REVIEW ARTICLE Disentangling the importance of intrinsic and extrinsic seed dispersal factors in forest restoration success: a global review Miriam Selwyn1,2,3 , Joan Pino1,2, Josep Maria Espelta1,2 Animal-mediated seed dispersal is envisaged as a key process to promote forest restoration success. Yet, we lack a comprehensive knowledge of the contribution of intrinsic and extrinsic factors which directly and indirectly influence this process and, ultimately, restoration outcomes. Here, we reviewed 157 articles to identify the most relevant intrinsic (e.g., plant and animal species traits) and extrinsic (e.g., connectivity) factors which mediate the seed dispersal process. We analyzed the contribution of such factors to restoration success, as identified in the reviewed articles, globally and for tropical and temperate biomes. Generally, our analysis revealed that the main factors affecting restoration success were the extrinsic ones (i.e., connectivity, type of restoration conducted, forest protection, and disturbance degree), while the ability of animals to achieve long-distance dispersal was the main intrinsic factor. Differences among biomes were observed for extrinsic factors, specifically for restoration approach, as more intense efforts seem to be needed for successful restoration in tropical systems (i.e., reconstructive restoration), whereas assisted restoration (e.g., perch placements) was mostly enough for temperate ones. Within intrinsic factors, rodent frugivory and “large-sized” seeds were related to unsuccessful restoration in tropical forests, while frugivory by carnivores and “small-sized” seeds favored resto- ration in temperate ones, and the main constraint to success in this biome was ungulate frugivory. Our results highlight the impor- tance of distinguishing between intrinsic and extrinsic factors mediating animal seed dispersal in forest restoration in different biomes. Furthermore, this paper will help to promote adequate measures when planning restoration actions. Key words: active and passive restoration, frugivory, functional restoration, plant–animal interactions, tropical and temperate forests forests has become a major global concern, underpinning inter- Conceptual Implications national initiatives to reverse degradation and destruction of for- • Animal-mediated seed dispersal is a key process to ensure est ecosystems worldwide (e.g., United Nations Decade on forest restoration success in both tropical and temperate Ecosystem Restoration (2021–2030), Aichi Biodiversity Tar- biomes. gets, E.U. Forest Strategy, E.U. Green Deal) (Crouzeilles • Differences among biomes in the importance of intrinsic et al. 2016; Meli et al. 2017; Aronson et al. 2020). Although a and extrinsic dispersal factors influencing seed dispersal critical need to assess the outcomes of forest restoration actions must be considered to specifically apply adequate restora- tion strategies regionally. • Plant dispersal syndromes and seed disperser traits are the Author contributions: MS, JP, JME conceived and designed the research; MS performed the review search, acquisition and analysis of the data and wrote the most relevant intrinsic factors determining restoration manuscript; JP, JME verified the analytical method, contributed to the improvement success stressing the importance of actions to ensure their and edition of the manuscript; MS, JP, JME discussed the results and contributed to the final manuscript. matching. [Correction added on 2 February 2023, after first online publication: The affiliations • Our review suggests that strategies differing in their effort have been renumbered in the affiliation section.] may be needed in temperate and tropical biomes to ensure 1 Universitat Autònoma de Barcelona, E08193, Bellaterra (Cerdanyola del Vallès), restoration success. Catalonia, Spain 2 CREAF, E08193, Bellaterra (Cerdanyola del Vallès), Catalonia, Spain 3 Address correspondence to M. Selwyn, email m.selwyn@creaf.uab.cat Introduction © 2023 The Authors. Restoration Ecology published by Wiley Periodicals LLC on Habitat destruction and land use changes over the past centuries behalf of Society for Ecological Restoration. have contributed to the loss and fragmentation of landscapes that This is an open access article under the terms of the Creative Commons Attribution- NonCommercial-NoDerivs License, which permits use and distribution in any medium, show a substantial impact on habitat structure and biodiversity, provided the original work is properly cited, the use is non-commercial and no leading to severe ecological disruptions and to the loss of impor- modifications or adaptations are made. doi: 10.1111/rec.13868 tant ecosystem processes and services (Kurten 2013; Bello Supporting information at: et al. 2015). Society’s current awareness of its dependence on http://onlinelibrary.wiley.com/doi/10.1111/rec.13868/suppinfo May 2023 Restoration Ecology Vol. 31, No. 4, e13868 1 of 12 1526100x, 2023, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/rec.13868 by Readcube (Labtiva Inc.), Wiley Online Library on [04/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Intrinsic versus extrinsic factors in restoration has been repeatedly claimed, the global forest restoration com- intrinsic factors as they will also directly determine the strength munity still lacks a comprehensive review for the evaluation of this plant–animal interaction and thus, the capacity of seeds and verification of the efficacy of such initiatives (FAO 2020). to reach restoration sites (Orrock et al. 2006; Kuprewicz 2013; Evaluating restoration is not straightforward, with extensive Naoe et al. 2018). On the other hand, “extrinsic” factors have been debates surrounding what characterizes successful restoration defined as those indirect factors that may (or may not) affect ani- and how it is best to measure (Wortley et al. 2013). Usually, mal seed dispersal, and ultimately restoration success through attempts to evaluate the success in recovering disturbed habitats their potential effects on the disperser community composition have focused on compositional and structural aspects of biodiver- and behavior (Molin et al. 2018; Wolfe et al. 2019). These extrin- sity such as species richness and abundance (Ruiz-Jaen & sic factors are related to the environment, including the amount Aide 2005; Forup et al. 2008; Brudvig 2011) assuming that they and proximity of forest cover (i.e., connectivity), microsite avail- can be proxies for inferring the extent of the recovery of ecological ability, conditions for seed deposition and climatic constraints. functions (Majer & Nichols 1998; Young 2000). Following the Other extrinsic factors may be related to human intervention, such basic guidelines for restoration planning provided by SER as past land use type, restoration approach (i.e., passive vs. active) (2021), three general attributes (i.e., vegetation structure, species and context (e.g., disturbance degree [i.e., initial conditions of the diversity and abundance, and ecological processes) have been pro- study sites]), time since restoration and forest protection, which posed to help practitioners identify appropriate indicators for resto- can also influence restoration success (Crouzeilles et al. 2016; ration success (Ruiz-Jaen & Aide 2005). Thus, an alternative is to Molin et al. 2018; Koelemeijer et al. 2021). evaluate the re-establishment of complex ecological interactions Finally, the effects of intrinsic and extrinsic dispersal factors in and the ecological functions that they provide as the starting point determining restoration success may potentially differ across for rebuilding community structure (Forup et al. 2008; Ribeiro da biomes, as they are subjected to different biotic, abiotic and socio- Silva et al. 2015). One of such functions is seed dispersal, which economic conditions which shape fauna and flora composition, may play a key role in ecosystem restoration, transporting seeds land use and management types and, even, restoration approaches into disturbed sites and linking the end of the reproductive cycle (Clark et al. 1999; Aide et al. 2013). Regarding intrinsic factors, dif- of source adult plants and their subsequent life cycle stages ferences may relate to dispersal agents’ composition, animal seed (Bakker et al. 1996; Jordano & Godoy 2002; Forget et al. 2011). dispersal distance ability or seed size (Hanya & Aiba 2010). While Within seed dispersal