Science 63, 567–652.

Landry, J.S., Pa rrott, L., Price, D.T., Ra man kutty, N., Matthews, H.D., 2016. Modelling long -term impacts of mountain p ine beetle outbreaks on me rchantable bio mass, ecosystem carbon, a lbedo, and radiat ive forcing. Biogeosciences 13, 5277–5295.

Langley, J.A., Megonigal, J.P., 2010. Ecosystem response to elevated CO2 levels limited by nitrogen-induced plant species shift. Nature 466, 96–99.

Laurance, W.F., 2002. Hyperdynamism in fragmented habitats. Journal of Vegetation Science 13, 595–602.

Laurance, W.F., Olive ira, A.A., Laurance, S.G., Condit, R., Nascimento, H.E.M., Sanchez -Thorin, A.C., Love joy, T.E., Andrade, A., et al., 2004. Pervasive alteration of tree co mmunities in undisturbed Ama zonian forests. Nature 428, 171–175.

Laurance, W.F., Nascimento, H.E.M ., Laurance, S.G., Andrade, A., Ewe rs, R.M., Harms, K.E., Lu izão, R.C.C., Ribeiro, J.E., 2007. Hab itat frag mentation, variable edge effects, and the landscape -divergence hypothesis. PLoS ONE 10, e1017.

Laurance, W.F., Goose m, M ., Laurance, S.G., 2009. Impacts of roads and linear clea rings on tropical forests. Trends Ecol. Evol. 24, 659–669.

Laurance, W.F., Ca margo, J.L.C., Lu izão, R.C.C., Laurance, S.G., Pimm, S.L., Bruna, E.M., Stouffer, P.C., Willia mson, G.B., et a l., 2011. The fate of A ma zonian forest frag ments: A 32-year investigation. Bio logical Conservation 144, 56–67.

Laurance, W.F., Cle ments, G.R., Sloan, S., O'Connell, C.S., Mue lle r, N.D., Goosem, M., Venter, O., Ed wards, D.P., et al., 2014. A global strategy for road building. Nature 513, 229–232.

Le Quéré, C., Raupach, M.R., Canadell, J.P., Marland, G., Bopp, L., Ciais, P., Conway, T.J., Doney, S.C., et al., 2009. Trends in the sources and sinks of carbon dioxide. Nature Geoscience, doi: 10.1038/ngeo689.

Le Quéré, C., Andrew, R.M., Canadell, J.G., Sitch, S., Korsbakken, J.I., Peters, G.P., Manning, A.C., Boden, T.A., et al., 2016. Global Carbon Budget 2016. Earth Syst. Sci. Data 8, 605–649.

Lee, X., Goulden, M.L., Hollinger, D.Y., Barr, A., Blac k, T.A., Bohrer, G., Bracho, R., Dra ke, B., et al., 2011.

Observed increase in local cooling effect of deforestation at higher latitudes. Nature 479, 384–387.

Lenoir, J., Gégout, J.C., Marquet, P.A., de Ruffray, P ., Brisse, H., 2008.A significant upward shift in plant species optimum elevation during the 20th century. Science 320, 1768–1771.

Lenton, T.M., He ld, H., Kriegler, E., Hall, J.W., Lucht, W., Rah mstorf, S., Schellnhuber, H.J., 2008 . Tipping elements in the Earth's climate system. PNAS 105, 1786–1793.

Lenton, T.M., 2012. Arctic climate tipping points. Ambio 41, 10–22.

Le wis, S.L., Brando, P.M., Phillips, O.L., van der He ijden, G.M.F.,Nepstad, D., 2011.The 2010 A ma zon d rought.

Science 331, 554.

Le wis, S.L., Edwa rds, D.P., Ga lbraith, D., 2015. Increasing human do minance of t ropical forests. Science 349, 827–

832.

Linderholm, H.W., 2006. Growing season changes in the last century. Agricultural and Forest Meteorology 137, 1–

14.

Lindner, M., Maroschek, M., Netherer, S., Kre mer, A., Barbati, A., Garc ia-Gonza lo, J., Se idl, R., Delzon, S., et a l.,

2010. Climate change impacts, adaptive capacity, and vulnerability of European forest ecosystems. Forest Ecology and Management 259, 698–709.

Lloyd, J., Farquhar, G.D., 2008. Effects of rising temperatures and [CO2] on the physiology of tropical forest trees.Phil. Trans. R. Soc. B 363, 1811–1817.

Loarie, S.R., Duffy, P.B., Ha milton, H., Asner, G.P., Fie ld, C.B., Acke rly, D.D., 2009. The velocity of c limate change. Nature 462, 1052–1055.

Lobell, D.B., Sch lenker, W., Costa-Roberts, J., 2011. Climate trends and global c rop production since 1980. Sc ience 333, 616–620.

