Ecosystem services of ‘Trees Outside Forests (TOF)’ and their contribution to the contemporary sustainability agenda: a systematic review

Trees Outside Forests (TOF) were recognized in the United Nations Food and Agricultural Organization’s 2000 Global Forest Resource Assessment as an essential component of sustainable development. Today, however, TOF remain an undervalued resource, with no comprehensive review of the wide spectrum of ecosystem services that are provided by TOF nor of the diversity of tree species that comprises TOF globally. Thus, a global analysis of TOF is vital to quantify their contribution to sustainable development and international climate initiatives. We reviewed the scientific literature to (1) classify and compile a repository of information on TOF resources, including the types of ecosystem services provided by TOF, geographic distribution, land-use type, and spatial pattern; and (2) document the diversity of tree species that comprises TOF globally. After screening the literature, 203 species of TOF were investigated across 20 countries in the 28 articles that we analyzed; another 15 articles across 57 countries did not include taxonomic information. Our results showed that the global distribution of TOF studies was biased toward Europe, revealing data deficiencies across the Global South, despite its contribution to biodiversity and ecosystem services. We also observed a bias in ecosystem service type toward regulating and supporting services, with studies lacking in cultural and provisioning services provided by TOF. Furthermore, studies conducted in urban areas were also lacking. To realize the full potential of TOF for sustainable development and climate initiatives, we conclude that a more holistic understanding of their ecosystem services must be established under national and intergovernmental reporting mechanisms.


Introduction
Trees, especially forested trees, are central to international climate initiatives (Köhl et al 2009, Krug 2018, Turner-Skoff and Cavender 2019, Silva Junior et al 2021, UNFCCC 2021, providing ecosystem services necessary to sustain both humans and wildlife (Lindenmayer 2009, Gibson et al 2011, Pan et al 2011, Thurner et al 2014, Pugh et al 2019. However, not all trees are found in forests (Guo et al 2014, Schnell et al 2015a, 2015b, Zomer et al 2009, 2016. Trees Outside Forests (TOF) constitute tree formations ranging from scattered trees to systematically managed trees in agroforestry systems and from woody patches within urban systems to trees along rivers, streams and canals. The United Nations (UN) Food and Agricultural Organization's (FAO) Global Forest Resource Assessment (FRA) of 2000 was one of the first reports to define and recognize TOF (Pain-Orcet and Bellefontaine 2004). The FAO defines TOF as all trees on land that do not meet the definitions of forests and other wooded lands due to size and/or land use. Here, forests are defined by the FAO as land no less than 0.5 hectares with a tree canopy cover of more than 10%, further stipulating that the trees should be able to reach a minimum of 5 meters in height at maturity in situ. 'Other wooded land' (OWL) is defined by similar Bowler et al 2010, Loughner et al 2012, Norton et al 2015, Estoque et al 2017. Trees in urban areas may also bridge the growing divide between humans and nature, increase bioaffinity and enrich urban life altogether (Turner et al 2004, Miller 2005, McKinney 2006, Wolf et al 2020, Dasgupta et al 2022. In addition to the direct benefits to humans, trees in developed areas also enhance landscape connectivity for wildlife (Zhang et al 2019).
Although TOF provide services similar to those provided by formal forests and forests in protected areas (Skole et al 2021), conservation policies over prioritize the protection of formal forests and protected areas (Ravindranath et al 2014). Not all countries consider TOF in conventional forest assessments. Recently, a range of international initiatives however required the inclusion of TOF to meet adequate levels of reporting, including the Convention on Biological Diversity, the Global Forest Resource Assessment, and the United Nations Framework Convention on Climate Change (UNFCCC). For example, the UNFCCC requires information on TOF woody biomass for greenhouse gas reporting (Eggleston et al 2006).
In this paper, we reviewed the scientific literature to explore the current understanding of TOF. Our aim was to evaluate the contribution of TOF to the contemporary sustainability agenda by classifying and compiling information on TOF, including the types of ecosystem services provided by TOF, geographic distribution, landuse type, spatial pattern, and the species that comprise TOF globally. Based on this review, we provide recommendations for future research by identifying knowledge and policy gaps in the current disclosures.