mechanisms, we chose to focus on animal- for extrinsic factors it would be likely to a priori assume similar mediated seed dispersal, provided by the interactions between effects within biome types, yet it cannot be discarded that differ- fruiting plants and frugivorous animals (Jordano & Godoy 2002; ences may arise from the actions performed (i.e., type of distur- Forget et al. 2011). This dispersal type has been acknowledged bance, forest protection, restoration approach) which would to largely influence plant distribution patterns by facilitating the ultimately affect restoration success. establishment of seeds in new habitats away from the mother plant Thus, a global analysis of forest restoration outcomes deter- (Traveset et al. 2007; Rubalcava-Castillo et al. 2021). Moreover, in mined by intrinsic and extrinsic dispersal factors, which to our many plant communities, seed dispersal by animals is often the knowledge, have seldom been surveyed simultaneously, is main form of dissemination of propagules and thus, essential for appropriate given that most case studies have mostly focused the long-term maintenance of forests where most plant species rely on specific forest types (de la Peña-Domene et al. 2014; Mittel- on this mechanism rather than wind, water, or other forms of dis- man et al. 2020). persal (Forup et al. 2008; Viani et al. 2015; Rubalcava-Castillo Here, we conducted a literature review to assess the role of et al. 2021). For example, on average, 90% of woody plant species animal-mediated seed dispersal in forest restoration efforts in tropical forests and up to 62% of temperate forest species pro- worldwide, and independently for tropical and temperate duce fruits that are dispersed by animals, usually birds and mam- biomes. Specifically, we determined which intrinsic and extrin- mals (Howe & Smallwood 1982; Herrera 1984). sic factors were assessed in the restoration studies reviewed and Analyzing the factors affecting seed arrival and survival in then, we evaluated their importance for the outcome of restora- restored forests can provide a useful tool for evaluating and mon- tion globally and across biomes. To this end, we addressed itoring the recovery of seed dispersal functions and thus, forest two main questions: (1) which is the relative contribution of restoration (Tylianakis et al. 2010). On the one hand, seed dis- intrinsic and extrinsic dispersal factors in forest restoration suc- persal can be directly affected by several “intrinsic” factors, which cess? and (2) does the importance of such factors in the restora- relate to the traits of the dispersed plants and their dispersal agents tion outcome differ among forest regions (i.e., tropical (e.g., species life-history traits, animal behavior) as well as spe- vs. temperate biomes)? The answers will be important for prior- cies richness and abundance (Schupp et al. 2019). These intrinsic itizing restoration actions and for their smart design in the recov- factors include fruit handling methods (e.g., legitimate seed dis- ery of different ecosystem types. persal by animals that swallow entire seeds and defecate or regur- gitate them in viable conditions for germination, sensu Traveset 1994, seed predation, scatter hoarding), animal seed dis- Methods persal distance ability and plant dispersal traits (e.g., seed size, amount of fruit reward) (Baños-Villalba et al. 2017; Li Literature Search et al. 2019). In addition, the richness and abundance of the dis- We searched for published studies of forest restoration that persed plants and dispersal agents have been considered as focused on seed dispersal functions as a key ecological process 2 of 12 Restoration Ecology May 2023 1526100x, 2023, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/rec.13868 by Readcube (Labtiva Inc.), Wiley Online Library on [04/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Intrinsic versus extrinsic factors in restoration towards ecosystem restoration. For this review we followed reference details, study period, region, country, continent, patch Pickering and Byrne’s (2014) recommendations for systematic size (ha), and biome type. As biome information was not stan- quantitative approaches to literature reviews, and Romanelli dardized across papers, we classified the study cases for climate et al. (2021) for repeatability of reviews of restoration outcomes. based on the geographic position of the study site using the Two global databases were selected to carry out our literature global Köppen-Geiger climate classification map (http:// search: Web of Science and Science Direct. For both databases koeppen-geiger.vu-wien.ac.at/present.htm). we searched for the following combination of words in the title, Intrinsic dispersal factors were considered those related to the abstract or keywords: seed dispersal AND (forest OR ecosystem attributes of the dispersed plants and their dispersal agents OR ecological) AND (restoration OR recovery OR regenera- which directly affect the outcomes of seed arrival and survival, tion). The search was limited to the above-mentioned terms to while extrinsic factors were those related to habitat conditions target papers not only focusing on active restoration strategies or restoration approaches which could indirectly affect the seed (i.e., restoration, recovery) but also passive (i.e., regeneration). dispersal process and thus, forest restoration success (see Fig. 1 We chose to include the aforesaid terms rather than others such and Table S1 for full variable list and descriptions). The qualita- as “rehabilitation” to be more consistent with the terms provided tive attributes used for intrinsic and extrinsic dispersal factors’ in the SER definition of ecological restoration (SER 2021). We classification were based on the description and terms provided also chose to include “forest” rather than “woodland” or “tim- in the articles selected for the review. A total of 27 indicators berland” as our experience with the literature and keywords (15 intrinsic, 12 extrinsic) were selected to evaluate their effects search showed that most papers that use these other terms also on the restoration outcome (Table S1). included “forest” and would be acquired by the above search. Within extrinsic factors, restoration approaches were classi- We obtained 600 articles matching the selected keywords. Then, fied into (1) “passive,” when after cessation of the disturbance, we excluded all publications that did not describe forest restora- natural (i.e., spontaneous) regeneration proceeds (Gann tion processes by means of seed dispersal interactions. More et al. 2019); or (2) “active,” when the potential regeneration is precisely, this scrutiny met the following established criteria: mediated by human intervention. Active restoration was further (1) the main objective of the study was the assessment of seed categorized into: (i) “assisted regeneration,” when interventions dispersal functions in disturbed landscapes affected either by harness the regeneration capacity of the remaining biota active or passive restoration, (2) the study included a compari- (e.g., removal of alien species) or they aim to reintroduce species son to either a reference ecosystem (forest similar to the project that cannot migrate into restoration sites without assistance by, site but nondegraded) or control (degraded forest without resto- for instance, installing resources to prompt colonization ration treatments) used for the definition of restoration success (e.g., perches), or (ii) “reconstruction,” when major proportions (see definition below), and (3) the articles included information of biota are introduced (e.g., reintroduction of early successional of the specific intrinsic and extrinsic dispersal factors used for species) (Gann et al. 2019). When possible, for each factor’s cate- the assessment of the restoration success (see next section). gory, the exact terms provided in the reviewed articles were used We obtained 157 articles published between 1996 and 2021 (e.g., disturbance category: “deforestation,” “fragmentation,” which matched the above-mentioned criteria. When the same “defaunation”). But for uniformity, some variables were converted article presented results from two or more study locations that into simpler ones, based on the information compiled from each were sufficiently distant (e.g., different continents) and sub- study (e.g., for seed size: “large,” “medium,” “small”; for animal jected to clearly differentiated conditions (e.g., biome type), seed dispersal distance ability: “long,” “medium,” “short”; for for- the results of each site were considered as independent cases, est cover: “continuous,” frequent,” “fragmented”; see Supplement resulting in a total of 161 items. Thus, the database compiled S2). Similarly, categories represented by factors limiting or favor- for this study is new and based on the literature search conducted ing seed dispersal and survival (e.g., distance to reference forest) specifically for this study. The greatest number of works were were binary classified as one (1) indicating a limitation on dis- conducted in tropical and subtropical forests (N = 94 and persal, or zero (0) indicating no limitation (see Supplement S2). 