Lôbo, D., Leão, T., Me lo, F.P.L., Santos, A.M.M., Tabarelli, M ., 2011. Forest fragmentation drives Atlantic forest of northeastern Brazil to biotic homogenization. Diversity and Distributions 17, 287–296.

Logan, J.A., Régniè re, J., Powe ll, J.A., 2003. Assessing the impacts of global warming on forest pest dynamics.

Front. Ecol. Environ. 1, 130–137.

Lo mbardozzi, D., Lev is, S., Bonan, G., Sparks, J.P., 2012. Predict ing photosynthesis and transpiration responses to ozone: decoupling modeled photosynthesis and stomatal conductance. Biogeosciences Discuss. 9, 4245–4283.

Lo mbardozzi, D., Levis, S., Bonan, G., Hess, P.G., Sparks, J.P., 2015. The influence of chronic ozone e xposure on global carbon and water cycles. J. Clim. 28, 292–305.

Loranty, M.M., Be rner, L.T., Goetz, S.J., Jin, Y., Randerson, J.T., 2014. Vegetation controls on northern high latitude snow-albedo feedback: observations and CMIP5 model simulations. Global Change Biology 20, 594–606.

Lovett, G.M., Tea r, T.H., Eve rs, D.C., Findlay, S.E.G., Cosby, B.J., Dunscomb, J.K., Driscoll, C.T.,Weathers, K.C., 2009. Effects of air pollution on ecosystems and biological d iversity in the eastern United States. Ann. N.Y. Acad. Sci. 1162: 99–135.

Lu, X., Mao, Q., Gillia m, F.S., Luo, Y., Mo, J., 2014. Nitrogen deposition contributes to soil ac idificat ion in t ropical ecosystems. Global Change Biology 20, 3790–3801.

Luo, Y., Su, B., Cu rrie, W.S., Du kes, J.S., Finzi, A., Hartwig, U., Hungate, B., McMurtrie, R.E., et al., 2004.

Progressive nitrogen limitation of ecosystemresponses to rising atmospheric carbon dioxide. BioScience 54, 731–739.

Luyssaert, S., Schulze, E.D., Bö rner, A., Knohl, A., Hessenmölle r, D., Law, B.E., Ciais, P., Grace, J., 2008. Old – growth forests as global carbon sinks. Nature 455, 213–215.

MacDic ken, K.G., 2015. Global Forest Resources Assessment 2015: What, why and how?. Forest Ecology and Management 352, 3–8.

MacDonald, G.M., Kre menetski, K.V., Be ilman, D.W., 2008. Climate change and the northern Russian treeline zone. Phil. Trans. R. Soc. B 363, 2285–2299.

Macias-Fauria, M., Forbes, B.C., Zetterberg, P., Ku mpula , T., 2012. Eurasian Arctic g reening reveals teleconnections and the potential for structurally novel ecosystems. Nature Climate Change 2, 613–618.

Magnani, F., Mencuccini, M., Borghetti, M., Berbig ier, P., Bern inger, F., De lzon, S., Gre lle, A., Hari, P., et a l., 2007. The human footprint in the carbon cycle of temperate and boreal forests. Nature 447, 848–850.

Maisels, F., Strindberg, S., Blake , S., W ittemyer, G., Ha rt, J., W illia mson, E.A., Aba'a, R., Abitsi, G., et al., 2013.

Devastating decline of forest elephants in Central Africa. PLoS ONE 8, e59469.

Malhi, Y., Aragão, L.E.O., Ga lbraith, D., Huntingford, C., Fisher, R., Zela zo wski, P., Sitch, S., Mc Sweeney, C., et al., 2009. Exp loring the like lihood and mechanism of a c limate -change-induced dieback of the Ama zon ra inforest. PNAS 106, 20610–20615.

Malle z, S., Castagnone, C., Espada, M., Vie ira, P., Eisenback, J.D., Harrell, M ., Mota, M., Aikawa , T., et a l., 2015. Worldwide invasion routes of the pinewood nematode: What can we infe r fro m population genetics analyses?. Biologica l Invasions 17, 1199–1213.

Matyssek, R., Karnosky, D.F., Wieser, G., Percy, K., Oksanen, E., Gra ms, T.E.E., Kubiske, M., Hanke, D., et a l., 2010a. Advances in understanding ozone impact on forest trees: Messages from novel phytotron and free -a ir fu migation studies. Environmental Pollution 158, 1990–2006.

Matyssek, R., W ieser, G., Ceule mans, R., Rennenberg, H., Pretzsch, H., Haberer, K., Low, M ., Nunn, A.J., et a l., 2010b. Enhanced ozone strongly reduces carbon sink strength of adult beech (Fagus sylvatica) – Resume fro m the free -air fumigation study at Kranzberg Forest. Environmental Pollution 158, 2527–2532.

