Forest Feology and Management 462 (2020) 117972 Journal homepage: wwe! Content lists available at ScienceDisest Forest Ecology and Management comilocaterforeco Brazil’s forest restoration, biomass and carbon stocks: A critical review of the knowledge gaps Fernando Ravanini Gardon’ “pepe of Esl, tint of isis, Une of So Pal, Ra de Mate, re Rozely Ferreira dos Santos", Ricardo Ribeiro Rodrigues” aera of ore Elo and Reeraon, Dart ef Bole! Soc, Une fio Po, oa Spar de Aria “ts te Quer A Baie Die 1, CEP 1341830, P.O Ba 9 Praca So Pa, Bras 1. Introduction Globally, tropical forests have been regarded as diversity pools and ‘carbon sinks, and its conservation plays an important role in the context ‘of climate change mitigation (IPCC, 2005; Chazdon etal. 2016; Poorter ‘al, 2016), Estimates suggest that forests store up to 80% of terrestrial carbon (Houghton, 2008), and throughout tropical regions, forest re stration has become a conservation tool to reestablish ecological processes as biomass and carbon accumulation, and the delivery of regulating ecosystem services, as reducing the effects of climate change (Chazdon, 2008; Locatelli tal. 2015; Mei etal. 2017). ‘The potential of restored forests in sequestering atmospheric carbon fs closely attained to ecological aspects, as forest type, restoration method and management, landscape context, and environmental con: ditions (Chazdon, 2014; Chazdon etal. 2016; Crouzeilles etal. 2017; “James el, 2018). Despite the particular blot and abiotic features of ‘each restoration, restoring forests by any method (planting seedlings ‘assisting natural regeneration) drives to a local increment in biomass and carbon stocks (Benini and Adeodato, 2017), Since biomass accu mulation ean be the main driver of plant community changes during tropical forest succession (Lohbeck et al. 2015) and forest restoration is ‘global strategy for climate change mitigation, ensuring the re- ‘establishment of this ecological processes could afford both restoration success and vital regulating ecosystem services. In Brazil, country which holds the second largest amount of natural forest inthe globe (SFB, 2013), many restoration projects have been implemented due to the legal obligation imposed by the Forest Code (Cesderal Lav, 12.651), and this activity beeame a potential agribusiness productive chain instead of just a visionary and romantic concern Genini and Adeodato, 2017). Ths, a robust scientific knowledge re- lated to biomass accumulation in forest restoration is expected to sup Port robust evaluations of the real contribution of these systems to carbon ecosystem services (Van der Gaast etal. 2018) In this context, we aimed to conduct a systematic review in the field ‘of aboveground biomass recovery and eatbon sequestration by forest, restorations in Brazil. We highlight the limitations of the accumulated "Corresponding authors. theoreti methodological slentifc knowledge for generalizations ofthe biomass stock in restored forests throughout the Brazilian territory 2. Methods 2.1, Studies searches and selection We conducted a systematic and integrative review to search for biomass or carbon evaluations conducted in Brazilian Passive Restoration (PR) and Active Restoration (AR) sites. A systematic que litative and quantitative literature review is a robust and replicable ‘method, based on a comprehensive and detailed search strategy, able to Identify boundaries around generalizations and important geographic, scalar, theoretical and methodological gaps in the literature (Pickering and Byrne, 2014; Torraco, 2016). We compiled information related to {he geographic distribution of the studies, che methods used to estimate [AGB/carbon stocks, existing bias in the researches, the main factors Influencing AGB/earbon recovery, and trends and limitations of the researches carried out in Brazil. We selected only scientific studies available in indexed bibliographic databases and published in English or Portuguese ‘The searches were performed in Scopus and Web of Science data bases for all years available, with the last view on October 2019. We selected the search terms to account for stdies that evaluated AGEY/ carbon stocks in restoration sites implemented by active (planting seedlings and seeding) and passive (natural regeneration and asisted regeneration) restoration methods in Brazil. The selected texmis were forese* (ecosystem), regrowth OR regener* OR restr OR recover* OR refores™ OR reveget® (processes involved in the ecosystem recovery), Biomass OR carbon (targeted ecological process), and Brasil (studies geographic distribution). As Worley et al. (2013) the tems “reclama- ion” or “rehablation” were not included, in order to focus on studies that were in line with the Society for Bologcal Restoration (SER) def nition of ecological restoration. Besides that, standardization of re storation terms still in process (hfeDonald etal. 2016), and, for this reason, terms as “reforestation” and “recovery” were also included to ‘mal adresses gardonh.aspr (ER. Gardon), oreo com be (LEA Santos), reslguspbe (A. Rodegues) tp: /d.org/10.1016/ ore. 2020.117972 Received 1 November 2019; Received in revise form 19 January 2020; Accepted §Febranry 2020 (0978-1127/ © 2020 Fever BV. Al ght reserved‘account for more published studies. Because we used the query “forest”, we highlight that our results are representative ofthe Brazilian forest formations embedded in Amazon, Atlantic Forest, Caatinga (Tropical Dry Forest) and Cerrado (Savanna) biomes ‘We refined the search by identifying the studies that addressed the following eriteria: (D studies cactied out exclusively in the Brazilian territory; (i) empirical studies, field surveys of observational data (excluding reviews, meta-analyses, perspectives, modeling, GIS assess ments, and scenarios papers; (il) studies that directly or indirectly ‘estimated biomass or carbon in restoration sites; (iv) studies conducted in areas previously forested, where native vegetation was removed for human interventions (agriculture, pasture, clear-cut, ete.) and then naturally regenerated or restored by other methods; and (¥) excluded eos of selective logging, assessment of gap dynamic, studies con: ‘ducted in greenhouses, experimental sites, commercial plantations, ‘monoculture, and agroforestry We selected only the studies that evaluated live and/or dead AGB ‘and/or carbon (ie. littrfll and plants aerial parts). We decided (0 restrict our study to aboveground stock because most of the below: ‘ground (soil studies found did not included roots contribution, but soil ‘carbon contents and microbial biomass. For this reason, we kept our review aligned with aboveground compartments, since plants structures were not the main target of belowground biomass studies. 