23, respectively), while temperate and mediterranean-type cli- Finally, for the purpose of this study, and although there is exten- mate forests accounted for 22 and 17 publications, respectively. sive debate on what defines restoration success (Wortley We found three assessments in steppe forests and one in savanna et al. 2013), the degree of forest restoration success (i.e., “high,” and desert forests. Publications were found in studies conducted “medium,” “low,” or “unsuccessful”) was established according in 40 countries and the ones with higher number of publications to the judgment reported by the authors of each publication on were Brazil (N = 35), Spain (N = 16), Mexico (N = 13), and the level of recovery of the ecological characteristics addressed in China (N = 12) (see Supplement S1). the particular study (e.g., vegetation structure, species diversity and abundance or ecosystem functioning) (Ruiz-Jaen & Aide 2005). From the four categories of restoration success identi- Rationale for Parameters and Categories Used in Classification fied in the primary studies, we then constructed a binary response The information recorded was obtained from each primary study variable which accounted for those studies that comprised “high” and was classified into general information and intrinsic and or “medium” (“YES”) and “low” or “unsuccessful” (“NO”) resto- extrinsic dispersal factors involved in the restoration process ration success to increase the robustness of our further analysis (see Supplement S2 for detailed description of the database (Cutler et al. 2007). See Supplement S2 for a more detailed ratio- and how it was compiled). General information included full nale of the classification of forest restoration success. May 2023 Restoration Ecology 3 of 12 1526100x, 2023, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/rec.13868 by Readcube (Labtiva Inc.), Wiley Online Library on [04/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Intrinsic versus extrinsic factors in restoration Figure 1. Main intrinsic and extrinsic dispersal factors used to evaluate forest restoration success by means of animal-mediated seed dispersal. See Supplement S2 for complete list of variables and descriptions. Data Analysis which indicate how accurate the CTA classification was per- The data analysis was divided into two steps; the first one to formed (Cutler et al. 2007). characterize the dataset and the second one to investigate which The above-mentioned analyses were conducted globally. Yet, intrinsic and extrinsic dispersal factors promote restoration suc- to detect differences between the most studied biomes, they cess globally and individually for tropical and temperate were separately carried out for tropical and subtropical (hereaf- biomes. Accordingly, to characterize the dataset, we first evalu- ter tropical), and temperate and mediterranean (hereafter tem- ated the distribution of case studies by continent and type of perate) systems following The Köppen Climate Classification. biome, the progression of publications over time, the percentage All analyses were performed using free software RStudio of cases within each type of disturbance category and the main (v. 1.4.1106) and packages within: dplyr (Mailund 2019), tidy- intrinsic and extrinsic factors used to assess restoration success verse (Wickham & Wickham 2017), rpart (Therneau in each case study. Second, to investigate how intrinsic and et al. 2015), and FactoMineR (Husson et al. 2016). extrinsic dispersal factors may drive restoration success, we con- ducted two complementary analyses. First, two-way correspon- dence analyses (CA) were performed to explore the association Results between the restoration outcome and the intrinsic and extrinsic dispersal factors identified. For this test, quality of representa- Characterization of the Dataset tion tests (measured by the square cosine, cos2) were carried Worldwide, few works were published between 1996 and 2005 out to measure the degree of association between the variables (N = 26) regarding forest restoration by means of seed dispersal; and a particular axis (i.e., how good the representation of the however, published works increased considerably between variables on each component is) and variable contribution tests 2006 and 2021 (N = 135). This trend showed some differences (measured by Contrib function) were performed to filter the between published works for tropical and temperate biomes: the most representative variables involved in the success of the res- former was present since 1995 and accounted for a total number toration process (Husson et al. 2016). Second, classification of 117 case studies until nowadays, whereas publications for the tree analyses (CTA) were performed as complementary anal- later mostly begun in 2002 and accounted for a total number of ysis to further explore the associations obtained in the previ- 39 case studies (see Supplement S3). The disturbance categories ous CA. For the CTA tests, class proportions allowed us to which caused forest degradation and decline prior to the restora- explore the relation between the predictor intrinsic and extrin- tion efforts showed that 99% of the case studies held forest frag- sic dispersal variables, inferring which ones were more rele- mentation, 87% were previously deforested and 28% were vant in the overall probability of restoration success defaunated. This scenario was mainly due to nine types of dis- (Therneau et al. 2015). The CTA tests were built by the rule turbance categories, being agricultural practices, the presence of binary partitioning in respect of the broader category of of alien species, inappropriate grazing and pasture lands, and restoration success (“YES/NO”) response variable (see game the most frequent ones (see Supplement S3). above). Finally, the Gini index was used to split and optimize Regardless of the type of biome, extrinsic factors were the nodes and to perform quality of representation tests, slightly more addressed in the reviewed articles (Supplement 4 of 12 Restoration Ecology May 2023 1526100x, 2023, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/rec.13868 by Readcube (Labtiva Inc.), Wiley Online Library on [04/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Intrinsic versus extrinsic factors in restoration S3; Table S2). Moreover, all extrinsic factors were similarly to “unsuccessful” restoration were game and passive restoration used aside from microenvironmental conditions and game. Con- approach, while ungulate frugivory was the main intrinsic fac- trarily, among intrinsic dispersal factors, disperser richness, animal tor (Fig. 4). seed dispersal distance ability, seed size and number of seeds dis- For all biome types, the classification tree layout showed that persed were the most addressed variables (Supplement S3; the overall probability of restoration success was 81% Table S2). (Supplement S3) and that the quality of representation of the CTA was 92%. Out of the extrinsic factors checked, the ones with major importance involving a positive restoration outcome, Drivers of Restoration Success defined by the Gini index, were close distance to reference hab- Globally, the two-way CA explained 92.6% of the cumulative itat, “improvement” microenvironmental conditions, higher res- variance (χ 2 = 536.90; p = 2.62  10 20) in the patterns of res- toration age, not being subjected to seed attack, assisted plant toration success, including all intrinsic and extrinsic dispersal restoration approach and “frequent” or “continuous” forest factors. The cos2 quality of representation test showed that most cover. Furthermore, the most important