McIntyre, P.J., Thorne, J.H., Do lanc, C.R., Flint, A.L., Flint, L.E., Ke lly, M., Ackerly, D.D., 2015. Twentieth – century shifts in forest structure in Ca lifornia : Denser forests, smaller trees, and increased dominance of oaks. PNAS 112, 1458–1463.

McLachlan, J.S., Clark, J.S., Manos, P.S., 2005. Molecular indicators of tree migration capacity under rapid climate change. Ecology 86, 2088–2098.

MEA, 2005. Ecosystems and human well-being: Synthesis. Washington, DC, United States.

Meunier, C.L., Gundale, M .J., Sánchez, I.S., Liess, A., 2016. Impact of nitrogen deposition on forest and lake food webs in nitrogen-limited environments. Global Change Biology 22, 164–179.

Meyerholt, J., Zaehle , S., 2015. The role of stoichio metric fle xib ility in mode lling forest ecosy stem responses to nitrogen fertilization. New Phytologist 208, 1042–1055.

Michaelian, M., Hogg, E.H., Ha ll, R.J., Arsenault, E., 2011. Massive morta lity o f aspen following severe drought along the southern edge of the Canadian boreal forest. Global Change Biology 17, 2084–2094.

Millar, C.I., Stephenson, N.L., 2015. Te mpe rate forest health in an era of e merg ing megadisturbance. Science 349, 823–826.

Millar, R.J., Fuglestvedt, J.S., Friedlingstein, P., Rogelj, J., Grubb, M.J., Matthews, H.D., Ske ie, R.B., Forster, P.M., et al., 2017. Emission budgets and pathways consistent with limiting warming to 1.5 °C. Nature Geoscience 10, 741–747.

Minteer, B.A., Collins, J.P., 2010. Move it or lose it? The ecological ethics of relocating species under climate change. Ecological Applications 20, 1801–1804.

Molina, M., Zae lke , D., Sarma , K.M ., Andersen, S.O., Ra manathan, V., Kania ru, D., 2009. Reducing abrupt c limate change risk using the Montreal Protocol and other regulatory actions to co mple ment cuts in CO 2 e missions. PNAS 106, 20616–20621.

Mooney, H., La rigauderie , A., Cesario, M ., Elmqu ist, T., Hoegh -Gu ldberg, O., Lavorel, S., Mace, G.M., Palme r, M., et al., 2009. Biodiversity, climate change, and ecosystem services. Current Opin ion in Environ mental Sustainability 1, 46– 54.

Morris, J.L., Mc Lauchlan, K.K., Higuera, P.E., 2015. Sensitivity and co mplacency of sedimentary biogeochemical records to climate-mediated forest disturbances. Earth-Science Reviews 148, 121–133.

Moritz, M.A., Morais, M.E., Su mme rell, L.A., Carlson, J.M., Doyle, J., 2005. Wildfires, co mple xity, and highly optimized tolerance. PNAS 102, 17912–17917.

Myers-Smith, I.H., Forbes, B.C., W ilmking, M., Ha llinger, M., Lantz, T., Blok, D., Tape, K.D., Mac ias -Fauria , M., et al., 2011. Sh rub expansion in tundra ecosystems: dynamics, impacts and research priorit ies. Environ. Res. Lett. 6, doi:10.1088/1748-9326/6/4/ 045509.

Myers-Smith, I.H., Elmendorf, S.C., Bec k, P.S.A., Wilmking, M., Hallinger, M., Blok, D., Tape, K.D., Rayback, S.A., et al., 2015. Climate sensitivity of shrub growth across the tundra biome. Nature Climate Change 5, 887–891.

Naudts, K., Chen, Y., Mc Grath, M.J., Ryder, J., Va lade, A., Otto, J., Luyssaert, S., 2016. Europe's forest management did not mitigate climate warming. Science 351, 597–600.

Ne mani, R.R., Kee ling, C.D., Hashimoto, H., Jolly, W.M., Piper, S.C., Tucker, C.J., Myneni, R.B., Running, S.W., 2003. Climate-driven increases in global terrestrial net primary production from 1982 to 1999. Science 300, 1560–1563.

Nepstad, D, McGrath, D., Stic kler, C., Alencar, A., A zevedo, A., Swette, B., Be zerra, T., DiGiano, M., et al., 2014. Slowing Ama zon deforestation through public policy and interventions in beef and soy supply chains. Science 344, 1118 – 1123.

Norby, R.J., De Luc ia, E.H., Gie len, B., Ca lfapietra , C., Giardina, C.P., King, J.S., Led ford, J., McCa rthy, H.R., et al., 2005. Forest response to elevated CO2 is conserved across a broad range of productivity. PNAS 102, 18052–18056.

Nu mata, I., Cochrane, M.A., Souza Jr, C.M., Sa les, M.H., 2011. Ca rbon emissions fro m deforestation and forest fragmentation in the Brazilian Amazon. Environ. Res. Lett. 6, doi:10.1088/1748-9326/6/ 4/044003.