Database and review protocol
To explore the diverse ecosystem services of TOF, we adopted the Preferred reporting items for systematic review and meta-analysis protocol (PRISMA) (Page et al 2021). We searched the Scopus literature database (https://www.scopus.com/) on July 26, 2021 using the Boolean search string: TITLE-ABS-KEY (tree AND outside AND forests) AND TITLE-ABS-KEY (ecosystem AND services). Scopus, among the largest academic database, presents an extensive repository of scientific literature, including forestry, environment, social sciences, arts and humanities (Falagas et al 2008, Basu andDasgupta 2021). The search was not restricted by publication date nor language. However, only articles that provided full text in English were considered for analyses. The search yielded 100 articles (see appendix 1 for a full list of articles). Eight articles that did not provide full text in English were excluded. Thereafter, we manually screened the resulting article abstracts to include only primary literature that explicitly evaluated ecosystem services in TOF and included access to taxonomic information of tree(s) under investigation (hereafter species studies), which the author(s) referred to as 'TOF', did not explicitly refer to as forests, and/or matched the definition of the term described in the Global Forest Resources Assessments of the FAO (de Foresta et al 2013). We also included studies that investigated ecosystem services in TOF for which taxonomic information of tree(s) under investigation was not accessible (hereafter non-species studies). Large-scale remote sensing investigations and studies evaluating regulating ecosystem services, for example, often did not report on taxonomic information and were thus included in nonspecies studies. We retrieved the full text of the articles meeting these criteria and manually screened out the articles that did not provide evidence on the same criteria in full-text. After screening abstracts and full-texts, 57 articles were excluded, leaving 43 relevant articles (figure 1).

Meta-analysis
We extracted the following information from the relevant species studies that we identified: (1) type of ecosystem services provided by TOF, (2) location of investigation, (3) land-use type and spatial pattern, and (4) species of trees investigated. Using the Millennium Ecosystem Assessment (Reid et al 2005), ecosystem services were categorized as follows: provisioning, regulating, cultural, and supporting. Studies that evaluated species richness and/or biodiversity of TOF were categorized under supporting services (Schwartz et al 2000, Balvanera et al 2006, Lohbeck et al 2016, Soliveres et al 2016. We categorized land-use type in accordance with FAO (Kleinn 2000, de Foresta et al 2013, Schnell et al 2015a, 2015b as follows: agricultural areas, urban or developed areas, non-agricultural/non-urban areas, and other/not specified. Similarly, we categorized spatial patterns in accordance with FAO (Kleinn 2000, de Foresta et al 2013, Schnell et al 2015a, 2015b as follows: scattered trees, trees in line, trees in strands, and other/not specified. However, the spatial pattern of TOF resources was seldom reported. The same information was compiled for relevant non-species studies excluding species of trees investigated. For meta-analysis of the extracted information from the relevant species studies, we counted each evaluation of ecosystem service type from each tree species collected from each country, associated with each land-use type and with each spatial pattern, in each article as a case. For meta-analysis of the extracted information from the relevant non-species studies, we counted each evaluation of ecosystem service type from each country, associated with each land-use type and with each spatial pattern, in each article as a case. The results presented in the result section derive from cross-tabulation of the attributes of all cases.

Data analysis
All figures were constructed in R (v.3.6.2) (R Core Team 2014). Only data on taxonomic information at the species level were included in phylogenetic analyses. We used the package 'rotl' (v.3.0.11) (Michonneau et al 2016) to match taxonomic names to the Open Tree of Life (Hinchliff et al 2015). The Open Tree of Life Taxonomy synthesizes taxonomic information and assigns each taxon a unique numeric identifier or Open Tree Taxonomy identification tags (OTT ID). Using these OTT ID's of species that were compiled in our analyses as search properties, a phylogenetic tree from previous studies that included all of these unique taxonomic tags could not be identified. Therefore, we extracted an induced tree from the Open Tree of Life corresponding to the taxonomic tags of the TOF species that were compiled. The resulting phylogeny contained 203 tips and was rooted without branch lengths. We used the package 'ape' (v.5.6) to compute arbitrary branch lengths for the resulting tree (Paradis and Schliep 2019) and the package 'phytools' (v.1.0-1) to visualize comparative analyses (Revell 2012). The same process was used to construct phylogenetic trees at the genera and familiae levels.