intrinsic dispersal fac- variables were highly represented by these two dimensions tors involving effective restoration were “medium” seed size, (Supplement S4). Extrinsic factors which contributed the most “high” or “medium” plant species richness and “medium” ani- to the definition of dimensions and that were associated to a mal seed dispersers’ abundance (Supplement S5). “high” and “medium” restoration success were, in decreasing For tropical systems, the overall probability of restoration order of contribution: time since restoration was initiated, “fre- success was 82% and the quality of representation was 93%, quent” forest cover, “improvement” of microenvironmental whereas for temperate systems, it was 79%, and 90%, respec- conditions, proximity to the reference habitat, not been sub- tively (Supplement S5). Extrinsic factors that showed major jected to seed attacks (e.g., fungi or infestations), “least importance in the effective restoration outcome of tropical for- degraded” sites and protected forest status. Within intrinsic dis- ests were higher restoration age, “improvement” of microenvi- persal factors, animal and plant abundance and richness and ronmental conditions, having a close reference habitat and an “long-distance” dispersal ability contributed the most to restora- assisted or reconstructive restoration approach (Supplement tion success (Fig. 2; Supplement S4). Contrarily, frugivory by S5), while only “frequent” or “continuous” forest cover was rel- rodents contributed to “unsuccessful” restoration efforts evant for temperate biomes (Supplement S5). Regarding intrin- (Fig. 2; Supplement S4). sic factors, for tropical and temperate forests, “medium” and For the two-way CA of tropical (Cumulative % of “long” animal seed dispersal distance ability were the only var. = 88.7; χ 2 = 448.63; p = 9.77  10 12) and temperate sys- intrinsic factors in common favoring restoration success. Across tems (Cumulative % of var. = 84.6; χ 2 = 203.40; p = 0.98), biomes, seed size, bird frugivory type, and “high” and most variables were also highly represented by the first two “medium” plant species richness were important in tropical for- dimensions (Figs. 3 & 4; Supplement S4). For both tropical ests (Supplement S5) while lack of hoarder activity and “high” and temperate biomes (Figs. 3 & 4; Supplement S4), extrinsic animal dispersers abundance were relevant in temperate ones factors which contributed the most to the definition of dimen- (Supplement S5). sions and that were associated with a “high” and “medium” res- toration success were, in decreasing order of contribution: higher restoration age, short distance to reference habitat, Discussion “improvement” microenvironmental conditions, “frequent” for- To our knowledge, this review is the first to assess the contribu- est cover, and “least degraded” sites. In addition, two other tion of different intrinsic and extrinsic factors affecting seed dis- extrinsic factors were further associated to restoration success persal in forest restoration success globally, and separately in in tropical systems: absence of seed attacks and reconstructive tropical and temperate biomes. Interestingly, most extrinsic fac- restoration approach (Fig. 3). For temperate systems, specific tors had similar effects in restoration success worldwide extrinsic factors that contributed to successful restoration were (e.g., forest connectivity, time since restoration), while differ- protected forests and assisted restoration approach, while pas- ences were observed for intrinsic dispersal factors in tropical sive restoration approach contributed to unsuccessful restoration and temperate forests. Specifically, those intrinsic dispersal fac- (Fig. 4). Intrinsic dispersal factors contributing to both tropical tors related to dispersal syndromes and seed disperser traits were and temperate restoration success were “higher” animal and the most relevant in determining restoration in tropical and tem- plant abundance and richness, whereas “long-distance” dis- perate forests. Yet, although our study was based on a thorough persal ability contributed to tropical forests and carnivorous revision of the scientific literature (157 articles), we must and “small-medium” sized seeds were more relevant for temper- acknowledge some limitations owing to the lack of studies com- ate ones (Figs. 3 & 4). Within the variables which contributed paring sites with and without animal-mediated dispersal. the most to a “low” and “unsuccessful” restoration, we found We observed that the number of publications on this topic has few differences between biome types (Figs. 3 & 4). For instance, gradually increased over time from 1996 until nowadays and for tropical biomes, the extrinsic factor that was mostly associ- specifically in tropical forests, whereas for temperate systems, ated to “unsuccessful” restoration was climatic constraints, this trend begun in 2002. This could be due to researchers taking while intrinsic factors were “large-sized” seeds and rodent fru- higher action in forest restoration case studies as commitments givory (Fig. 3). For temperate biomes, extrinsic factors related and initiatives to restore ecosystems and the services that they May 2023 Restoration Ecology 5 of 12 1526100x, 2023, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/rec.13868 by Readcube (Labtiva Inc.), Wiley Online Library on [04/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Intrinsic versus extrinsic factors in restoration Figure 2. Variable contribution for “high,” “medium,” “low,” and “unsuccessful” restoration success (blue labels) in all biomes, defined by two-way correspondence analysis (see Supplement S2 and Table S1 for variable codes and descriptions). Only the 32 variables with the highest contribution are shown. provide are becoming more widespread (Pereira & Africa, and Oceania, that have also experienced forest loss and a Navarro 2015; Ockendon et al. 2018; Romijn et al. 2019). More- rapid decline in forest area (Betts et al. 2017), showed fewer num- over, publications were significantly higher in tropical systems, ber of publications. Future development of national and interna- where up to 90% of trees have traits adapted to animal-mediated tional networks of ecological restoration scientists and seed dispersal (i.e., zoochory, Howe & Smallwood 1982; practitioners may promote such efforts in less considered regions. Howe 2016). Even so, there was a high variability in the number Beside specific socioeconomic attributes of each country and of publications among regions. For instance, publications were region, intrinsic and extrinsic dispersal factors have also shown notably high in Brazil as it is one of the five countries encounter- to determine the result of the entire seed regeneration process ing more than half of the world’s forest, but also the country with (e.g., Vallejo et al. 2012; Crouzeilles et al. 2016; O’Donnell the highest average annual net loss of forest area (FAO 2020). et al. 2018). In these lines, we found that the most used factors This scenario has made Brazil lead South America in public pol- to assess the effects of seed dispersal in forest restoration success icies and legislation aimed at increasing the effectiveness of forest were the extrinsic ones. This could be due to the clear advan- restoration (e.g., Atlantic Forest Restoration Pact) (Moreira da tages of using extrinsic indicators that are easily measured, cost Silva et al. 2017). In contrast, other regions in Europe have also effective, and easily interpretable (Dey & Schweitzer 2014). For experienced forest loss but such rates have declined substantially instance, frequently used factors such as information on restora- following the land abandonment that began in the mid-twentieth tion approaches, protection types and climatic conditions are century (FAO 2020; Palmero-Iniesta et al. 2020). Although forests easily accessible, provided by government agencies and other are expanding in some regions, primary forests continue to disap- organizations involved in natural resource management. pear in Europe (FAO 2020; Sabatini et al. 2020) together with Regarding intrinsic factors, plant species richness and abun- important ecosystem services that they provide (e.g., carbon stor- dance are relatively easy-to-measure indicators of plant species age, high levels of biodiversity, contribution to resilience) diversity (Lindenmayer et al. 2000) and were also widely used. (Watson et al. 2018). Thus, ensuring protection and restoration of Conversely, other intrinsic dispersal factors were not simple or such forests has prompted restoration initiatives in Europe from a inexpensive to measure, and usually required large investments conservation perspective (Sabatini et al. 2020), and together with of time. A good example is the richness and abundance of ani- natural forest expansion has boosted restoration assessments in mal species, and the measurement of dispersed, germinated Europe in the past decade. Other territories in Southeastern Asia, and established seeds (Aavik & Helm 2018; Iijima 2020). 