Olden, J.D., 2006. Biotic homogenization: a ne w research agenda for conservation biogeography. Journ al of Biogeography 33, 2027–2039.

Osuri, A.M., Ratna m, J., Varma, V., A lvarez-Loayza, P., Astaiza, J.H., Bradford, M., Fletcher, C., Ndoundou – Hocke mba, M., et a l., 2016. Contrasting effects of defaunation on aboveground carbon storage across the global tro pics. Nature Communications 7, doi:10.1038/ncomms11351.

Page, S.E., Siegert, F., Rieley, J.O., Boehm, H.D.V., Jaya, A., Limin, S., 2002. The a mount of carbon re leased fro m peat and forest fires in Indonesia during 1997. Nature 420, 61–65.

Pan, Y., Birdsey, R.A., Fang, J., Houghton, R., Kauppi, P.E., Kurz, W.A., Phillips, O.L., Shvidenko, A., et a l., 2011.

A large and pers istent carbon sink in the world's forests. Science 333, 988–993.

Pan, Y., Birdsey, R.A., Phillips, O.L., Jackson, R.B., 2013. The structure, distribution, and bio mass of the world's forests. Annu. Rev. Ecol. Evol. Syst. 44, 593–622.

Pardo, L.H., Fenn, M .E., Goodale , C.L., Geiser, L.H., Driscoll, C.T., Allen, E.B., Ba ron, J.S., Bobbink, R., et a l., 2011. Effects of nit rogen deposition and e mpirica l n itrogen critica l loads for ecoreg ions of the Un ited States. Ecologica l Applications 21, 3049–3082.

Pastor, J., Cohen, Y., Hobbs, N.T., 2006. In Large He rbivore Ecology, Ecosystem Dynamics and Conservation .

Cambridge University Press, Cambridge, UK.

Pearson, R.G., Phillips, S.J., Loranty, M.M ., Bec k, P.S.A., Da moulas, T., Kn ight, S.J., Goet z, S.J., 2013. Shifts in Arctic vegetation and associated feedbacks under climate change. Nature Climate Change 3, 673–677.

Pelt zer, D.A., A llen, R.B., Lovett, G.M., Whitehead, D., Wardle, D.A., 2010. Effects of biologica l invasions on forest carbon sequestration. Global Change Biology 16, 732–746.

Peng, C., Ma, Z., Le i, X., Zhu, Q., Chen, H., Wang, W., Liu, S., Li, W., et a l., 2011. A drought-induced pervasive increase in tree mortality across Canada's boreal forests. Nature Climate Change 1, 467–471.

Peñuelas, J., Llusià, J., 2003. BVOCs: plant defense against climate warming?. Trends in Plant Science 8, 105–109. Peñuelas, J., Rutishauser, T., Filella, I., 2009. Phenology feedbacks on climate change. Science 324, 887–888.

Peñuelas, J., Canadell, J.G., Ogaya, R., 2011. Increased water-use efficiency during the 20th century did not translate into enhanced tree growth. Global Ecol. Biogeogr. 20, 597–608.

Peñuelas, J., Sardans, J., Rivas -Ubach, A., Janssens, I.A., 2012. The human-induced imbalance between C, N and P in Earth's life system. Global Change Biology 18, 3–6.

Peñuelas, J., Poulter, B., Sardans, J., Cia is, P., van der Ve lde, M., Bopp, L., Boucher, O., Godderis, Y., et al., 2013. Hu man-induced nitrogen–phosphorus imbalances alte r natural and managed ecosystems across the globe. Nature Communications, doi: 10.1038/ncomms3934.

Peres, C.A., Emilio, T., Schietti, J., Des mouliè re, S.J.M ., Levi, T., 2016. Dispersal limitation induces long-term biomass collapse in overhunted Amazonian forests. PNAS 113, 892–897.

Pfeifer, M., Le febvre, V., Peres, C.A., Banks -Le ite, C., Wearn, O.R., Marsh, C.J., Butchart, S.H.M., Arroyo – Rodriguez, V., et al., 2017. Creation of forest edges has a global impact on forest vertebrates. Nature 551, 187–191.

Phelps, J., Webb, E.L., Agrawa l, A., 2010. Does REDD+ threaten to recentralize forest governance?. Science 328, 312–313.

Phillips, O.L., A ragão, L.E.O., Le wis, S.L., Fisher, J.B., Lloyd, J., López-Gon zále z, G., Ma lhi, Y., Monteagudo, A., et al., 2009. Drought sensitivity of the Amazon rainforest. Science 323, 1344–1347.

Piao, S., Ciais, P., Fried lingstein, P., Pey lin, P., Re ichstein, M., Luyssaert, S., Ma rgolis, H., Fang, J., et al., 2008. Net carbon dioxide losses of northern ecosystems in response to autumn warming. Nature 451, 49–52.