Results
We classified and compiled all four types of ecosystem services as describe in the Millennium Ecosystem Assessment from the relevant studies. Studies evaluating regulating services commonly focused on carbonrelated benefits, quantifying carbon stock and sequestration of TOF resources; supporting services were often described in the context of biodiversity, detailing species abundance and distribution; while provisioning services commonly focused on TOF as timber and/or food resources, especially in rural areas; and cultural services exclusively focused on the monetary value of trees in urban settings in the United States, with the exception of one study that focused on recreational benefits of trees in Israel.
For the 28 species studies, 203 species of trees were referenced in 20 countries (figure 2(a)), totaling 439 cases (see appendix 2 for a full list of the specific cases and the type(s) of ecosystem service evaluated in each case). Several species were reported in more than one article; a few studies were conducted in numerous countries and across different land-use types; and a number of studies evaluated multiple types of ecosystem services, namely regulating and supporting services. Accordingly, 52.2% of the cases (229/439 cases) evaluated supporting services; 40.3% (177/439 cases) evaluated regulating services; 6.4% evaluated provisioning services (28/439 cases), and 1.1% (5/439 cases) evaluated cultural services ( figure 3(a)). Regarding land-use type, agricultural areas were most frequently reported (48.3%; 212/439 cases); non-agricultural, non-urban areas accounted for 13.4% (59/439 cases); and urban areas for 10.9% (48/439 cases). Land-use type was not specified in one study, accounting for 27.3% (120/439) of the cases (figure 3(b)). Sweet cherry (Prunus avium) was the TOF species cited the most times with 8 cases (figure 4; and see appendix 3 for a full list of the number of cases and type(s) of ecosystem services for each tree species). As such, Prunus and Rosaceae were the most cited TOF genus and family with 29 and 48 cases, respectively (figures S1 and S2; and see appendix 4 and 5 for a full list of the number of cases and type(s) of ecosystem services for each tree genera and familiae, respectively).

Learning from the results
Despite recent interests, TOF remain an overlooked and undervalued resource that are seldom accounted for from an academic perspective. To the best of our knowledge, there is no comprehensive review of the wide spectrum of ecosystem services that are provided by TOF nor of the diversity of tree species that comprise TOF globally. Understanding ecosystem services can demonstrate the importance of TOF to help achieve the SDGs and climate targets. At the same time, recognizing biodiversity of TOF can also support conservation efforts. Combined, this information can inform policy decisions and quantify the contribution of TOF to the contemporary sustainability agenda.
Ecosystem services provided by TOF were investigated in 66 counties (figure 2). However, we found that the global distribution of studies was biased towards Europe. Our analyses clearly revealed studies lacking across the Global South, despite its overwhelming contribution to global biodiversity and ecosystem services . This geographical bias was also observed in the species of TOF studied, with nearly a third of the cases for tree species that are native to temperate regions or the Northern Hemisphere, including Rosaceae, Betulaceae, and Pinaceae. These geographical biases may have exacerbated biases in the types of ecosystem services that were evaluated.
We also observed a bias in the types of ecosystem service that were evaluated, with a majority of studies evaluating regulating and supporting services (figures 3(a) and (c)). Studies that evaluated regulating services typically focused on carbon sequestration; while studies that evaluated supporting services typically focused on biodiversity. Our analyses clearly revealed studies lacking evaluations of cultural and provisioning ecosystem services provided by TOF. Studies that focus on cultural ecosystem services provided by TOF can help quantify factors that determine connectedness to nature and place attachment