6 of 12 Restoration Ecology May 2023 1526100x, 2023, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/rec.13868 by Readcube (Labtiva Inc.), Wiley Online Library on [04/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Intrinsic versus extrinsic factors in restoration Figure 3. Variable contribution for “high,” “medium,” “low,” and “unsuccessful” restoration success (blue labels) in tropical biomes, defined by two-way correspondence analysis (see Supplement S2 and Table S1 for variable codes and descriptions). Only the 32 variables with the highest contribution are shown. Within intrinsic dispersal factors, our review showed that dis- (Selwyn et al. 2020). Also, considering that hunting mostly perser and plant species richness and abundance highly contrib- occurs where ungulates are present (Rooney et al. 2015), the uted to successful restoration. Perhaps, such taxonomic richness negative effect of game on forest restoration might arise from reflects the diversity of dispersal traits among species (e.g., fruit the link between these two factors. Similarly to Galetti et al. handling methods and gut passage effects) and thus a broader (2015), we observed that although rodents may sometimes act spectrum of seeds to be dispersed in viable conditions for germi- as seed dispersers (Vander Wall et al. 2005), they were mostly nation (Samuels & Levey 2005). Regarding animal seed dis- associated with an unsuccessful restoration outcome in tropical persal distance ability, restoration resulted successful when forests. This could be due to population growth and activity animals dispersing seeds at long distances were involved, espe- increase of rodents owing to the reduced competition with larger cially for tropical forests. This may be due to the presence of dis- seed disperser mammals, because of defaunation in altered land- persers with higher movement capacity (e.g., birds), able to scapes or hunting pressure (Wright et al. 2000; Galetti travel long distances and disperse seeds, affecting the spatial et al. 2015). Also, rodents may differ in the way in which they patterns of seed depositions and, ultimately, recruitment interact with the plants, for instance, squirrels disperse a rela- (Gonzalez-Varo et al. 2019). Interestingly, relevant differences tively high proportion of seeds, whereas mice disperse a smaller were found among tropical and temperate biomes regarding dis- fraction as they tend to predate most seeds in situ (Gomez persal syndromes and frugivory types. For instance, we found et al. 2019). Simultaneously, we also found that “large”-sized that in temperate systems, carnivores highly contributed to seed seeds were more often associated with an “unsuccessful” resto- dispersal, probably as they are known to act as legitimate seed ration outcome in tropical forests. Again, this pattern could be disperser for many species (Gonzalez-Varo et al. 2013; Selwyn related to defaunation and extreme hunting pressure of the larger et al. 2020; Rubalcava-Castillo et al. 2021). Contrarily, large frugivorous mammals and birds responsible of the dispersal of seed consumers (i.e., ungulates) were related to “low” and large sized seeds (Wright et al. 2000; Wotton & Kelly 2012). “unsuccessful” restoration as they usually grind ingested seeds Regarding temperate forests, “small” and “medium-sized” seeds May 2023 Restoration Ecology 7 of 12 1526100x, 2023, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/rec.13868 by Readcube (Labtiva Inc.), Wiley Online Library on [04/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Intrinsic versus extrinsic factors in restoration Figure 4. Variable contribution for “high,” “medium,” “low,” and “unsuccessful” restoration success (blue labels) in temperate biomes, defined by two-way correspondence analysis (see Supplement S2 and Table S1 for variable codes and descriptions). Only the 32 variables with the highest contribution are shown. were related to higher restoration success. This fact may suggest been required to actively accelerate forest regeneration (Shoo that in temperate biomes, seed dispersal is mostly driven by a et al. 2016; Holl 2017; Mittelman et al. 2020) and to recover stable, more generalized avian seed dispersers’ array which feeds on fully functional communities with reasonable investment require- fruits with small seeds that are produced in large quantities and ments (Gann et al. 2019; SER 2021). As for the negative effects are consumed by a wide range of frugivorous birds of passive restoration in temperate forests, there may exist several (Herrera 1984; Howe 1986). These finding underline the impor- limiting factors (e.g., high seed predation, seedling herbivory, and tance of considering the variety of the plant dispersal network stressful microclimatic conditions) which may limit and slow down models reflecting the outcome of restoration success and the par- passive restoration (Rey-Benayas et al. 2015; Cruz-Alonso ticularities of different biomes to undertake the most suitable et al. 2019; Martínez-Muñoz et al. 2019) and should be taken into restoration approaches in different forest ecosystems. consideration. Although these results suggest the advantage of Beyond the relevance of intrinsic dispersal factors, the analy- active against passive restoration interventions, trade-offs between sis of the extrinsic factors related to socioeconomic characteris- different interests and restoration objectives need to be examined to tics of each study site allowed us to detect that forest protection select the best restoration practices (e.g., Prach & del Moral 2015; played a fundamental role in achieving restoration in forests Meli et al. 2017). worldwide, as it may be crucial for promoting the functioning Furthermore, we found that studies covering forests with in regenerating and managed forests (Sabatini et al. 2020). lower restoration time periods, independently of their biome Regarding forest restoration approaches, differences were found type, tended to show “low” or “unsuccessful” restoration suc- between tropical and temperate forests: successful restoration cess suggesting that restoring conditions closer to those of refer- occurred in temperate forests in which “assisted” restoration ence forests face many challenges, not the least of which is the took place (e.g., non-native species removal, clear-cuts), while long timeframes involved (Sabatini et al. 2020). Longer restora- “reconstructive” restoration (e.g., reforestation) was the most tion timeframes entail higher plant abundance, and thus, fruit successful option in tropical ones. For both biome types, these resource, for which animal dispersers would be expected to active restoration approaches suggest that these may have been spend more time in such areas, increasing their dispersal capac- highly disturbed areas, with a significant extent of deforestation, ity (Crouzeilles et al. 2016) and comprising cascading effects on in combination with inadequate ecological conditions, resulting restoration success. Long-term monitoring and evaluation of in slow or no recovery and thus, human intervention may have restoration actions will therefore provide higher quality to 8 of 12 Restoration Ecology May 2023 1526100x, 2023, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/rec.13868 by Readcube (Labtiva Inc.), Wiley Online Library on [04/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Intrinsic versus extrinsic factors in restoration restoration assessments as the effects of restoration timeframes factors may contribute to such process globally, and specifically, have shown to contribute to restoring forest complexity. for both tropical and temperate forests. While similar outcomes In addition, other extrinsic factors related to the landscape were observed regarding most extrinsic