2.2, Studies classification ‘Studies included inthe final list were classified within eriteria en: closed in Table 1 ‘The geographic position ofall sites sampled in the selected studies ‘was mapped t0 the most detailed scale provided in the article (co: ‘ordinates, locality, municipality, state). Where it was possible to iden ty the same site in more than one artee it was mapped onee, thus we found 25 replicated sites. The biome was determined based on the Brazilian Biomes Map provided by the Brazilian Institute of Geography and Statistics (IBGE 2004) in 1:5 000 000 seale. All mapping procedres were conducted in QGIS (Version 3.412), Past land-use was determined ‘only ifthe information contained inthe article was local, not the su roundings. Some articles defined thresholds in the plant individuals size (diameter and height) to be included in their database, and ve ‘bsained this information for all articles. We defined as passive re storations sites areas where forest was cleared, used oF not for human ‘etivities, and then abandoned for natural regeneration recovery, that is, secondary forests, On the other hand, active restorations were de fined as sites previously forested, used or not for human activities, but recovered by human interventions, as planting seedlings and direct seeding 123. Aboveground biomass temporal analysis We extracted the AGH/carbon values and age ofthe restoration sites from the texts, tables, and figures available in the seleted publications. We included in the analyses AGB values obtained by the use of allo metric and volume equations or desiictive methode to estimate the AGB stored in the plant individuals. We only attained information fom restorations of age < 30y, since ths isthe most common age class of the restoration sites found among the studies (> 90%). Due tothe impossibility in determining the exact age or AGB stock for some of the restorations sites studied in the articles, the dataset is ‘composed of individual and grouped (average age ar AGB) values. ‘Thus, we obtained both AGB and age for 177 events. Data presented as cearbon values were converted to biomass by the carbon conversion factor given in the particular study. Because AGB recovery overtime Is a non-linear process and is ex pected to be different among restoration methods (ioll sax Zalawi, 2014), we estimate ste’ AGB stocks (MgAGB.ha~?) as function of age ‘ensuring an asymptotical shape in the AGB accumulation pattern. We table 1 (Categories and subcategories of studies clasleatlon. Factor of ines on AGB cory 5 i 2 Pubation dea Tandscpe ore abuts Reoraton inpleentatin, as, nd herbaceous Pet AGB or estan sqution; Mot nae ad ueforme erated te, hab, pal 98 Year of publeron Joumal of Subestgn es egy and Managemen 482 (2020) 117972es egy and Managemen 482 (2020) 117972 6 g 5 2 4 oc | a1 oe g22552289828 28 SRE SARE RRR aaRe ARS {= tT 1 5 ae g 38 = i aB3 8 butt second- 50% of the whole AR sites Identified (Fig. 2). Only 89% of the articles evaluated ARs and PRs ofsimilar conditions and at the same time (Ferreira et al 2015; César el 2018), ‘We identified a variety of human land-uses practiced among sites, before restorations implementation, ranging from clearcut t0 4 com: bination of agriculture, pasture, and burning (Table A.2). Agriculture and pasture were the most common past land-uses among restoration sites, handled in $79 ofthe articles, Almost 23% of the studies eval: uated bummed sites, and all of them were conducted in PRs, mainly in ‘Amazon (758%). PRs derived from silviculture a past and.se aetvity ‘were also found in Atlantic Forest (César ct al. 2018; Almeida et al 2019; Rosenfeld and Miller, 2019s; Rosenfield and Miller, 2019), ‘and an actively restored site previously degraded by hydropower dam ‘constriction was also found in this biome (Siva etal. 2015). 3.3, Methodological criteria adopted by the stdies to estimate AGB To estimate AGB of plants aerial parts, studies used an averaged, total area sampled per site of 2363.2 m®. Among stdes, the total area sampled per site ranged from 75 m* to 13000 m# (Table 4.3). Some of them used additional sample plots to calculate the biomass of shrubs, herbs and small plants (<1 m height) by destructive methods. The ‘rteria for inclusion ofthe plant individuals used inthe articles were diameter and height. For plant diameter, the minimum size ranges from > 0.2 to = 10 em, and for height from > 1 to > 2m (Table A.) Moreover, 37% of the articles that evaluated AGB in plant aerial pats Included only individuals of DBH = 5 em. Fifty-five allometrie equations were used among sues to estimate AGB, and one to direct estimate carbon (Table A.4) Of these allometric ‘equations, 45% were developed based on weighing mixed species and 54% of them derived from Brazilian forests, distributed in the Amazon (= 6), Allantie Forest (n= 3), Caatinga (n= 4), and Cerrado (n = 1) biomes (Table A.4). Only 28% of the mixed species equations used among the studies derived from restored forests, all of them were de veloped in PRs (naturally regenerated) of Amazon and Atlantic Forest biomes. We observed that Amazon was the Brazillian biome with the largest set of equations, with 43% of the Brazilian mixed species ‘equations found in the studies, and the two mixed species equations most used among the articles were derived from this biome (Uhl e 1088; Nelson et al. 1999). We found seven mixed species equations developed in PRs, and no equation derived from sites restored by AR methods (Table A.) ‘We also found 30 equations specific for 16 species, four plant genus, different life-forms (lianas, palms, and shrubs), and the Araliaceae bo: taneal family (Lable Ad). However, only 41% ofthe studies included non-tee life-forms in the AGB estimates (eg. shrubs, lianas, palms, bamboo, and herbaceous plants) (Table A.9). Almost all studies that ‘accounted for diferent life-forms vere conducted in PR sites, with only ‘one study conducted in ARS. We also found an equation to estimate the leaf biomass in Caatinga plants (de Souza etal 2012). ‘The allometri equations used to estimate AGB in the articles are based in plant's atrbutes, as trunk diameter, diameter plus height, Dasal area plus height, basil area, erown area or a combination of diameter, height and wood specific gravity (Table A.4). Among the mixed species equations used in the articles, the diameter ofthe felled ‘Table 2 actors of influence and parameters telat to AGB recovery fund inthe articles, es egy and Managemen 482 (2020) 117972 and weighed plants ranged from > 0 to 212 em and the number of individuals from 14 to 4004, but only 32% ofthese equations presented plant individuals with diameter 2 100 em, and 56% felled and weighed =100 plant individuals “Just 21% of the articles that estimated AGB in plant individuals accounted for wood specific gravity (Table A.9, a8 a constant factor (indner and Sattler, 2012; da Silva et al. 2014; Galvdo et al. 2018; Silva etal. 2016; da Siva etal. 2017), or averaged values for specific species, genus and botanical family (Uhl etal. 1988; Sieininges, 2000; de Souza et al. 2011; Imafa-Encinas etal. 2012; Robinson etal. 2015; Silva etal. 2015; Sansevero etal. 2017; César etal. 2018). Carbon stored in plants biomass was evaluated only in 25% of the studies, whose assumed that 4586, 47%, or 50% of the AGB values correspond to carbon elements (Markewitr etal, 2010; de Melo and Dutrigan, 2006; de Souza etal 2011; Berenguer et al. 2014 Sattler etal. 2014; Robinson et al. 2015; Moura et al. 2016). Destructive methods and laboratory procedures to calculate carbon concentration in di ferent species were also observed (Foldpausch al. 2004; Fearnside ral 2007; Silva et al. 2015; Pereira et al. 2016) Biomass stored in liter compartment has also being evaluated at diferent forest types, conditions and restoration methods. We found {har litter biomass was evaluated {n 23% of the articles (Table A.3), 75% of them conducted in PRs (Gamo-Rodrigues tal 2003; Markewitz er al, 2010; Gama-Rodrigues et al. 2007; Vendrami et al. 2012; Berenguer et al. 2014; Moura etal. 2016; Pereira etal. 2016; Peixoto etal 2017; de Azevedo et al. 2018; Froufe etal. 2019; Rosenfeld and Maller, 20198, 20195). Liter stocks were most evaluated in Atlantic Forest, at both PR ard ARS. 34, Factors of influence on AGB recovery ‘We observed that 70% ofthe articles related AGB to one, or more, biotic and abiotic factors. Forest attributes were the factors of influence ‘most related to AGB among the studies, including forest age, vegetation ‘ype, plant community condition and composition (Table 2). Restora- Udon age, that is, the time elapsed since restoration implementation, was the factor most evaluated among restoration sts, and our temporal analysis of AGB revealed that sites restored by different methods pre sent divergent pattems of AGH accumulation in the middle/long term ig. 9. 