Pinto, S.R., Melo, F., Tabarelli, M., Padovesi, A., Mesquita, C.A., de Mattos Scaramuzza, C.A., Castro, P., Carrascosa, H., et al., 2014. Govern ing and delive ring a bio me-wide restoration in itiative : The case of Atlantic Forest Restoration Pact in Brazil. Forests 5, 2212–2229.

Post, E., Forchha mmer, M.C., Bret-Harte, M .S., Ca llaghan, T.V., Christensen, T.R., Elberling, B., Fo x, A.D., Gilg, O., et al., 2009. Ecological dynamics across the Arctic associated with recent climate change. Science 325, 1355–1358.

Poulsen, J.R., Clark, C.J., Pa lme r, T.M., 2013. Ecologica l e rosion of an Afrotropical forest a nd potential consequences for tree recruitment and forest biomass. Biological Conservation 163, 122–130.

Prăvălie, R., Sîrodoev, I., Peptenatu, D., 2014a. Changes in the forest ecosystems in areas impacted by a rid ization in South-Western Romania. Journal of Environmental Health Science and Engineering 12:2.

Prăvălie, R., Sîrodoev, I., Peptenatu, D., 2014b. Detecting climate change effects on forest ecosystems in South – Western Romania using Landsat TM NDVI data. Journal of Geographical Sciences 24, 815–832.

Prăvălie, R., 2014. Nuclea r weapons tests and environmental consequences: A global perspective. Amb io 43, 729–

744.

Prăvălie, R., 2016. Dry lands extent and environmental issues. A global approach. Earth -Science Rev iews 161, 259–

278.

Prăvălie, R., Bandoc, G., Patriche, C., To mescu, M., 2017a. Spatio -te mporal trends of mean air te mperature during

1961–2009 and impacts on crop (ma ize) yie lds in the most important agricultural region of Ro mania . Stoch. Environ. Res. Risk Assess. 31, 1923–1939.

Prăvălie, R., Patriche, C., Bandoc, G., 2017b. Quantificat ion of land degradation sensitivity areas in Southern and Central Southeastern Europe. New results based on improving DISM ED methodology with new climate data. Catena 158, 309–320.

Prăvălie, R., Bandoc, G., 2018. Nuclea r energy: Bet ween global electricity demand, wo rld wide decarbonisation imperativeness, and planetary environmental implications. Journal of Environmental Management 209, 81–92.

Pütz, S., Groeneveld, J., Henle, K., Knogge, C., Martensen, A.C., Metz, M.,Metzger, J.P., Ribe iro, M .C., et al., 2014.

Long-term carbon loss in fragmented Neotropical forests. Nature Communications 5, doi: 10.1038/ncomms6037.

Raftery, A.E., Zimmer, A., Frie rson, D.M.W., Start z, R., Liu, P., 2017. Less than 2 °C warming by 2100 unlikely.

Nature Climate Change, doi:10.1038/nclimate3352.

Randerson, J.T., Liu, H., Flanner, M.G., Cha mbers, S.D., Jin, Y., Hess, P.G., Pfister, G., Mack, M.C., et al., 2006.

The impact of boreal forest fire on climate warming. Science 314, 1130–1132.

Redford, K.H., 1992. The empty forest. Bioscience 42, 412–422.

Re is, S., Grennfelt, P., Klimont, Z., A mann, M., ApSimon, H., Hettelingh, J.P., Holland, M., Le Ga ll, A.C., et a l., 2012. From acid rain to climate change. Science 338, 1153–1154.

Richardson, A.D., Black, T.A., Ciais, P., De lbart, N., Friedl, M.A., Gobron, N., Ho llinger, D.Y., Kutsch, W.L., et a l., 2010. Influence of spring and autumn phonological transitions on forest ecosystem productivity. Ph il. Trans. R. Soc. B 365, 3227–3246.

Ripple, W.J., Ne wsome, T.M ., Wolf, C., Dirzo, R., Everatt, K.T., Galetti, M., Hayward, M.W., Kerley, G.I.H., et a l., 2015. Collapse of the world's largest herbivores. Science Advances 1, e1400103.

Rockströ m, J., Steffen, W., Noone, K., Persson, Ǻ., Chapin III, F.S., La mb in , E.F., Lenton, T.M., Scheffer, M., et a l., 2009. A safe operating space for humanity. Nature 461, 472–475.

Rockströ m, J., Ga ffney, O., Rogelj, J., Me inshausen, M., Nakicenovic, N., Schellnhuber, H.J., 2017. A roadmap for rapid decarbonization. Science 355, 1269–1271.

Rogelj, J., den Elzen, M., Höhne, N., Fransen, T., Fe kete, H., Win kle r, H., Schaeffer, R., Sha, F., et al., 2016. Paris Agreement climate proposals need a boost to keep warming well below 2 °C. Nature 534, 631–639.