Challenges in TOF monitoring and ways forward
Despite the articulation of TOF by the FAO as early as the 1990s, the definition and application of TOF vary by country. For example, in India, TOF are defined as trees growing outside recorded forest areas irrespective of patch size (Forest Survey of India 2019). A study in Italy defined TOF units in accordance with the Italian national forest inventory (Marchetti et al 2018). Similarly, a study in China defined TOF based on data collected via China's forest inventory (Guo et al 2014). In most cases, TOF are defined vis-à-vis forests. Therefore, one can only identify, locate, quantify, and monitor TOF if the forest boundary has been defined and delineated. But we know that forest definition varies across countries and geographic regions; and, as previous work has highlighted, the UN's definition of forests can be problematic and have major implications on measuring greenhouse gas emissions and monitoring forest degradation (Sasaki and Putz 2009, Putz and Redford 2010, Estoque et al 2021.
References to TOF were inconsistent in the articles resulting from the Scopus search string used in our study. While certain studies explicitly referenced TOF (albeit the differing definitions of TOF applied as discussed above) ( (Hinchliff et al 2015), in which the bars at the tips represent the total number of cases by type of ecosystem service evaluated in relevant studies. Each tip corresponds to a unique tree species identified in our meta-analyses. See appendix 3 full list of the number of cases and type(s) of ecosystem services for each tree species. (Lisnawati et al 2017, Tanhuanpää et al 2017, Timilsina et al 2017, Ongole et al 2018. Synonyms of TOF may include a plethora of colloquial terms, such as hedgerows, tree lines, and street trees. Thus, we focused our Scopus search to only standardized terms defined by FAO, which may have limited our analysis by excluding articles that referenced non-forested trees without using the term, 'trees outside forests.' Altogether, we recommend a standardization of TOF reporting to allow for more comprehensive and comparative analyses. Nonetheless, recent advances in Earth observation (EO) technology, advanced image analysis algorithms including intelligent image segmentation and object-based classification techniques, alongside increasing availability of spatially explicit data can help delineate TOF effectively and uniformly. Advances in EO technology have further enabled the mapping of tree species. For instance, the combination of LiDAR and hyperspectral data are effective to identify the tree species in urban areas. Altogether, this provides the opportunity for better landscape planning and integration of various ecosystem services of TOF and allows a more global perspective to integrate TOF into the international framework policy. A Brazilian city displayed an increase in green space under urbanization (Sperandelli et al 2013), a phenomenon also observed in China (Kong and Nakagoshi 2006). However, without clear intervention, urban forests may continue to disappear as in Malaysia (Kanniah and Siong 2017). TOF may also help ensure connection to nature, sense of place, and environmental justice (Gosling and Williams 2010, Soga and Gaston 2016, Ko and Son 2018, Vaz et al 2018, Schilling et al 2020, Basu et al 2021. Urban green infrastructure is unequally distributed across income and race geographies in both private and public spaces in South Africa (Venter et al 2020), a phenomenon also seen across the United States (Schwarz et al 2015, McDonald et al 2021. In fact, while landowners in Bangladesh benefit from the secondary environmental benefits that TOF provide, economic drivers have been primarily responsible for the documented increase of TOF (Salam et al 2000). TOF may act as an indicator to sustainable development and environmental justice, while providing critical public health interventions (Nath et al 2018), especially in cities experiencing rapid population growth and urbanization.
As stricter conservation policies and regulations have restricted the exploitation of resources from forests, TOF has emerged as an important resource, especially for countries with low forest cover. In Bangladesh, a country scale study on TOF, for example, mapped 31% and 53% more TOF than existing estimates of TOF and forest, respectively (Thomas et al 2021). Enabling legislation can help realize the full potential of TOF in India (Ghosh and Sinha 2018, 2019), where TOF is the main source of wood (Pandey 2008). Altogether, a more wholistic accounting of TOF resources can highlight the role of emerging countries in realizing the full potential of TOF in the contemporary sustainability agenda. Combined with additional field data, TOF inventories can have major implications on implementing carbon accrediting schemes such as REDD+and measuring progress for the SDGs.

Conclusion
Despite providing a plethora of important ecosystem services, TOF are still undervalued resources and are seldom included in national inventories and international reporting frameworks. At the same time, the scope of improving formal forests remains limited in many countries, exacerbating the need to integrate TOF and its ecosystem services in national and regional level sustainability goals and targets. In this paper, we provided a synthesis of the TOF literature based on a systematic review of scientific literature. Despite data deficiencies across the Global South, we identified 203 species of TOF and provided a rich repository of information, including ecosystem services provided by TOF, geographic distribution, land-use type and spatial pattern. These findings demonstrate that TOF may be an important resource of OECM. As such, TOF should be further integrated into efforts to achieve conservation and sustainable use of biodiversity. We further observed that TOF are fairly established for their impeccable regulating and supporting services, while still undervalued for provisioning and cultural services. In conclusion, we argue that in order to realize the full potential of TOF in the contemporary sustainability agenda, proper national and intergovernmental assessment and reporting framework should be established.

Data availability statement
The data that support the findings of this study are available upon reasonable request from the authors.