4. Discussion 4.1, Diaribution of AGB recovery studies among brasilian biomes and restoration methods: The impacts to nationa-seale generalizations International commitments forthe recovery of degraded lands, the rise of carbon credit markets, the pressure to reduce OO, emissions by forest conservation and restoration has been on the government's agendas for more than 40 years, sinee the First United Nations Conference on the Human Environment in 1972, and the First World Climate Conference in 1979. Along this period, political marks as the United Framework Convention on Climate Change (1992), the Kyoto Protocol (1997) and the Intergovernmental Panel on Climate Change Taco of auence on AGB emery Parameters aries anda lg ts aroun of ace eet slop, a aration to w= 5) oe ates ‘ge ovat per sucesso! phase; ven (chess Sion, an haan plan ey and eforn ASH (8 = 23) esoraton implementation ‘etotin bod (piv, ate ad ue! aon Bean) et landse| ‘bp atime of nd ae ad aie degradation evel ew) Sal Sol ype and si properties wea 8) Plan teraction Testor. Boa 1) inate Annual reiation a== Passive R=02001 RSENS 300 Biome var + + saa sca, 200 AGB stock (Mg.ha) 100 o 10 20 30 Restoration age (years) Fig. 5. Predictive polynomial models Fite to estimate AGB tock MAG.) in active (adjusted RE = 080: p < 0.05; Residual Standard Error = 2633 MgAGR.ha Nin= S7:AGR = ~1935058 + 4.00902 "age + 021681" (age2) and posve (adjusted R= 02% p < 0.05 Residual Standard Boe RSE = 4915 MgAGRna n= 14; AGH = 10079 + 7.3956 * age ~ 01158 * (2) restorations 3s 3 response of ae. The dak sd lin the trend of AG ck covery overtime ‘observe in passive restorations among biomes, andthe lightly shaded line isthe tread for acverstoraons. The shaded ara along the predive Uns represent the conidenoe interval (05) forth model Data pos were categorized by biome [A (ele) ~ Asai ores, AM tangle) ~ Aras CA (are) ~ Catia: and (CE rss) Cerrado. (PCC) lead global governments to prioritize climate change mitigation andl to preserve and recovery natural ecosystems. For this reason, we ‘were expecting to observe an earlier increase in the number of studies related to AGB recovery by restored forests, improving the knowledge ‘and supporting political decisions over the yeas, instead of only in the last decade. In 2008-2009 the National Plan on Climate Change and the National Politi on Climate Change were implemented in Brazil (Br, 2008, 2009), This ean explain alate and slight increase inthe number fof publications/year, reaching the peak in 2019 (Fig. 1). Although ‘ecological restoration in Brazil present a solid and growing scientific production, with almost 300 articles published over the last 30 years (Guerra etal. 2020), specific evaluations of AGBYcarbon recovery in these systems does not keep up with the inereasing political decisions ‘and appealing discussions related to forest restoration and climate ‘change. Furthermore, beeause the Brazilian post graduation network present solely almost 300 programs directly related to biodiversity and ‘environmental sciences researches (Souza and Fernandes, 2013) we ‘expected to find a more diversity of institutions improving scientific knowledge related to AGB recovery ‘We argue thatthe continuous improvement of the general knowl ‘edge could be attained to the broad range geographic distribution of scientific studies, allowing to predict AGB recovery patterns of different forests restored by diferent methods. However, we observed an wn: balance in the restoration methods evaluated among biomes, where ‘outcomes of AGB recovery from restored forests lack for some parts of Brazil and only 23% of the studies evaluated AGB recovery in ARS. ‘Guerra etal (2020) observed that AR (seedling planting) isthe most, wed restoration method among Brazilian biomes, but we found an ‘expressive predominance of PR sites among the studies of AGB/carbon, (ig. 2) In addition, studies have shown that 78% ofthe and allocated for restoration projets in the Atlante Forest, Atlantic Forest/Cerrado fecotone and Amazon was designed for AR (Brancalion et al. 2016), “Together, these arguments evidence that scientific knowledge related to es egy and Managemen 482 (2020) 117972 AGB recovery isnot aligned with the restoration methods most needed and practiced in Brazil ‘The fact chat we found only one study that evaluated AGB stock in ARS implemented by direct seeding, an efficient restoration method cheapest than planting seedlings and able to trigger initial successional processes (Freias etal 2019), evidence another bottleneck in the field However, this result ean be a consequence of the focus of our study forests, seen that direct seeding i a restoration method used in most of the Brazilian biomes, but mainly implemented in non forest ecosystems. Due to agricultural expansion, Brazil is one of the major cor- tributors to deforestation and to land-use change emissions (Ze st l. 2016) and the trend is that Brazilian biomes will continue to be de forested (Soares-Filho etal. 2014; Rochedo etal. 2018), inereasing the red for restoration efforts, In this sense, our results suppor the need for scientific evaluations of AGB accumulation in regions with weak information about the recovery of this ecological process, beyond Amazon and Allantie Forest biomes. This is erucal to the effective achievement ofthe restoration goals, based on realistic site conditions. Better understanding temporal changes in the process of AGB/carbon accumulation among restored sites, could leverage the improvement of |AGH/carbon monitoring protocols, alloting a confident interpretation bout the large-sale potential of forest restorations in mitigating ell mate change, 4.2, Methodological divergences in AGB and carbon stocks etimates in ‘Brullian restored forests ‘The rigorous selection of the appropriate allometrie equation is fundamental toa reliable AGR estimate and studies have shown tha the main source of eror is attained to the ehoice of the allomettie model, lonce not representative equations of a targeted forest can lead to sub stantial errors in AGB estimates. Even in areas with similar conditions, as climate, sol, forest type, restoration method, and age, major errors can be introduced (Chave et al.2004; Lindner and Sater, 2012; Sileshi, 2014; Silva etal. 2015). In tropical forests from Central Ameriea, Chav tal (2004) found that different allometric models can estimate AGB witha variation of 246 MgAGB.ha, even when applied to the same plot In Amazonian PR sites, using eight different models co estimate AGB stock, Wandelli and Fearnside (2015) observed that the mean error oF estimates fr accumulated biomass varied from 5.6% to 57-59%. These studies highlight the need fora representative set of allomettic equa: tions ofthe many tropieal forest formations. “Thus, i eto be assumed tha those fsues weve taken into sesount by AGB recovery studies, but our results show major divergences. Studies of AGB recovery in restored forests of Brazil have used eque- ons decived from Brazilian natural forests and also fom tropa for- ests of other countries to estimate AGB (Table A), forests which may present evident differences from the targeted one. The few equations derived from restored forests (2896) are not representative of the many forest formations and restoration methods used in Brazil, and