Ro mijn, E., Hero ld, M., Kooistra, L., Murd iyarso, D., Verchot, L., 2012. Assessing capacities of non -Annex I countries for national forest monitoring in the context of REDD+. Environmental Science & Policy 19–20, 33–48.

Rypdal, K., Berntsen, T., Fuglestvedt, J.S., Aunan, K., Torvanger, A., Stordal, F., Pacyna, J.M., Nygaard, L.P., 2005. Tropospheric ozone and aerosols in climate agree ments: scientific and polit ical cha llenges. Environ mental Sc ience & Policy 8, 29–43.

Sardans, J., Rivas -Ubach, A., Peñuelas, J., 2012. The C:N:P stoichiometry of organis ms and ecosystems in a changing world: A review and perspectives. Perspectives in Plant Ecology, Evolution and Systematics 14, 33–47.

Schimel, D., 2010. Drylands in the Earth system. Science 327, 418–419.

Schime l, D., Stephens, B.B., Fisher, J.B., 2015.Effect of increasing CO2 on the terrestrial carbon cycle. PNAS 112, 436–441.

Schmitt, C.B., Burgess, N.D., Coad, L., Be lokurov, A., Besançon, C., Boisrobert, L., Ca mpbell, A., Fish, L., et a l., 2009. Global analysis of the protection status of the world's forests. Biological Conservation 142, 2122–2130.

Schoennagel, T., Veb len, T.T., Negron, J.F., Smith, J.M ., 2012. Effects of mountain pine beetle on fuels and expected fire behavior in lodgepole pine forests, Colorado, USA. PLoS ONE 7, doi:10.1371/journal.pone.0030002.

Schuur, E.A.G., Mc Guire, A.D., Schädel, C., Grosse, G., Harden, J.W., Hayes, D.J., Hugelius, G., Koven, C.D., et al., 2015. Climate change and the permafrost carbon feedback. Nature 520, 171–179.

Seddon, P.J., Griffiths, C.J., Soorae, P.S., Armstrong, D.P., 2014. Reversing defaunation: Restoring species in a changing world. Science 345, 406–412.

Seddon, A.W.R., Macias -Fauria, M., Long, P.R., Benz, D., Willis, K.J., 2016. Sensitiv ity of g lobal terrestrial ecosystems to climate variability. Nature 531, 229–232.

Seidl, R., Tho m, D., Kautz, M., Mart in-Benito, D., Pe ltonie mi, M., Vacchiano, G., Wild, J., Ascoli, D., et al., 2017.

Forest disturbances under climate change. Nature Climate Change 7, 395–402.

Seto, K.C., Güneralp, B., Hutyra, L.R., 2012. Global forecasts of urban expansion to 2030 and direct impacts on biodiversity and carbon pools. PNAS 109, 16083–16088.

Sherry, R.A., Zhou, X., Gu, S., Arnone III, J.A., Sch ime l, D.S., Verburg, P.S., Wallace, L.L., Luo, Y., 2007.

Divergence of reproductive phenology under climate warming. PNAS 104, 198–202.

SilviStrat, 2005. Manage ment of European forests under changing climat ic conditions, Research notes 163, University of Joensuu, Faculty of Forestry, Joensuu, Finland.

Sitch, S., Co x, P.M., Collins, W.J., Huntingford, C., 2007. Indirect radiative forc ing of climate change through ozone effects on the land-carbon sink. Nature 448, 791–794.

Sloan, S., Sayer, J.A., 2015. Forest Resources Assessment of 2015 shows positive global trends but fo rest loss and degradation persist in poor tropical countries. Forest Ecology and Management 352, 134–145.

Staver, C.A., Archibald, S., Lev in, S.A., 2011a . The global e xtent and determinants of savanna and forest as alternative biome states. Science 334, 230–232.

Staver, C.A., Archiba ld, S., Levin, S.A., 2011b. Tree cover in sub -Saharan Africa: Ra infa ll and fire constrain forest and savanna as alternative stable states. Ecology 92, 1063–1072.

Staver, C.A., Levin, S.A., 2012. Integrating theoretical climate and fire e ffects on savanna and forest systems.

American Naturalist 180, 211–224.

Steffen, W., Persson, Ǻ., Deutsch, L., Zalasie wic z, J., Willia ms, M ., Richardson, K., Cru mley, C., Crut zen, P., et a l., 2011. The Anthropocene: From global change to planetary stewardship. AMBIO 40, 739–761.

Strauss, J., Schirrmeister, L., Grosse, G., Fortie r, D., Hugelius, G., Knoblauch, C., Ro manovsky, V., Schädel, C., et al., 2017. Deep Yedo ma perma frost: A synthesis of depositional characteristics and carbon vulnerability . Earth-Science Reviews, doi: 10.1016/j.earscirev.2017.07.007.

Suding, K., Higgs, E., Pa lmer, M., Callicott, J.B., Anderson, C.B., Ba ker, M ., Gutrich, J.J., Hondula, K.L., et a l., 2015. Committing to ecological restoration. Science 348, 638–640.