factors, some major dif- scale contributed the most to restoration for all biome types. This ferences were found within intrinsic dispersal factor across trop- is the case of forest cover and connectivity, which were the major ical and temperate biomes, suggesting the outermost importance factors limiting or favoring seed dispersal. Such factors are pos- of the variety of dispersal syndromes and traits in forest restora- itively correlated with the disperser community abundance and tion. Thus, our findings provide valuable information on the most composition (Bael et al. 2013; Falcão et al. 2018), and many significant intrinsic and extrinsic dispersal factors influencing authors suggest the benefits for seed rain and seedling recruit- restoration success, while also informing on potential differences ment (Chazdon et al. 2009; Tambosi et al. 2014). Accordingly, in the best restoration strategy to be applied in temperate and trop- assuring connectivity between reference habitats and restoration ical biomes to achieve restoration goals. However, we acknowl- sites would be essential as the lack of propagules could impede edge that distinguishing between levels of restoration success is restoration of taxonomic and functional diversity (Hewitt & Kell- a challenging task and future empirical research would be essen- man 2002; Ibañez et al. 2014). In this line, as expected, our anal- tial for understanding the full benefits of including animal- ysis also revealed that “highly degraded” sites often showed mediated seed dispersal in forest restoration. “unsuccessful” restoration for all biome types, as duration and severity of the disturbance reduce propagule availability and cre- ate competitive environmental conditions for regeneration ACKNOWLEDGMENTS (Martínez-Ramos et al. 2016). In addition, environmental effects This study was financially supported by the Spanish Ministry of at a more local scale were also found, as microenvironmental Science, Innovation and Universities (MCIU) by the NEWFOR- conditions of seed deposition sites showed major importance LAND project (RTI2018-099397-B-C22) and benefited from for the restoration outcome. This fact could mainly be due to the framework of the wildE Horizon Europe Project (GAP- stressful conditions such as high light and temperature levels 101081251). and poor soils quality in disturbed sites, which limit seeding sur- vival and growth, particularly of late successional species which are adapted to regenerate in the forest’s understory (Holl 2012). LITERATURE CITED Also, seed attacks, particularly in tropical biomes, seem to limit Aavik T, Helm A (2018) Restoration of plant species and genetic diversity restoration success. This is not surprising as, for instance, post- depends on landscape-scale dispersal. Restoration Ecology 26:S92–S102. dispersal predation by insects and seed-infecting fungi are some https://doi.org/10.1111/rec.12634 of the main drivers of seed mortality in tropical forests Aide TM, Clark ML, Grau HR, L opez-Carr D, Levy MA, Redo D, Bonilla- (Hulme 1998; Sarmiento et al. 2017; Kulikowski et al. 2022). Moheno M, Riner G, Andrade-Núñez MJ, Muñiz M (2013) Deforestation Our results suggest that to manage ecosystem restoration, and reforestation of Latin America and the Caribbean (2001–2010). Biotro- pica 45:262–271. https://doi.org/10.1111/j.1744-7429.2012.00908.x plant animal interactions need to be addressed; specifically, Aronson J, Goodwin N, Orlando L, Eisenberg C, Cross AT (2020) A world of accounting for the abundance of frugivorous animals as well possibilities: six restoration strategies to support the United Nation’s as fruit resource availability should be considered as indicators Decade on Ecosystem Restoration. Restoration Ecology 28:730–736. of the quality of the dispersal function. Moreover, landscape fea- https://doi.org/10.1111/rec.13170 tures such as forest cover (connectivity) and habitat quality need Bael SAV, Zambrano R, Hall JS (2013) Bird communities in forested and human- to be considered to favor animal movements into restoration modified landscapes of Central Panama: a baseline survey for a native spe- sites. Regarding differences among biomes, the results of our cies reforestation treatment. International Journal of Biodiversity Science, Ecosystem Services & Management 9:281–289. https://doi.org/10.1080/ review would recommend that generalist seed dispersers 21513732.2013.842187 (e.g., birds) should be promoted in temperate systems while Bakker JP, Poschlod P, Strykstra RJ, Bekker RM, Thompson K (1996) Seed large seed dispersers (e.g., primates) in tropical ones. Regarding banks and seed dispersal: important topics in restoration ecology. Acta the restoration approach to accomplish these recommendations, Botanica Neerlandica 45:461–490. https://doi.org/10.1111/j.1438-8677. our results indicate that assisted restoration would suffice for 1996.tb00806.x temperate biome while more reconstructive efforts would be Baños-Villalba A, Blanco G, Díaz-Luque JA, Dénes FV, Hiraldo F, Tella JL needed in tropical ecosystems. (2017) Seed dispersal by macaws shapes the landscape of an Amazonian ecosystem. Scientific Reports 7:7373. https://doi.org/10.1038/s41598- Determining the efficacy and effects of seed dispersal on forest 017-07697-5 restoration is a complex task, but one that is essential both for Bello C, Galetti M, Pizo MA, Magnago LFS, Rocha MF, Lima RAF, Peres CA, evaluating returns on restoration investments, as well as for Ovaskainen O, Jordano P (2015) Defaunation affects carbon storage in improving the quality of future efforts to repair degraded forests. tropical forests. Science Advances 1:e1501105. https://doi.org/10.1126/ As nations across the globe are working to develop forest restora- sciadv.1501105 tion plans (e.g., Aichi Biodiversity Targets, E.U. Forest Strategy) Betts MG, Wolf C, Ripple WJ, Phalan B, Millers KA, Duarte A, Butchart SH, Levi T (2017) Global forest loss disproportionately erodes biodiversity in it is important to use standardized methods for assessments of the intact landscapes. Nature 547:441–444. https://doi.org/10.1038/ restoration outcomes (e.g., Restoration Project Information Shar- nature23285 ing Framework). Here, we highlight the importance of the part Brudvig LA (2011) The restoration of biodiversity: where has research been and that animal-mediated seed dispersal plays in forest ecosystem where does it need to go? American Journal of Botany 98:549–558. https:// recovery, and how different intrinsic and extrinsic dispersal doi.org/10.3732/ajb.1000285 May 2023 Restoration Ecology 9 of 12 1526100x, 2023, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/rec.13868 by Readcube (Labtiva Inc.), Wiley Online Library on [04/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Intrinsic versus extrinsic factors in restoration Chazdon RL, Peres CA, Dent D, Sheil D, Lugo AE, Lamb D, Stork NE, Miller SE Holl KD (2017) Research directions in tropical forest restoration. Annals of the (2009) The potential for species conservation in tropical secondary forests. Missouri Botanical Garden 102:237–250. https://doi.org/10.3417/2016036 Conservation Biology 23:1406–1417. https://doi.org/10.1111/j.1523- Howe HF (1986) Seed dispersal by fruit-eating birds and mammals. Pages 123– 1739.2009.01338.x 183. In: Murray DR (ed) Seed dispersal. Academic Press, Wollongong. Clark JS, Silman M, Kern R, Macklin E, HilleRisLambers J (1999) Seed dispersal near https://doi.org/10.1016/B978-0-12-511900-9.50009-7 and far: patterns across temperate and tropical forests. Ecology 80:1475–1494. Howe HF (2016) Making dispersal syndromes and networks useful in tropical https://doi.org/10.1890/0012-9658(1999)080[1475:SDNAFP]2.0.CO;2 conservation and restoration. Global Ecology and Conservation 1:152– Crouzeilles R, Curran M, Ferreira MS, Lindenmayer DB, Grelle CEV, Rey 178. https://doi.org/10.1016/j.gecco.2016.03.002 Benayas JM (2016) A global meta-analysis on the ecological drivers of for- Howe HF, Smallwood J (1982) Ecology of seed dispersal. Annual Review of est restoration success. Nature Communications 7:11666. https://doi.org/ Ecology and Systematics 13:201–228. https://doi.org/10.1146/annurev.es. 10.1038/ncomms11666 13.110182.001221 Cruz-Alonso V, Ruiz-Benito P, Villar-Salvador P, Rey-Benayas JM (2019) Hulme PE (1998) Post-dispersal seed predation: consequences for plant demog- Long-term recovery of multifunctionality in Mediterranean forests depends