we did not find in our database any equation derived from AR sites. ‘The similarity between the ecosystem thatthe equation derived and, where it will be applied is important, bat the equation inputs mist alsa be considered, The allometric equations wed inthe studies are based on plants trunk diameter and height, but basal area, crown area or a ‘combination of diameter, height and wood specie gravity areas used {o estimate AGB (Saldarriaga et al. 1988; Deans etal. 1996; Chave etal. 2005; Chave et al. 2014; Robinson et al. 2015). Chave et al. (2005) observed that the inclusion of hefght inthe equations reduced the etror of individual tree biomass estimation from 16% to 6%. In contrast, uncer et al. (2013) argue that height measurements in tropeal forests tre labor and present potentially lage errors, related to forest cond tions (eg. dense understory vegetation, tll canopies, and elosed-« nopies), observer experience, and the equipment use. “The performance of allometric models is also highly associated with the diameter range and the number of weighted plants ineluded in theirdatabase, and equations developed from a small numberof individuals ‘and small diameter range should be avoided (Chave et al. 2004), ‘Among studies, we found seven equations encompassing = 100 Weighed plants and individuals with large diameters (= 100 em), but ‘only one of these equations was derived from a Brazilian forest, an Amazonian natural forest (dos Sontos, 1996; Ferreira and Prance, 1099), Despite some equations enable the inclusion of small-sized plants, 3796 ofthe studies included in their AGB estimations only plant individuals = 5 em diameter Of course, individuals within this range represent the most share of AGB in mature tropieal forests, but in young restorations (2-4y re ‘generation) 100% of the AGB ean be stored in the 1-S em DBH class, ‘decreasing to less than 10% in older restorations (12-14y) (Fldpauseh ft al. 2005). Thus, biomass evaluations in restored forests should be aware of the contribution of these smalsized individuals to AGB re covery and, depending on restoration method and the successional sage, neglecting smal plants can cause misleading Interpretations of the AGB stored in restored Forests ‘This fae can be extended to the limitation of using wood specifi gravity, due co the searce availability of experts able to identity species in field campaigns. n spite ofthat, besides wood specific gravity be also recognized as an important species attribute to reliably estimating AGB stock on forest stands (Brown etal, 1989; Chave etal. 2006; Chave eal. 2014), the majority of the studies in Brazilian restored forests have neglected this attribute ‘Another highlight we found is that the number of species used to develop the Brazlian’s mixed species allometrie equations Is not even lose to the plant diversity of these topical forests. The importance of Including more species an lfeorms inthe allometrie models relies on the processes of weighing plants components (teunk, eanopy, leaves), while species store different amount of biomass among their compo nents. The concern of including more species in the allometeie equ tons also involves accounting for diferent life-forms. Besides trees share the highest proportion of AGB stocks in tropical forests (Ligot tal. 2018; Meakem et al. 2018), other life-forms ean play important roles in ecological processes and functions, as forest productivity (Gerwing and Farias, 2000; Schnitzer and Bongers, 2002; Gehring etl 2005; Alves etal. 2012). For example, in Amazon, shrubs can corre spond to 74% of the total AGR stored in 2-y regenerated pastures (eldpausch et al. 2005), and among different forests in the Atlantic Forest biome, lianas, palms, tree fers, bamboo, and epiphytes can share more than 10% ofthe AGB (Vieira etal. 2008). Despite the share of diferent life-forms on AGB stocks, only four ‘equations and less than half of the studies ineludest non-tre life-forms in their database. Moreover, no AR site was evaluated within this ‘concern. The conteibution of diferent species and life-forms to AGB ‘accumulation in tropical forests cannot be neglect, and this ise fs ‘accentuated by the urgent need to include more diversity of species and life-forms in AR projects (Rosirigues et al. 2009) Studies are selecting equations not well ited with thir objectives, ‘or even their forest types or stage, suggesting the lack of multiple ‘equations adjusted for restored forests and at different conditions, ‘which can lead to unreliable AGB estimates, In a first effre to overlap this gap, wide and robust allometsie models must be available for ‘generalizations in each biome. Whenever possible, these general models should encompass data from the different vegetation types of each biome, as the Atlantic Forest formations recognized as Orbrophilous (Dense, Mixed, and Open) and Seasonal (Semidecidual and Decidual) forests, or even the forested Savannas formations (Cerradio). Other studies could regard the inclusion of allometric equations fited for vegetation assorted with water courses, and also mangroves located at costal zones (Ferreira et al. 2015; Ferreira etal. 2019), since this proximity directly influence the vegetation patterns (structure and composition) at different biomes (Coutinho, 2016; IBGE, 2012), Because evaluations of AGB stocks in forests are based on the ex: trapolation of data collected at plot scale, plot size and plot position are es egy and Managemen 482 (2020) 117972 also important issues to ensure robust AGB estimates, Chave et al. (2004) proposed that plots should size 2500 m to estimate AGB with good confidence, representing the spatial variability of forests AGB. In general, the average plot size found is smaller than that for studies conducted in Brazil, and this pattem seems to be widespread among twopieal restoration sites — 947 m, average size for ~1468 plots of Neotropical secondary forests (Poorier etal. 2016). ‘The conversion of AGB on earbon stocks also need tobe highlighted. Wie observed that carbon concentration in dry AGB, used as an earbon conversion factor among studies, varied from 45% t0 50%, However, i is known that diferent forests, with different species composition, ean present disparities in average value of carbon (IPCC, 2006). Obviously, generalizations are important, especially to conduct large-scale studies, but these methodological aspects might be rigorously thought, ance i fan result in not confident estimates of carbon stocks and CO» se ‘questation potential by restorations. Ie is known that technical and legal issues ean bar Dullding allo metric equations in restored forests and, for a while, authors are re stricted to the few equations indicated for AGB estimations in these areas. For example, costs related to destructive methods of AGB est ‘mation can reach US$ 11.00 per plant individual (Silva, 2007), af feeting the viability of fitting new equations based on large datasets Nevertheles, some advances for estimating AGB stocks in AR sites has "sen in the Atlantic Forest biome (Miran eal. 2011; Nogueira Jénior etal. 2014; Fer etal, 2015; Ré etal, 2015; Zanini, 2019). We state that our goal was not to evaluate the whole set of equations available for Brazilian forests, but those that have been most used among studies to-estimate AGB in restoration sites, 43, The main drivers of AGB recovery sued among Brazilian restored Forest atributes (45%), past land-use (16%), soll properties (16%), and landscape metrics (108), were the most factors related to biomass, comprised in 65% of the studies. Among the forest attributes, time elapsed since restoration starts (eg. age) was