Swart, R., Ma rinova, M., 2010. Po licy options in a worst case climate change world. M itig. Adapt. Strateg. Glob.

Change. 15, 531–549.

Tabarelli, M., Ca rdoso da Silva, J.M., Gascon, C., 2004. Forest frag mentation, synergisms and the impoverishment of neotropical forests. Biodiversity and Conservation 13, 1419–1425.

Tan, M., Li, X., 2015. Does the Green Great Wall effectively decrease dust storm intensity in China? A study based on NOAA NDVI and weather station data. Land Use Policy 43, 42–47.

ter Steege, H., Pit man, N.C.A., Sabatie r, D., Bara loto, C., Sa lo mão, R.P., Guevara, J.E., Phillips, O.L., Castilho, C.V., et al., 2013. Hyperdominance in the Amazonian tree flora. Science 342, doi: 10.1126/science.1243092.

Tera mage, M.T., Onda, Y., Kato, H., 2016. Sma ll scale temp ora l distribution of radiocesiu m in undisturbed coniferous forest soil: Radiocesium depth distribution profiles. Journal of Environmental Management 170, 97–104.

Thiry, Y., Colle, C., Yoschenko, V., Levchuk, S., Van Hees, M., Hurtevent, P., Kashparov, V., 2009. Impact of Scots pine (Pinus sylvestris L.) p lantings on long term 137Cs and 90Sr recycling fro m a waste burial site in the Chernobyl Red Forest. Journal of Environmental Radioactivity 100, 1062–1068.

Thuiller, W., Lavergne, S., Roquet, C., Boulangeat, I., La fourcade, B., Araujo, M.B., 2011. Consequences of climate change on the tree of life in Europe. Nature 470, 531–534.

Trumbore, S., Brando, P., Hartmann, H., 2015. Forest health and global change. Science 349, 814–818. United Nations, 2016. The Sustainable Development Goals Report 2016, New York, United States.

Urban, M.C., 2015. Accelerating extinction risk from climate change. Science 348, 571–573.

USDA, 2000. The impacts of climate change on America 's forests: A technical document supporting the 2000 USDA Forest Service RPA assessment, US Depart ment of Agricu lture Forest Service, Roc ky Mountain Research Station, United States.

van der Plas, F., Manning, P., Soliveres, S., Allan, E., Scherer-Lorenzen, M., Ve rheyen, K., Wirth, C., Zavala, M.A., et al., 2016. Biotic homogenization can decrease landscape-scale forest multifunctionality. PNAS 113, 3557–3562.

van der Werf, G.R., Morton, D.C., DeFries, R.S., Olivie r, J.G.J., Kasibhatla, P.S., Jackson, R.B., Collatz, G.J., Randerson, J.T., 2009. CO2 emissions from forest loss. Nature Geoscience 2, 737–738.

Van Drunen, S.G., Schutten, K., Bowen, C., Bo land, G.J., Husband, B.C., 2017. Populat ion dynamics and the influence of b light on A merican chestnut at its northern range limit: Lessons for conservat ion. Forest Eco logy and Management 400, 375–383.

Van Mantgem, P.J., Stephenson, N.L., Byrne, J.C., Danie ls, L.D., Fran klin, J.F., Fu lé, P.Z., Ha rmon, M.E., Larson, A.J., et al., 2009. Widespread increase of tree mortality rates in the western United States . Science 323, 521–524.

Van Nes, E.H., Hirota, M., Holmg ren, M., Scheffer, M., 2014. Tipping points in tropica l tree cover: linking theory to data. Global Change Biology 20, 1016–1021.

Vaughan, N.E., Lenton, T.M., 2011. A review of climate geoengineering p roposals. Climatic Change 109, 745–790.

Vida l, M.M., Pires, M.M., Guimarães Jr., P.R., 2013. Large vertebrates as the missing components of seed -dispersal networks. Biological Conservation 163, 42–48.

Vors, L.S., Boyce, M.S., 2009. Global declines of caribou and reindeer. Global Change Biology 15, 2626–2633. Walther, G.R., 2010. Co mmunity and ecosystem responses to recent climate change. Phil. Trans. R. Soc. B 365,

2019–2024.

Walther, G.R., Roques, A., Hulme , P.E., Sykes, M.T., Pyšek, P., Kühn, I., Zobel, M ., Bacher, S., et al., 2009. Alien species in a warmer world: risks and opportunities.Trends in Ecology & Evolution 24, 686–693.

Wang, Y.P., Houlton, B.Z., 2009. Nitrogen constraints on terrestrial carbon uptake: Imp lications for the global carbon-climate feedback. Geophysical Research Letters 36, doi:10.1029/2009GL041009.