raphy and evolution. Perspectives in Plant Ecology, Evolution and System- on restoration strategy and forest type. Journal of Applied Ecology 56:745– atics 1:32–46. https://doi.org/10.1078/1433-8319-00050 757. https://doi.org/10.1111/1365-2664.13340 Husson F, Josse J, Le S, Mazet J, Husson MF (2016) Package ‘FactoMineR’. An Cutler DR, Edwards TC Jr, Beard KH, Cutler A, Hess KT, Gibson J, Lawler JJ R package 96: 698 (2007) Random forests for classification in ecology. Ecology 88:2783– Ibañez I, Katz DSW, Peltier D, Wolf SM, Connor Barrie BT (2014) Assessing the 2792. https://doi.org/10.1890/07-0539.1 integrated effects of landscape fragmentation on plants and plant communi- de la Peña-Domene M, Martinez-Garza C, Palmas-Perez S, Rivas-Alonso E, ties: the challenge of multiprocess-multiresponse dynamics. Journal of Howe HF (2014) Roles of birds and bats in early tropical-forest restoration. Ecology 102:882–895. https://doi.org/10.1111/1365-2745.12223 PLoS One 9:e104656. https://doi.org/10.1371/journal.pone.0104656 Iijima H (2020) A review of wildlife abundance estimation models: comparison Dey DC, Schweitzer CJ (2014) Restoration for the future: endpoints, targets, and of models for correct application. Mammal Study 45:177–188. https:// indicators of progress and success. Journal of Sustainable Forestry 33:S43– doi.org/10.3106/ms2019-0082 S65. https://doi.org/10.1080/10549811.2014.883999 Jordano P, Godoy JA (2002) Frugivore-generated seed shadows: a landscape view of Falcão LA, Espírito-Santo MM, Fernandes GW, Paglia AP (2018) Effects of hab- demographic and genetic effects. Pages 305–321. In: Seed dispersal and fru- itat structure, plant cover, and successional stage on the bat assemblage of a givory: ecology, evolution, and conservation. CABI, São Pedro tropical dry forest at different spatial scales. Diversity 10:41. https://doi. Koelemeijer IA, Tack AJ, Zewdie B, Nemomissa S, Hylander K (2021) Manage- org/10.3390/d10020041 ment intensity and landscape configuration affect the potential for woody FAO (2020) Global Forest Resources Assessment 2020: Main report. FAO, plant regeneration in coffee agroforestry. Agriculture, Ecosystems & Envi- Rome. ronment 313:107384. https://doi.org/10.1016/j.agee.2021.107384 Forget PM, Jordano P, Lambert JE, Böhning-Gaese K, Traveset A, Wright SJ Kulikowski AJ, Zahawi RA, Holl KD (2022) Effects of insect herbivory on seed- (2011) Frugivores and seed dispersal (1985-2010); the ‘seeds’ dispersed, ling mortality in restored and remnant tropical forest. Restoration Ecology established and matured. Acta Oecologica 37:517–520. https://doi.org/10. 30:e13467. https://doi.org/10.1111/rec.13467 1016/j.actao.2011.09.008 Kuprewicz EK (2013) Mammal abundances and seed traits control the seed dis- Forup ML, Henson KS, Craze PG, Memmott J (2008) The restoration of ecolog- persal and predation roles of terrestrial mammals in a costa Rican Forest. ical interactions: plant–pollinator networks on ancient and restored heath- Biotropica 45:333–342. https://doi.org/10.1111/btp.12014 lands. Journal of Applied Ecology 45:742–752. https://doi.org/10.1111/j. Kurten EL (2013) Cascading effects of contemporaneous defaunation on tropical 1365-2664.2007.01390.x forest communities. Biological Conservation 163:22–32. https://doi.org/ Galetti M, Bovendorp RS, Guevara R (2015) Defaunation of large mammals 10.1016/j.biocon.2013.04.025 leads to an increase in seed predation in the Atlantic forests. Global Ecol- Li N, Wang Z, Zhang S, Yan C, Li X, Lu C (2019) Importance of bird traits for ogy and Conservation 3:824–830. https://doi.org/10.1016/j.gecco.2015. seed dispersal patterns of co-fruiting trees in a patchy forest. Integrative 04.008 Zoology 14:470–478. https://doi.org/10.1111/1749-4877.12374 Gann GD, McDonald T, Walder B, Aronson J, Nelson CR, Jonson J, et al. (2019) Lindenmayer DB, Margules CR, Botkin D (2000) Indicators of forest sustainabil- International principles and standards for the practice of ecological restora- ity biodiversity: the selection of forest indicator species. Conservation Biol- tion. Restoration Ecology 27:S1–S46. https://doi.org/10.1111/rec.13035 ogy 14:941–950. https://doi.org/10.1046/j.1523-1739.2000.98533.x Gomez JM, Schupp EW, Jordano P (2019) Synzoochory: the ecological and evo- Mailund T (2019) Pages 109–160. Manipulating data frames: dplyr. R Data Sci- lutionary relevance of a dual interaction. Biological Reviews 94:874–902. ence Quick Reference. Apress, Berkeley, CA https://doi.org/10.1111/brv.12481 Majer JD, Nichols OG (1998) Long-term recolonization patterns of ants in west- Gonzalez-Varo JP, Arroyo JM, Jordano P (2019) The timing of frugivore- ern Australian rehabilitated bauxite mines with reference to their use as mediated seed dispersal effectiveness. Molecular Ecology 28:219–231. indicators of restoration success. Journal of Applied Ecology 35:161– https://doi.org/10.1111/mec.14850 182. https://doi.org/10.1046/j.1365-2664.1998.00286.x Gonzalez-Varo JP, Lopez-Bao JV, Guitian J (2013) Functional diversity among Martínez-Muñoz M, G omez-Aparicio L, Pérez-Ramos IM (2019) Techniques to seed dispersal kernels generated by carnivorous mammals. Journal of Ani- promote tree regeneration in Mediterranean savannah-like ecosystems. mal Ecology 82:562–571. https://doi.org/10.1111/1365-2656.12024 Ecosistemas 28:142–149. https://doi.org/10.7818/ECOS.1798 Hanya G, Aiba SI (2010) Fruit fall in tropical and temperate forests: implications Martínez-Ramos M, Pingarroni A, Rodríguez-Velazquez J, Toledo-Chelala L, for frugivore diversity. Ecological Research 25:1081–1090. https://doi.org/ Zermeño-Hernandez I, Bongers F (2016) Natural forest regeneration and 10.1007/s11284-010-0733-z ecological restoration in human-modified tropical landscapes. Biotropica Herrera CM (1984) Adaptation to frugivory of Mediterranean avian seed dis- 48:745–757. https://doi.org/10.1111/btp.12382 persers. Ecology 65:609–617. https://doi.org/10.2307/1941423 Meli P, Holl KD, Rey Benayas JM, Jones HP, Jones PC, Montoya D, Moreno Hewitt N, Kellman M (2002) Tree seed dispersal among forest fragments: Mateos D (2017) A global review of past land use, climate, and active II. Dispersal abilities and biogeographical controls. Journal of Biogeogra- vs. passive restoration effects on forest recovery. PLoS One 12: phy 29:351–363. https://doi.org/10.1046/j.1365-2699.2002.00679.x e0171368. https://doi.org/10.1371/journal.pone.0171368 Holl KD (2012) Restoration of tropical forests. Pages 103–114. In: van Andel J, Mittelman P, Kreischer C, Pires AS, Fernandez FA (2020) Agouti reintroduction Aronson J (eds) Restoration ecology: the new frontier. 2nd edition. Black- recovers seed dispersal of a large-seeded tropical tree. Biotropica 52:766– well Publishing, Hoboken, NJ 774. https://doi.org/10.1111/btp.12788 10 of 12 Restoration Ecology May 2023 1526100x, 2023, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/rec.13868 by Readcube (Labtiva Inc.), Wiley Online Library on [04/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Intrinsic versus extrinsic factors in restoration Molin PG, Chazdon R, de Barros F, Ferraz S, Brancalion PH (2018) A landscape Sabatini FM, Keeton WS, Lindner M, Svoboda M, Verkerk PJ, Bauhus J, et al. approach for cost-effective large-scale forest restoration. Journal of Applied (2020) Protection gaps and restoration opportunities for primary forests in Ecology 55:2767–2778. https://doi.org/10.1111/1365-2664.13263 Europe. Diversity and Distributions 26:1646–1662. https://doi.org/10. Moreira da Silva AP, Schweizer D, Rodrigues Marques H, Cordeiro Teixeira AM, 1111/ddi.13158 Nascente dos Santos TV, Sambuichi RH, Badari CG, Gaudare U, Samuels IA, Levey DJ (2005) Effects of gut passage on seed germination: do Brancalion PH (2017) Can current native tree seedling production and infra- experiments answer the questions they ask? Functional Ecology 19:365– structure meet an increasing forest restoration demand in Brazil? Restoration 368. https://doi.org/10.1111/j.1365-2435.2005.00973.x Ecology 25:509–515. https://doi.org/10.1111/rec.12470 Sarmiento C, Zalamea PC, Dalling JW, Davis AS, Stump SM, U’Ren JM, Naoe S, Masaki T, Sakai S (2018) Effects of temporal variation in community- Arnold AE (2017) Soilborne fungi have host affinity and host-specific level fruit abundance on seed dispersal by birds across woody species. effects on seed germination and survival in a lowland tropical forest. Pro- American Journal of Botany 