related to AGB stocks in 29%6 of the studies condueted in Brazil. Restoration age Is important because enough time is needed to forests succession and ecological processes, as AGB accumulation, to reestablish (Crouzelles eta. 2016) Besides that, age is more correlated to AGB stocks in ARS (R= 0.82) than PRs (R? = 0.22) (Fig. 3), suggesting that in PR other factors present high influences on the AGB stock Plant community composition has a relevant importance on AGB stocks in restoration, sine fast-growing species presenta higher con: tribution during the early years (< 37 years), but the contribution of slow-growing is significant at later stages of succession (Shimamoto cal 2014), This suggests chat AGB necumlation patters i not con: stant over time, an is related to the changes in the plant composition long the restorations seeesstonal trajectory. In this context, forest type and successional stage are also related to AGB stocks (Lu et al. 2003; Santos et al. 2003). Costa et sl. (2014) observed that Caatinga PR sites of dense forests and at advanced suc fessional stages accumilate more than 2-fld the AGB stock of open forests at early regeneration stages (45.8 and 20.7 Mg ha, respec- tively). Similar results were also observed for different vegetation types at different successional stages in Cerrado biome, where sites of Oer- ado (the Brazilian Forested Savanna) presented more AGB stocks than grasslands and typical Savannas (de Miranda et sl. 2014; Roquett, 2018), Because Brian biomes comprise a mosaic of vegetation types, these results imply in a huge challenge to understand how AGB iste covered by restoration aetions ina full range of situations. ‘The influence of plant diversity (species richness, Shannon and Simpson index) on AGB stocks was accessed in 99% of the studies, conducted in Amazon, Atlantic Forest and Caatinga biomes, pointing to «a positive relationship with AGB recovery (Vieira etal. 2003; de Aguiar cecal 2013; Robinson eta. 2015; Rosenfield et al. 2019b). The postiverelationship between plant community diversity and biomass. pro: ductivity oecurs du tothe coexistence of species that ensure a greater ‘complementarity and facilitation in forested ecosystems, and conse ‘quently a better efficiency in the use of resources (Wrigit, 2002 Nakamura, 2008; de Aguiar etal. 2013). For example, de Aguiar etal (2013) estimated that enhancing species richness from 1 to 10 species, PR sites of Caatinga (693) annual AGB increment can increase 5.1 Mgha~!y", These results have a direet effect on restoration str ‘egies, providing the subsidy for restoration ecology, which isthe basis for any ecological restoration. ‘The AGB recovery rate also depends onthe site's conditions, as past, land-use, adjacent forest cover, climatic water deficit, and other social ‘and political drivers (Chardon et al. 2016; Crouzeilles et al. 2016; ‘Chazdon eta 2017). Studies have found tha, in restoration sites at fat terrains and embedded in landscapes with high forest coverage, the AGB recovery is more efficient for both PR and AR methods (Settler et al. 2014; Robinson et al, 2015; Toledo et al. 2018), Sailer et al (2014 found evidence that AR sites located on sloped terrains stock less than half the earbon (17.5 Mg.ha~*) store by similar restorations on fat terrains (37.6 Mgiha~") in Atlantic Forest biome. ‘Throughout the landscapes, the amount and distribution of sur rounding forests are the main regulators of seed dispersal witha strong effect in early-successional trajectories, while flat terains are asso ciated with high water retention, being both relevant drivers of AGB. ‘accumulation (Osman and Barakbah, 2011; Uriarte etal. 2011; Holl and “Zalnawi 2014; Sattler etal 2014; Robinson etal, 2015). In 30y Atlantic Forest PRs, Robinson etal. (2015) observed that the amount of adjacent forest and slope are predictors with the large effets on biomass tock, Positively and negatively correlated to AGB, respectively. Edge effects were also studied, and sample plots near tthe forest ‘edge presented low AGB stocks (Sorenguer etal 2014), result that ean be explained by the decline observed in AGB. within 100 m from forest fragments edges, where tree mortality is increased by mleroclimatie ‘changes and wind disturbances (Laurance et al. 1997). Regarding these ‘evidences and the fact that a signifiant par of the environmental debt ‘encompasses legally protected forests at riparian and sloped areas (Gederal Law, 12.651; SoaresFilho et al. 2014), which canbe limited to ‘rips of m width and steep slopes, restorations success and goals can be hampered. Land-use type, degradation level, time since last disturbance and under landuse, were observed to have significant effects on AGB. Time since last fire event presented postive effects on AGB recovery, while {grazing, time under land-use, and land-use intensities negatively af fected AGB recovery in Amazonian PR sites (Uhl etal, 1988; Feldpauseh ‘etal. 2007; Gehring etal 2005; Berenguer etal. 2014). in Amazon, PRL sites (« 15y) previously used for agriculture presented faster AGB ac ‘cumulation than PR sites regenerated from pasture lands (Wael and Feamside 2015). However, 7-20y PRs regenerated from pastures in Adlantie Forest can store 439% more AGB of native species than re ‘generated sites of similar age derived from Eucalypus sp. plantations (César etal. 2018). The number of burns also negatively affect AGB (Wandollt and Fearnside 2015; Sansovero et al. 2017) ‘Soll parameters presented important effects on AGR stocks among. Brazilian restored forests. In PRS derived from pastures in. Adantie Forest, soil sand and clay content presented negative and positive ef fects on AGB recovery, respectively (Robinson etal. 2015; Toledo et 2018), These properties are related to soil water retention capacity, an Important driver for plant growth and AGB recovery (Lu etal. 2002) Higher AGB stocks were observed in Atlantic Forest ARS at fertile and layish soils than the same restorations at sandy and poor soils (de Nelo ‘al Durga, 2006). Moran etl. (2000) also found signffeant effects of sol type on AGB stocks in passively restored forests in Amazon. Besides sil fertilization play an important role on any plant growth, undesired ‘outcomes as benefiting aggressive and invasive species may negatively ‘affect AG accumulation, but control interventions present positive effets in these eases (Siddique et al. 2010, Fez etal 2015) es egy and Managemen 482 (2020) 117972 ‘The iter compartment and its contribution to total biomass stocks in restoration sites also depends on factors as restoration age, structure and composition. in Amazon (19y secondary forest), Adantie Forest (oy secondary forest), and Cerrado (22y secondary forest) biomass stocks in literfall was estimated to reach 67 Mg.ha~', 7.2 M.ha~!, and 39 Mga", respectively (Matkewitz et al, 2010; Gama. Rodrigues et al. 2007; Peixoto et al. 2017). In Caatinga (30y re generated forest), annual production of titterfall biomass ean reach 6.1 MgAGB.ha“!y" (Pereira et al. 2016). In secondary Allantic Forests at diferent regeneration stages, annual production of litterall biomass was higher in early stages of natural regeneration than in more ad: vanced forests, and this is attained to the high proportion of pioneer species in young secondary forests, where species present higher leaf ttumover than late successional species (Vendram’ etal. 2012). Besides we found afew studies ofitter stock, they are also high concentrated in PR, Only three studies evaluated litter stock in ARs, exclusively in the ‘Adantic Forest biome (cle Azevesto etal. 2018; Rovenfeld and Miller, 2019s; Rosenfeld and Miller, 2019). 