Wang, X.M., Zhang, C.X., Hasi, E., Dong, Z.B., 2010. Has the Three -North Forest Shelterbe lt Progra m solved the desertification and dust storm problems in arid and semiarid China? Journal of Arid Environments 74, 13–22.

Wang, B., Shugart, H.H., Shuman, J.K., Le rdau, M.T., 2016. Forests and ozone: productivity, carbon storage, and feedbacks. Scientific Reports, doi: 10.1038/srep22133.

Wieder, W.R., Cleveland, C.C., Smith, W.K., Todd -Brown, K., 2015. Future productivity and carbon storage limited by terrestrial nutrient availability. Nature Geoscience 8, 441–444.

Willia ms, A.P., Allen, C.D., M illa r, C.I., Swetna m, T.W., Michaelsen, J., Still, C.J., Leavitt, S.W., 2010. Forest responses to increasing aridity and warmth in the southwestern United States. PNAS 107, 21289–21294.

Wingfield, M.J., Broc kerhoff, E.G., Wingfie ld, B.D., Slippers, B., 2015. Planted forest health: The need fo r a global strategy. Science 349, 832–836.

Wittig, V.E., Ainsworth, E.A., Long, S.P., 2007. To what e xtent do current and projected increases in surface o zone affect photosynthesis and stomatal conductance of trees? A meta -ana lytic rev iew of the last 3 decades of e xperiments. Plant, Cell and Environment 30, 1150–1162.

Wittig, V.E., A insworth, E.A., Naidu, S.L., Karnosky, D.F., Long, S.P., 2009. Quantifying the impact of current and future tropospheric ozone on tree biomass, growth, physiology and biochemistry: a quantitative meta -analysis. Global Change Biology 15, 396–424.

Wuyts, B., Cha mpneys, A.R., House, J.I., 2017. A ma zonian fo rest-savanna bistability and human impact. Nature Communications 8, doi: 10.1038/ncomms15519.

Xu, L., Sa manta, A., Costa, M.H., Ganguly, S., Ne mani, R.R., Myneni, R.B., 2011. W idespread decline in greenness of Amazonian vegetation due to the 2010 drought. Geophysical Research Letters 38, doi:10.1029/2011GL046824.

Yang, X., Thornton, P.E., Ricciuto, D.M ., Post, W.M., 2014. The role o f phosphorus dynamics in tropical fo rests – a modeling study using CLM-CNP. Biogeosciences 11, 1667–1681.

Young, P.J., Archiba ld, A.T., Bowman, K.W., La ma rque, J.F., Na ik, V., Stevenson, D.S., Tilmes, S., Voulgara ki, A., et al., 2013. Pre -industrial to end 21st century projections of tropospheric ozone fro m the At mosph eric Che mistry and Climate Model Intercomparison Project (ACCMIP). Atmos. Chem. Phys. 13, 2063–2090.

Yu , Z., Loisel, J., Brosseau, D.P., Be ilman, D.W., Hunt, S.J., 2010. Globa l peatland dynamics since the Last Glac ial Maximum. Geophysical Research Letters37, doi:10.1029/2010GL043584.

Zaehle, S., Friedlingstein, P., Friend, A.D., 2010. Te rrestria l n itrogen feedbacks may accele rate future c limate change. Geophysical Research Letters 37, doi:10.1029/ 2009GL041345.

Zaehle, S.,Medlyn, B.E., De Kauwe , M.G., Wa lke r, A.P., Diet ze, M .C., Hickler, T., Luo, Y., Wang, Y.P., et a l., 2014. Eva luation of 11 terrestrial ca rbon–nitrogen cycle models against observations fro m two te mperate Free -Air CO2 Enrichment studies. New Phytol. 202, 803–822.

Zeng, N., Yoon, J.H., Marengo, J.A., Subra mania m, A., Nobre, C.A., Mariotti, A., Nee lin, J.D., 2008. Causes and impacts of the 2005 Amazon drought. Environ. Res. Lett., doi:10.1088/1748-9326/3/ 1/014002.

Zhang, T., Barry, R.G., Knowles, K., Heginbottom, J.A., Brown, J., 2008. Statistics and characteristics of permafrost and ground-ice distribution in the Northern Hemisphere. Polar Geography 31, 47–68.

Zhao, M., Running, S.W., 2010. Drought-induced reduction in global terrestria l net prima ry production fro m 2000 through 2009. Science 329, 940–943.

Zimov, S.A., Schuur, E.A.G., Chapin III, F.S., 2006. Perma frost and the global carbon budget. Science 312, 1612–

1613.

Zyryanova, O.A., Yaborov, V.T., Tch ikhacheva, T.L., Ko ike , T., Ma koto, K., Matsuura, Y., Satoh, F., Zy ryanov, V.,

2007. The structure and biodiversity after fire disturbance in Larix gmelinii (Rupr.) Rupr. forests, northeastern Asia. Eurasian J. For. Res. 10, 19–29.

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7