105:1792–1801. https://doi.org/10.1002/ ceedings of the National Academy of Sciences of the United States of ajb2.1173 America 114:11458–11463. https://doi.org/10.1073/pnas.1706324114 Ockendon N, Thomas DH, Cortina J, Adams WM, Aykroyd T, Barov B, et al. Schupp EW, Zwolak R, Jones LR, Snell RS, Beckman NG, Aslan C, et al. (2019) Intrin- (2018) One hundred priority questions for landscape restoration in Europe. sic and extrinsic drivers of intraspecific variation in seed dispersal are diverse and Biological Conservation 1:198–208. https://doi.org/10.1016/j.biocon. pervasive. AoB Plants 11:plz067. https://doi.org/10.1093/aobpla/plz067 2018.03.002 Selwyn M, Garrote PJ, Castilla AR, Fedriani JM (2020) Interspecific interactions O’Donnell FC, Flatley WT, Springer AE, Fulé PZ (2018) Forest restoration as a among functionally diverse frugivores and their outcomes for plant reproduc- strategy to mitigate climate impacts on wildfire, vegetation, and water in tion: a new approach based on camera-trap data and tailored null models. PLoS semiarid forests. Ecological Applications 28:1459–1472. https://doi.org/ One 15:e0240614. https://doi.org/10.1371/journal.pone.0240614 10.1002/eap.1746 SER (2021) Third global forum on ecological restoration. Society for Ecological Orrock JL, Levey DJ, Danielson BJ, Damschen EI (2006) Seed predation, not Restoration International, Tucson, Azusa seed dispersal, explains the landscape-level abundance of an early- Shoo LP, Freebody K, Kanowski J, Catterall CP (2016) Slow recovery of tropical successional plant. Journal of Ecology 94:838–845. https://doi.org/10. old-field rainforest regrowth and the value and limitations of active restoration. 1111/j.1365-2745.2006.01125.x Conservation Biology 30:121–132. https://doi.org/10.1111/cobi.12606 Palmero-Iniesta M, Espelta JM, Gordillo J, Pino J (2020) Changes in forest land- Tambosi LR, Martensen AC, Ribeiro MC, Metzger JP (2014) A framework to scape patterns resulting from recent afforestation in Europe (1990–2012): optimize biodiversity restoration efforts based on habitat amount and land- defragmentation of pre-existing forest versus new patch proliferation. scape connectivity. Restoration Ecology 22:169–177. https://doi.org/10. Annals of Forest Science 77:43. https://doi.org/10.1007/s13595-020- 1111/rec.12049 00946-0 Therneau T, Atkinson B, Ripley B, Ripley MB (2015) Package ‘rpart’. cran.ma. Pereira HM, Navarro LM (2015) Rewilding European landscapes. Springer ic.ac.uk/web/packages/rpart/rpart.pdf Nature, Cham. https://doi.org/10.1007/978-3-319-12039-3 Traveset A (1994) Influence of type of avian frugivory on the fitness of Pistacia Pickering C, Byrne J (2014) The benefits of publishing systematic quantitative lit- terebinthus L. Evolutionary Ecology 6:618–627. https://doi.org/10.1007/ erature reviews for PhD candidates and other early-career researchers. BF01237845 Higher Education Research & Development 33:534–548. https://doi.org/ Traveset A, Robertson AW, Rodríguez-Pérez J (2007) A review on the role of 10.1080/07294360.2013.841651 endozoochory in seed germination. Pages 78–103. In: Dennis AJ, Prach K, del Moral R (2015) Passive restoration is often quite effective: response Schupp EW, Green RJ, Westcott DA (eds) Seed dispersal: theory and its to Zahawi et al. (2014). Restoration Ecology 23:344–346. https://doi.org/ application in a changing world. CAB International, Cambridge, MA 10.1111/rec.12224 Tylianakis JM, Lalibertéb E, Nielsen A, Bascompte J (2010) Conservation of spe- Rey-Benayas JM, Martínez-Baroja L, Pérez-Camacho L, Villar-Salvador P, cies interaction networks. Biological Conservation 143:2270–2279. https:// Holl KD (2015) Predation and aridity slow down the spread of doi.org/10.1016/j.biocon.2009.12.004 21-year-old planted woodland islets in restored Mediterranean farm- Vallejo VR, Allen EB, Aronson J, Pausas JG, Cortina J, Gutiérrez JR (2012) Res- land. New Forests 46:841–853. https://doi.org/10.1007/s11056-015- toration of Mediterranean-type woodlands and shrublands. Pages 130–144. 9490-8 In: van Andel J, Aronson J (eds) Restoration ecology: the new frontier. Ribeiro da Silva F, Montoya D, Furtado R, Memmott J, Pizo MA, Rodrigues RR Blackwell Publishing, Hoboken, NJ (2015) The restoration of tropical seed dispersal networks. Restoration Vander Wall SB, Kuhn KM, Beck MJ (2005) Seed removal, seed predation, and sec- Ecology 23:852–860. https://doi.org/10.1111/rec.12244 ondary dispersal. Ecology 86:801–806. https://doi.org/10.1890/04-0847 Romanelli JP, Silva LG, Gonçalves MCP, Naves RP, de Almeida DRA, de Viani RAG, Vidas NB, Pardi MM, Castro DCV, Gusson E, Brancalion PH (2015) Resende AF, Rodrigues RR (2021) Repeatability of the searching process Animal-dispersed pioneer trees enhance the early regeneration in Atlantic in reviews of restoration outcomes. Restoration Ecology 29:e13496. Forest restoration plantations. Natureza & Conservação 13:41–46. https:// https://doi.org/10.1111/rec.13496 doi.org/10.1016/j.ncon.2015.03.005 Romijn E, Coppus R, De Sy V, Herold M, Roman-Cuesta RM, Verchot L (2019) Watson JE, Evans T, Venter O, Williams B, Tulloch A, Stewart C, et al. (2018) Land restoration in Latin America and the Caribbean: an overview of The exceptional value of intact forest ecosystems. Nature Ecology & Evo- recent, ongoing and planned restoration initiatives and their potential for lution 2:599–610. https://doi.org/10.1038/s41559-018-0490-x climate change mitigation. Forests 10:510. https://doi.org/10.3390/ Wickham H, Wickham MH (2017) Package tidyverse. Easily Install and Load the f10060510 ‘Tidyverse Rooney TP, Buttenschøn R, Madsen P, Olesen CR, Royo AA, Stout SL (2015) Wolfe BT, Macchiavelli R, Van Bloem SJ (2019) Seed rain along a gradient of Integrating ungulate herbivory into forest landscape restoration. Restora- degradation in Caribbean dry forest: effects of dispersal limitation on the tion of Boreal and Temperate Forests 4:69–84. trajectory of forest recovery. Applied Vegetation Science 22:423–434. Rubalcava-Castillo FA, Sosa-Ramírez J, Luna-Ruíz JDJ, Valdivia-Flores AG, https://doi.org/10.1111/avsc.12444 Íñiguez-Davalos LI (2021) Seed dispersal by carnivores in temperate and Wortley L, Hero JM, Howes M (2013) Evaluating ecological restoration success: tropical dry forests. Ecology and Evolution 11:3794–3807. https://doi. a review of the literature. Restoration Ecology 21:537–543. https://doi.org/ org/10.1002/ece3.7201 10.1111/rec.12028 Ruiz-Jaen MC, Aide TM (2005) Restoration success: how is it being measured? Wotton DM, Kelly D (2012) Do larger frugivores move seeds further? Body size, Restoration Ecology 13:569–577. https://doi.org/10.1111/j.1526-100X. seed dispersal distance, and a case study of a large, sedentary pigeon. 2005.00072.x Journal of Biogeography 39:1973–1983. https://doi.org/10.1111/jbi.12000 May 2023 Restoration Ecology 11 of 12 1526100x, 2023, 4, Downloaded from https://onlinelibrary.wiley.com/doi/10.1111/rec.13868 by Readcube (Labtiva Inc.), Wiley Online Library on [04/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License Intrinsic versus extrinsic factors in restoration Wright SJ, Zeballos H, Domínguez I, Gallardo MM, Moreno MC, Ibañez R Table S1. Codes and descriptions for intrinsic and extrinsic dispersal factors. (2000) Poachers alter mammal abundance, seed dispersal, and seed preda- Table S2. Usage of intrinsic and extrinsic dispersal factors used to assess forest resto- tion in a neotropical forest. Conservation Biology 14:227–239. https://doi. ration for all publications, within each type of biome studied. org/10.1046/j.1523-1739.2000.98333.x Supplement S1. Number of publications by biome, continent, and country and geo- graphic representativeness of the study area. Young TP (2000) Restoration ecology and conservation biology. Biological Con- Supplement S2. Description of variables. servation 92:73–83. https://doi.org/10.1016/S0006-3207(99)00057-9 Supplement S3. Evolution of publications over time, disturbance categories, and most addressed intrinsic and extrinsic factors. Supplement S4. Quality of representation tests and variable contributions of the Cor- Supporting Information respondence Analysis. The following information may be found in the online version of this article: Supplement S5. Classification Tree Analysis layout. Coordinating Editor: Pedro Brancalion Received: 14 October, 2022; First decision: 24 November, 2022; Revised: 11 January, 2023; Accepted: 11 January, 2023 12 of 12 Restoration Ecology May 2023