4.4, Temporal analysis of AGB recovery in Brasilia forest restorations In recent years, studles have focused on to determine which one, PR lor AR methods, is more successful in recovering AGB stocks, vegetation structure and biodiversity (Brancalion et sl. 2016; Croszeilles etal 2017, Meli etal 2017), We found only two articles that compared AGB Of AR and PR sites at similar conditions (Ferreira etal. 2015; César etal 2018). César eal. (2018) studying AGB recovery in Atlantic Forest, found that PR sites derived feom pasture stock 45% (91.3 MgAGB.ha~") less AGB of native species than AR sites (132.2 MgAGB ha”) from 7 t0 2o.yearold, In an Atlantic Forest mangrove, Ferreira etal. (2015) also found this patter of AGB recovery in S-year-old restored sites, but AGB stock in AR (60 Mg.ha™* was 3-fld the AGB of PR sites (20 Mg.ha~*) (Our model estimated that PRs of 30y can stock 38% (127 Mg.ha"") of the AGB estimated for ARS of the same age (334 Mg-ha~!) (ig. 3) Other studies have found that AGB recovery in 30y PR sites ean reach 98,5 MgAGB.ha~! in Amazon and 43.3 MgAGB ha! in Caatnga sites, while 25y PRs in Atlante Forest can accumulate 1588 MgAGBLha=? (de Souza et al. 2011; Galvao et al. 2015; Pereira etal 2016). ARS in Atlantic Forest and Cerrado can present AGB stock higher {han PRs of the region and similar age (de Melo an Durigan, 2006; de Souza etal. 20115 César tal 2018), While AGB stocks of AR ste ofthe Atlantic Forest can reach 222 MgAGB.ha-1 at 25 years after restoration Implementation (Durigan ct al. 2016), 25 years-old PRs of the same biome ean stock from 73 f0 159 MgAGB.ha~! (le Souza etal. 2011), “The results corroborates the pattem of AGH accumulation between restoration methods, where ARs present better outcomes than PRS. In general, it ean be observed that the processes of AGB accumulation in the fist years of restoration are similar at both active and passive re storations (Fig, 2), but after 8-10 year AGB accumilation restoration methods start to present divergences. However to compare PR and AR ‘outcomes depends on how restorations data sets are categorized as passive and active methods, and controlling for key factors as past Alisturance, landscape metres, climate pattems, and restoration age is fundamental (Crowzeilles etal. 2017; Reid etal. 2018). Nevertheles, comparative stadies of different restoration methods are important to know what to expect from exch strategy, but the selection of the re: storation method fora particular site must be based on its natural re silence, past landase history, and the surrounding landscape matrix Holt and Aide, 2011), It is important to highlight that remote sensing (e.g. LIDAR and optical sensors) are useful tols that have been utilized among Brazilian biomes to the monitoring of AGB recovery, especially due to the pos sibility of long-term temporal analysis (Luickman et al. 1997; Santos e¢al 2003; Vieita tal. 2003; da Silva etal, 2014; Galvao etal. 2015; ‘Almeida et a. 2019), Also, although allowing quick data collection of forest structural atiributes, remote methods are supported by flektmeasures, needed to the validation of the estimates. 4.5. Advancing the fld and implications for future studies We observed main researches bottlenecks related to studies dis. tribution and methodological isues, expecially the lack of studies ‘across the whole Brazilian territory and equations fitted for AGB est mation in restored sites of different forest types, The real contribution fof forest restorations to climate change mitigation stil weakly ad ‘dressed for some Brazilian regions, elther by the absence of AGB studles from ditferent biomes and restoration methods, or by the Weakness of the methodology used t0 estimate AGB and carbon stocks. Regarding the ecologial role of biomass asa dever for restoration success, stules of AGB should be conducted at restored forests in a ‘wide range of situations (restoration method, landscape forest cov ‘erage, soll and forest formation). AGB researches must be intensiied in restored forests, beyond Amazon, and at Doth passive and ative re storations. A robust set of allometric equations representative of Brazilian restored forests, and a database on species wood specific ‘gravity and earbon concentration, available for scientists and decision maker usage is urgently need for Brazilian ecosystems. In adtion, we defend that depending on the restoration method, environmental con ditions and the successional phase of the recovering sites, different lif. forms should be included, leading to an accurate estimate of AGB re- ‘covery, Yet, studies are limited to the methodological alternatives that account for this fsue, and efforts must be applied to overcome this barrier. An effective path to overcome this barrier would be the develop ment of Federal or State standardized monitoring protocols for forest restoration, where aboveground biomass estimates would be attached In their routines. A minimum se¢ of allometcie equations should be available for different forest formations, a least at biome level, re ‘garding the particular ecological aspects of each ecosystem, as plants architecture, size, wood specific gravity and carbon concentration ‘Together with these approaches, the inresing usage of could result ina ‘qualified database, helpful to develop robust scenarias able to predict, the real contribution of different restoration methods to AGB recovery, ‘carbon sequestration and climate change mitigation in Brazilian lands. Because Brazilian biomes present diferent demands for restoration methods, but the need for both passive and active restorations exists in ‘every biome comprised in this study, restoring these lands performs an lunprecedented opportunity to address the knowledge gaps highlighted here, Finally, forest restoration programs are global tools to mitigate biodiversity loss and the climate change effects, and, fr this reason, the ‘outcomes of these appronches must be deeply understand, and science ‘must continually move the knowledge forward, otherwise, already set ‘and future efforts can fil im achieving the expected target, Declaration of Competing Interest ‘The authors declare that they have no known competing financial Interests or personal relationships that could have appeared to inte ‘ence the work reported in this paper. Acknowledgment ‘This study was financed by the Coordenago de Aperfeigoamento de Pessoal de Nivel Superior ~ Brasil (CAPES - Programa de Exceléncia ‘Académica) ~ Finance Code: 88882,327891/2019-01, The authors have ‘no conflict of interest to declare. Appendix A. Supplementary data ‘Supplementary data to this article ean be found online at hisps:// doi.org/10.1016/)foreco.2020.117972. es egy and Managemen 482 (2020) 117972 References Atnala, DRA. Sah, $C, Chanfon Neon, BY Ces LG, Ml P Grass, Dante MAL. Valen, R Mateo, V3. Alsen AF, Amaro, N Goncalves NBL Sita CA Sth, Brncaon PHS, 2019 ecene of ier remote sein fr meray oe avec tists ad tine estar. Foe, Ee Manage tps our 101016 or 20190202. ‘EA day, CA 2512, Variation in Lt Adon ad Boma Along a eto! Gracie the Topi atic eves (Baz). Bea Res 27 525-92 bipdet org 1007/1 LaRe OT O Bunn A, AdoaatS, 2017. Foret Restoration zona. The Nature Comer ‘So Psi, Ba Berenguer Fermin J, Cane, TA, Arg, LEO.C, De Camara, Pa, Ce CE, ‘Durga. De lve, Ue, Vie, LEG, vow, J 204, angele eld ‘eset of an stocks i buna dled opal ets lb: change BL 29, 718-36, hos doy 1011 ub 12627 TET, Nave, AG., Rodelgues, LR, 2016, Balancing economic coats snd eccogieal ‘tomes of ae ntact rsortion nage nae he ceo Brwat Biovopien 4 80-007 frp "every T0110 ple [BRASIL Pian Naina sobte Mulan do Clin ~ PNM. Bas Besa, Deceto 26 "26 de 21 de novembre de 2007, 2008 ny goer se ates ‘Suter poe mona cana ca pol (atened 27 December 2019) ‘Mata do Cina = PNMC dus provides, Dito Oil da Un, rutin, Sogio Eager p. 10810, 208, lato sre ei Chateau? a0i0/2000 L417 hn aces 27 Dec 2009), Brown Sy lll, A, Lugo, AE, 196). Bits etalon ethos pieal Tires with pitino oe ier at ort Se 35 (0), B32 racion, PH, 218 tele! ote of mata eset de Danton reining are Insp Eel. App 28 73-384 Ns! (ntag/10 1002 gp 105 ‘ne, Adal, C, Brow, &, Cais, MA, Chamber, JQ, Eas, D, oats, rma ewe, Nit Ler Neton KN, Opa ‘ira BYemahur, T2008 Tee lime an ero esumaton aca Soaks a babe a opel foes Ocala 14,8799. ips /00.04/ 10" Tyraneez ons nb Chav, Cat, Aula Hemander, A, La, Pee, 2004, Err pope "ann an sali or pc oer ome etna PD. Ten. Ro. Soe Bok seb 405-400, /aoon/ 1010872005 1425, Care, J Maller Lands HLC, Baker, TR, Eade, A, Hane Sage, TER, Webb "C6, 2005 Repeal a pyiogenet Variation of Wood ds ass 2435 8 tropie te species. eal pp ips//u 70 esa 10st O76tzDI6) Srelaasenapvowiz0co (Cave. Rj cain, M, Braue, A, Cau, E, Clg, MS, Deli WBC, Degee, A Ee 1. Fea, PAL Goocman RC, Vy, My Martiea Vac Nogan WA, NalerLan, Hi Meneac, t,Revon BN. Needs, A Nop: Eat Oz Mavs, ser tan. Ra, CB Salers Set lsd, 204 Improved allan models ema eave ines of topical ees lab ange Be 20, 3177-3190. par ges 18 (Chutdon, EL, 2014 Second promt: the promise oop fore regeneration in ange of deta J Chem In Model pe ol 10 1017 CBOULU7A2% (huzdos L208, Beyond defrestcon: Restoring oes and ease eves on ‘esas inde Slence. t= kore Tossa Se, (hatdon Ri, raeaion, PHS. Lamb, Lasts L, cao, Muar C, 2017. ' poleydtvenkwiece weds rebel est a ladaps estort, Gone Let ps avy 10.111 sl 1220, oui, Power, Pe Garcia EA, Msgs, Dur SA Ponto, Nigear. Nunes ERP. Porter L, Detsow, 1S, Yan Beigel Mt, Baie ‘Mee, JA, Broan, EN, Suncher Aue, A, Masse, P- Dupuy. 38, Cara, CGAL Bangor ese fi, Labo, M. Brancallon PHS, Solaray 2c ora, Vila, LG, sling, Schwan NB, Bal), DeWalt 33 Sungei, AB Oia Beant, E, Pea Con St Aneids Caer 3 inane Senso, NG, Ae TA, Becktel LM Uetcer SC Rameo- Pete, LE, de Jong. by Rozndal, ALA, Cave, D, Willams, Gi. Bio Sto Mit Jakovae Ce, van der Wai, Vin, LDA, Ru. Ann amos, Fado, iC, 2016 Cate Squestaton pte of son owt fore generation athe nt Ameren roger Se Ag 2 SDS trio fas 301635. (cosa, TL, Sango, EV Si, saa, AME, esl, LJ, ANDO, ED. Prem, EGC, ‘Aibuore, ERG, Weneaes, RSC, 2014 Ro also lias In the {to dy fet fsa Nets Bee Pt Sel 6122-12 sh. tg. sr1104015 20001 oui, 26 Blomae brash Ofna de Ten, (ous, Caran i ere, ALS, Lindenmajer DA, Gree, CEV. Rey Bennyas JM, 2046. A goldtone espa vers of eres omion secest No Camaun 7 tar lula mmole ‘crovailes, Feria M.S, Charon, RL, Linden, DB Sasereo JB, ‘Mono; arent A Laawce AE, Stashure BAN, 2017 Ecologia re Sorton sess ahr fr natural eeoettn than ace estoratn n| ‘Specter tes om ALLE. or apo pears and condo al oad tin the las wh yea arcs SceneRemate Ses St 677-604, s/o oy/10 1081848165, 2014072865, "2017 Degrotan impact pin reso he lower Met esse Drier of Ament Fe Bok Mage #03, 92-101 pcre 1010167. eee 917.0705 dl Agi ML, Fah, 1, de Aj F das CS, Campa MLM, de Otis, TS, “13 Doe biomes predation depend on lat community est? Agrtes'y Sat 87, 690-711 ipso vs 100710487 01295909 deaabede AD, Canara Prtelin, ALR. Perea, ALG aes PSS. 2018 Estoque Ge catono en eas de strain foveal ata tics Rowe 8, ISSN Spud tossbureetone sat eto, ACG, Durga, G, 206, Carbon eqestatin by planed riparian fein Purana Va Sf, Brant PhaAtheo ue carbo en efecto de tas elas no Vale 60 Paranpatena, SP, Bes lesan, oC, Butane, he Pane Mages, ele, ML Feria, 211% Regal vara in las aston a Bella Sinan Wola Brg ipo or 111101.12095. Sout Ad Bois Ar Sore CPB, Visi By de Gp RO, de any 18, 200 ‘Sts ad rows flume, bos, ciboo aa eon eae Sesoal Semidecsunas Forest Resa Aoi a 1015900100, ‘yeaa ioo0?000 de Souza LQ. de Pras ADS, Sampaio, EV. de Sa, Moura, PML Meneaes RC, 72012 How uch alqen Sse by boll sao pica ry foes Tread Ras ng Arcot ROD ea eats, J.D, Moan, Gre, 1996, Biomass eatlonsips fo te spas or ‘eran om douou opel met fort Cameron ar Bea Manage 8 isc age /otorg 10 01ST 127 O0}E os Seo, 1966 Anise Ge modelos de regres pr estimar a toa da rest "roel mda de tea red Anan eases PD Dstertaton eral ‘Dusan, Supanima, MS, De Nelo, AC, 2016 Valves esperades pre anit de horas rips em resauraco em diferes ades: cea Fre Forest ‘ir aGn-474, gs /unsag 10.1667 sar 1015 ‘eect, PM, Barto, Rd Aeneas Grace, PALL, 2007 Burg of secondary ‘ert in Arnaz, Bens bur een an nal ation ding Land Preperation for apcue Apa, orm, Bri For Seal Mange 242, Eras. hp ao 101046 rca. 20070202 ‘Fel Law 12681, 22 May 2012. Native Vegetation Poeton Law Bra. Sr plana sv ec O/ Aen) 2004/2015) acme 14 December 2015. ‘ekg TR da Goneceo Presa. Ferandes, BGM, Ra SJ, 207 Scena fret gre evaton rm ernroson reine natal ‘zona. Got Change Ba 13, 957-978 hanson 10s} es 286, Noro Feldguset TR, Rs, SJ, Femapdes, GM, Wand, EV, 2005. Deslopnent of ert Sects an esc ie ccna) ess enero tuadoncd, {ures ect Amazbal Earth tert 9, ips von 10 1175/10 each FA Ro HA. Frans BM dt £208 itn ‘tania Eoot Appa 104-1: apes diy wolso Theos Fare, APG, Campo, 0, Mons, 6, Se, 2015, vet pparuales Manage 30,4045 hipe//dverg/ 1 1016s 50108015, Forel RC, Beso LEA. Mites Canon, I, 209, Aorearoand carton tock ‘Trenred neovopical nngrove inflcaos of hanugeent sa bauer Sssembage Weands Bz Manage 27, 223-22 ipl ory 101007 51273 tisancee, ‘sera neo! mangrove Beso plat rma rd era coe Dies, Ocean Cost Mane 10, 38-48 is coton/10 1016) cen, Fee, LV. ance, GT, 199, Eosstm recovery in eta ee forests ater cating ‘nd bung: A sompaion on specs ricines,frsticempstion and rest ‘Sacre inthe Jou Natona Pak Arazon Bot Lo, Soe 130,37 11095: rg 101006 oBRCZE Peta MEG. Roda, S., Campos lta, EM. do Carmo, GHP. da Vlg, ML Sangin RG ie DIM, 2015, nang the mice eet nding fo opt fret raring over enya Far Bea Mage 48, 226-205 ta? ‘itong/ 1006 freo 201002034, ‘CES, gai LA Vez, FA, 201, Nase ecg om le ite in Sat sicesional aprofrt) sons ad natural reer train Alte anor Ble Agrotr, Sue 11S poor 10.1007 1085701 ato, U5 os Seaton, J, da ha, RD, da Sa, C¥, Moura, YM. Bewng FA, 201s Following tect second succes ine Anon wing he Lens er pt el etc 3 Roms 4S ‘asa touigues, AC, Bars, NP, tos, NLL, 200, Deconpstion ad mute ease ier pret mined sans of ate tre Species nutes ‘Gama ata A, Bates, NF. Comet NB, 207. ls an tent ecg Ber Gna 87-2 You 0 PONT eta, Dh M, sk, PLC, 208, Rese fn oe ron ater aco 10.1017 asia aos. Z ‘ensing 3, Fas, Di, 200. Ieerting ana abundance an forest str ito an es egy and Managemen 482 (2020) 117972 iat ot above ome fo an en Amazin oT. Es feats tage canon anna Gun: A Re LK Borge, FG. Ota, PIA, Pda, DA. Minds CO. ‘idm’, Seo, TALK, Shibuya PS Nargus, MCh Lauran 8, Gara, 2030, cla eaten Beton Does ing aos ‘Seay Ro Eo aay’ 8 TRE pes Toe en aS Ho, Ale TM 2011, When an where acy ete cae? Fo. aap atts tne ntery Atos joesa ona a ta Ds Zan A, 20 aco expiry wont ave ound cer selon a eed opal fa Ror Eel Msnage 10S freaky tte fora ra ae oui, Som: he pense, 5 Fah BD. Es), eee yo er, 40-8 Huse hele Ve BMC, 2012 Tei pic et Tio xian Rogen 1 aI. e/Focy HOSTS TO Easton "8G ian bons Ro de a. na 15000 06. prj hokage oper pecencins nova toc ‘eivcanatatoe snot onan? at downs ee 23 Samay 219) snc OLY Mana sc avg hei. Manis icone essa 1. aval nignotce yegvnac ero hve te taf Encis be Pal, Conesen CDA, 2012 Wood volume ad Homes of wget of be Mat Alin Sau Marae coum, ate pat, Sty a | Pre voles eis Se sing de Ma ‘Sis mn det fp os 1 yaormental Fans on imate ge, 208 Spec pean Chon ude Cape Sere Perey Wenig Gi of {neon Peel ot Cine Cage Cambrg Ulery Pes, ming Ute Kingiom nah Yom NY, USA este = Sse. eres tance wekep 1p eed sy so {Paes Fae on Cmte ge, 208, lin fr aon ‘penne pur ment pr re: ose Se SeSpigsot pica! fe vuunea von eh eta pa cond 0 None 308). Sam IM ater ean GD, ies, CE, Mile C3 Ber, 3D, aan TD, ng, RD. 208 Te fet oe etn o com cb stem Noh Aros Amato or Er Msg. 10. Tiere aura “EE, 107 Romi clap a Rearing ce 278, NiPTiig eon olan SOY Lio coure Fey: Outta, Moh, Basen, Baa, Bea “stu J, el, A, 2018 i nd conten ge ws saa thames roto growth trol fore sl Appl 26, 1273-128, Line A Sir, 20:2 Rae enti in fort fren itn isnt ae oe of Ride an, re. Nw os 8, 39-90. Ieper 10 te FS Loc carer, CP bch, P Rar, Caso, R Macs , Meer, 1, Pome, Swit Nie, bl, 208. Tope wean a cst one By et: ar a 27S Jer NOT oe Lobbec At Port Ls Mates Raos M. Boge alt NJ, 205, ns i Sen Eoin 12421352 hse Jer To Te90 THOT ET ta Dror ©, News, 200 Linking amass sonar suceion oe fee ot rr ek ey a 1a Dy Maa Peon, Morn 200. Cain of eon fost ‘agin th ran fon nl orl Mang. 090131 oe 0 Siena sare aosocees Lucha A Bar, Kup, TA, Crna Cots, FY. Alen, CB 907. A ‘the litany betwen rade dene ond ener ep et Sth gucrorn a satus ete Sons 60, 13 ges? ‘ouor/10igrsnst4a57n0n ace, Dinn,y Mein,P, Neptad , 010,Naet Len s Tessin ser at rng ghy Watered ai i coon Poe by bec Socay of Ames NUTRIENT 15S aD EDSTRIOUTION APR FOREST CHEARNG ON A IGHIY WEATHERED SO1IN rn Bogie! Renmin nsdn nels st Ry Coere: Sty ‘Srila Rest, Wags, Be Mes Teply A, Goss, £8, Hern, Ml an 1, Wig abl Sh Cn Antiwar Teen kl 208 Role ie ‘nia mepe! oentn cng wer ei espns New Pa 219, Sere pe coorg Lor tv aston al; Ha Boman IAL, Jon Jom, 2, Monts, Maen, DA Di7 lel ew pt an, tesa cle: ae eaten recite ONE 7 eae de lana BL, de Ml, A.C, Sings, CR, 2011 Ea bn pare nat ds amc: rons retort Renan kiore 75-008 peso ers. ee20 10D Mora Ero, Tarkan 3M aba Foners MC, MeCranan, SF1, 200, fo erty an lands on oes scion nanan Fo, Et Mange 155, 92-08, pe Jonny or ove snare T2700 ‘Simon, St, 26, Caron and nai estou iter a eso sages of Catia dey feat a Noten Balt Nu, Cc Bonar $55 2 psa To too sloaorear7 Naka, 82008, pees ete an greaton eo he prod of oder plat eos unr cowpea, Bsa Ho 1 120185, ny lotto ona sonora tant ‘eton BW, Meng Ree JG, Gla Adio De Sous, 8, Tea Bat, 1 ovine Cot, 199. Alomei een or improved sma fe ‘ny ors om inthe cel Arwen. Fr Esl Manat 117 149-62 pw /doag/too/sosretta7 sens NontiaJiir R, Eng VL Part JA, de hl, ACG. Re DS, 2014 fore sandstones 4 here 101390 eae ea ‘aman: Xa $3,211. The ate of lant css on spe sai. Er Eng. 37, 19-10 pkey 101010 coe 201008002 Per KS, Minoo, BH, Mino, BS El, de av J, Fe es, Has en Neve, EPs, RCS. Mah 2017 Uneven oe “Since he Aso ran tn lat 3 ‘har corde eet uct loge uteaion At Og 62 S551, Wie ato 7 0800 Peres, Lt de And EM, de Quo Pa UA, Rayer, PCL, lo, LCR eka, 3 Cat seks a ope dye a a et Cea ‘Aarne 7, 2-40, hip oan 10935 100 eon, cies fr Pa cium here acer ger Ee Ros Der 3, S459 p71 101080079450 ano, ‘oor angers FA, TL, Aimed Zambrano, AM, Blane, P Beck Shi Butt Sansion P45 Brains EN, Chsdon, Ry Graven De Be ‘iCrea 3, Cara GAL, DeJong B43, Delo, 3 Det Bi Deva Due Barn. SM Ept-Sna, MA, Finda Mi la, ita ds Manda Sato una Cer uqaco, A, Reed, Beer, SG, Leo. 3, tanec, Marna Mee Rabon, Misc P awe, JA: Mequt,B Moa Mund Mee Nae, rere EA. Ht, b, Power 1S” Rodriguez Vez. ee Fee, Este J aden, Sue Aaa, Scat Nb Sng MK Scone, Usa, Man Bega Der Wale oy Rorndn Dai 201 mms iene of Nerepc ondny fren:Ne 530, 21216 holon 0.103812 ‘2c evra formation Stem Open Sue essai Foundation Proj eon S442 hip apsorrone ca Team, 2017.8 ate nd vient fer Stil Compt, eon Si. m oundain fr Sats compaug Vn, Ast 6.08 Vi Si EM, Slo a Een mn BV 1n-aa pe over to lon 0lberTeolaaiOn sii ag, ME, Za eA 28 Pose site selon is nt nat ‘oaring curl eerie ate oe ern Se. ohn, $8, von Gen Ber. Meets, GS. Ont. X20. Fc ntening tori cary uct er ad cent carbon ost in ran Atom Fore ‘Rodin BAUR, Se Fo, Schaefer RVs, Silay A, Lae, ABD. Rohr, Dai J. Raj Re Ratha, R218 The ha of piel bt ing cleanin aC, Change ps org 10.1087 Sis os ani» Roig, RR, Lins, RAF, Gado S, Nav, AC, 200, On the rstortn of ih 1a, ates hp nny 101016 acon 08 1208. Rost, 2018 bet =f aman cdo a I mparan fr hon sore, ena Forel 21960-1963 Hion solr 10.5902 SOSA, Going reriten Re Ea 2, 156167 pe kay vice Rosle NE le, 8, 2019. Pant Tras Rae than Species Ress Expain nang eeme! in Sipe Fone Erat CAE poser Iorfiwoat greene Soar 3, Wey Dey Tn ALL, UC 1988 Lange Ciconegeece of Fore econ i the Upc ho Nero of Combis Venice Ee 7, Sb. haps dtoa/10.25 ance Suse, J, PtP, Ste Tapa, Ay Braga, SMA, Rods, PEP Shr Pat noe nd Colgate nan aia ie Foe: impletons psa reseraton New Forest 8 373-30, Mino 10 itivrtiowsi ones som JP CE, Ar, LS, Due LV. Ma LC, Gua FF. Stet, LS. ‘Sharan 35, 2008 ore band SAR spd the bobegrund ams Ste ihe raan opel afro Reowte Sos Ene, 87 482-55 fp ovrg othe 20a 13.0 sb. c,d, A 2014 afc of ao tp above biomass accion ao ton ss alee paren Slutht eaa Bo Bag 72 aT hge’ ang iohe sey es egy and Managemen 482 (2020) 117972 002 Eel Eval hel 0 Lotwsareo AT (oaaN Se8 Seric Florotal rasr.Fres o Brn om reo 2013 dos de 207 2012 Serco Peal Bat, ~ Bele SE. 18, 2013, vale 7 22Seteizadoennovenbo 515 ceed 31 Sanur 2019). Shima, Boron, Pc Maris LCM. 24 How mchcaton eee hag neta faa at? xia ase specie Bes ‘nfo or, eo Manage 39, op“ 1 ne 0 Sills Vii LG, Shi, Lamb, Cava, IR, De Fg, RO. ‘ont 5, Buison E,S0 Stores an pss aoe neatly tect ee pcs dvey ia rpc oe egronth tector. Econ 9. ‘atts ind cree menses Fr Ek Mage. ps0 101) Sava, Ve de, Stn, 1R ds Gav LS, Sv, RD. our, YM, 2016 Pode a reo Aazorian pinay oe nds eros o cgay ‘Seecion. Ata Aronia 81ST hips /Soor 1015901608 Siva HE Ribein, SC, Bodo, SA, Fa, RAYS, Tee MBR, Nel, J ‘eturaca Boreal on Sis Cra Senn Fora Fre Snes 3, Ses 8 tn do tbo rao Asn 18 “unr na Regio de Maas (A) Pub Dasa, ia Nocial runt a Anti & Pndneto neride Fede So Anim aa Soars a, R Moced, M,C, A Ct, WG Medi Hy ‘ner A Zoe Gchng Bil Foren Code Scans 94388 ete/a tr scence Tae soutn', Fenn, 291. Ane «racer ds ici abe 0 Sueigre, MK. 2, Sonar fost ste ad oma flo hand ene le cena and southern Trop Bl, 68-708, Fite scorpio sosscr aaa “aa, Santon, RE etn, Ping ML Vrhyen, K 018. So ropes ‘ed nkitarng st oe aft above soe! Doms a tel Sango lt dg ep rot rn Aa eg Sea, aS ie S/ Tox), 2 Wein integnie Lirate Rei Us the Fst sn Pret Foie fe Fre Harm, avr Dew Re hips or 101077 (uC Basibct, Ry Seo, EAS, 18 Abandoned Paste Eastern Anan iS Bmereont Saremon. The Eo oe cb. p= neo. Ua, Atle, MDa Sa, MET, Rabi, hen, Bru, EM. 201 ‘nt he versa urd ang dan sed peel ty bd i x fem ngs ian 8-7 eh OH vn de Gant Sea R, ae, M201, The contin of ocd ce Poe aes cat hsge ge: Cmte Pay 8, -8. Frm /suar/10 1080 somo ote 12tae ena St, Sn, CP, Carano Leen Oe, AA, 2912 ‘ts ns sco no Palo Aine do nds de Sto Ps, Bs Bt Nesp 121814 peony 1 15006760829 DOE ‘Guerrero, J.B, 2003. Cassifving successional forests wsing Landsat spectral propes- ts and ecole neater Alena Se Eni. 7. srg SA, Ales LA MAS, Bake, Bat, JL, Camp, ‘imaro,P,Gv J. Deli WB, Figue N Hoor, Jay. Kal ‘eMail ek de ston EA, Mehr Pip, Oy dr Sato FAM, Sinutro- Sve tore SE” 3008 Eaton of ome a Carton sk hc ft Aa oe Bla Nr 2 trai ser ost anonsO Vand, Persie Pi, 2013 Stony Yee a cena ati: Lad heh eect onsboveound Boas, Fo Eco Manage 47 140-68 Mori Ha JA one M202 rating slog! ‘ Tee of the erate Rest. Ee 21 357-54 ip /dolog NTTT Tost Woh, 2002, verity in tpl oes: A eve och of pc ovate. Octo W/o a0 1020s 908 ‘ani Aa std atone esta ora com Sans da na aa anc. sae Baton Us ofS Pt atin Do, Hats Laci A, Asc, Herts ML, Azevedo Rate, ‘ore, hargoo, BA Alena AC, Cabin Allee A Popo, P Faro hi, ahr W'S, Ses, Traber an. ple Ga ete emis ton topical ean drop by 30% 8 8 Year? Gh Change Bl B17 ono 10 TNTN S