Ecosystem service flows: A systematic literature review of marine systems

Understanding and quantifying ES flows is essential for the sustainable management of social-ecological systems, as it directly captures the human-nature interactions within the system and not solely its individual elements. Especially in degrading marine systems, most ES assessments focus solely on either biophysical or socioeconomic elements of these social-ecological systems, failing to directly capture the human-nature interactions. This systematic literature review aims to capture the state of the art of ES flow studies to improve the knowledge base on marine ES flows while highlighting knowledge gaps and discussing future research pathways. Within the review we extract information on: i) the ES flow definitions, classification systems, and indicators; ii) the scales of assessment and methods used to assess marine ES flows; and iii) the types of assessment outputs. 82% of the reviewed ES flow assessment methods were spatially explicit. 63% of the studies assess marine ES flows locally. Across-scale ES flows are rarely taken into account. We detect a broad range of conceptualizations within marine ES flow literature. We thus propose an updated definition for ES flows in which they are defined as a spectrum within the social-ecological system, within which different ES flow indicators are placed depending on the relative contributions of biophysical or socio-economic attributes. Based on the extracted information and detected literature gaps, we propose a set of four criteria that should be the minimum required information when referring to ES flows: i) the relative contributions of biophysical and socio-economic attributes present in ES flow indicators; ii) identification of the supplying and receiving systems; iii) the direction and branches of flows; and iv) the spatial and temporal scales across which ES flows occur.


Introduction
Ocean's benefits reach people through the flow of Ecosystem Services (ES) to cover societal demand in a given area (Stuchtey et al., 2020). Assessing, quantifying, and mapping these flows in marine and coastal social-ecological systems (SES) 1 , from here on referred to as marine systems, is becoming significantly relevant within the increasingly connected global society. As part of a broader SES, marine ecosystems supply vital benefits to humans, such as food from fisheries, coastal protection, and climate regulation through blue carbon (Howard et al., 2017). "The remoteness of many marine ecosystems to beneficiaries" (Hattam et al., 2015a, pp. 63) triggers flows of international visitors to worldwide coastal areas and marine recreational zones (Cook et al., 2010;Rees et al., 2010;Ruskule et al., 2018), while many benefits obtained from the ocean reach beneficiaries beyond the boundaries of the provisioning ecosystem (Drakou et al., 2017b) due to, e.g., global trade. For instance, seafood biomass distributed to local communities or worldwide contributes significantly to global food security (Garcia and Rosenberg, 2010;Kittinger et al., 2015). Oceans are complex systems that extent to inter-regional and arguably even global boundaries and provide ES valuable to people irrelevant of their location (Barbier, 2017).
Understanding and measuring ES flows can reveal multifaceted differences or mismatches between the points of supply and demand (Serna-Chavez et al., 2014;Vergara et al., 2021); it can unravel the spatial distribution of nature's contributions to people and highlight societal inequities occurring between people living within supplying ecosystems and distant beneficiaries (Felipe-Lucia et al., 2015). Accurately measuring and mapping ES flows provides tangible and spatially explicit information that can be interpreted and used as scientific evidence to inform the decision-making of marine resources and societies dependent on them . In fact, without the quantification of ES flow information, ES values cannot be easily recognized and acknowledged (Bagstad et al., 2013). To this moment, the assessment and quantification of ES flows is not well established (Bagstad et al., 2014), and is mostly limited to specific ES, e.g., traded flows of agricultural products (Kastner et al., 2011;Boerema et al., 2016).
Within the concrete conceptualizations of ES concepts, along came the need to understand the relation between the ecosystems that supply the benefits and the SES in which benefits are received through the assessment of ES flows (Ruhl et al., 2007;Fisher et al., 2009;Bagstad et al., 2013). The ES flow concept became more robust when Syrbe and Walz, (2012) characterized ES flows as the spatial and temporal connections between a service providing area (SPA) and a service benefiting area (SBA) area through a service connecting area (SCA), setting the basis for the concept to be spatialized. Liu et al. (2013), within the telecouplings framework, conceptualized ES flows as movements of materials, energy, value, or information between systems, transferred due to actions taken by agents. O'Higgins et al., (2019) described ES flows as positive externalities that facilitate interdependencies between supply and beneficiary areas. This diversity of definitions and approaches created a conceptually diverse and rich landscape of ES flow conceptualizations. The added value of this heterogeneity improved the scientific understanding of the concept, while it also acted as a barrier towards sharing and using ES flow information within the communities of policy and practice (Bagstad et al., 2013).
Within this diverse landscape of assessments, most ES flow research targets terrestrial SES, while relevant work within the marine SES is more scarce despite its significance (Drakou et al., 2017b). Marine ES flow conceptualization varies significantly depending on the ES under investigation and the ecosystems of supply or the societal groups that receive or have a demand for the benefits (Garcia Rodrigues and Villasante, 2016;Drakou et al., 2017b). Although many studies outline the importance of quantifying and mapping marine ES flows (Bagstad et al., 2013;Liquete et al., 2013a;Drakou et al., 2017b), the quantification and mapping efforts are limited and relatively diverse, and heterogeneous. For instance, Liquete et al., (2013b) mapped coastal protection flows by quantifying combined indicators. Oleson et al., (2015) modeled cultural bequest values for ES flows within indigenous communities, and Drakou et al., (2018) mapped the service flow from the West and Central Pacific Ocean (WCPO) purse seine tuna fishery to worldwide beneficiaries. Such attempts explicitly assess marine ES flows, and the methods applied go beyond land-based adaptations, yet appear to be rather limited within ES assessments.
Commonly, most marine assessments focus solely on the quantification of biophysical or socio-economic ES indicators. As ES flows stand on the interface between the supply by nature and demand by people, they are ideally measured by indicators that include both biophysical and social aspects of ES. Frequently used biophysical ES flow indicators  Moher et al., (2009). in marine SES are fish catch (Miller et al., 2017) or fish landings (Neugarten et al., 2016). Socio-economic marine ES flow indicators often target regulating and cultural ES, such as the nutrient retention in sediments (e.g., Liquete et al., 2016) or the extent of recreational activities (Hattam et al., 2015b), accordingly. Integrated indicators that describe both socio-economic and biophysical attributes are used less throughout the literature (La Notte et al., 2017), but they become more and more relevant as they are essential for ecosystem accounting and integrated ES valuation . Capturing the spatial attributes of ES flow indicators remains a challenge, as our main understanding relies on broad categories of the spatial relations between ES supply and demand (Ronchi, 2018). This becomes even more demanding for the marine SES for which spatial attributes are complex and hard to define due to the mobile and dynamic nature of the marine environment and the inclusion of the third dimension of depth (Drakou et al., 2017a). This is tightly linked to the lack of homogeneous and spatially explicit data on both biophysical and socio-economic aspects of marine ES flows (Liquete et al., 2013a). Although we live in a data-rich era, data for marine SES are limited (Townsend et al., 2018). Satellite imagery is commonly used to cover broad geographical areas and is more often used within marine zones (e.g., Belhabib et al., 2020). Despite recent efforts towards this direction (Höschle et al., 2021), it remains difficult to acquire data for species distribution, and marine habitat condition to model ecological processes within marine SES (Maes et al., 2012), let alone ES flows. Spatial data needed to address the distribution of marine ES flows are still scarce. Acknowledging and reporting data collection methods, information gaps, and different visualization approaches for interpretation and communication of marine SES data (Drakou et al., 2015;Drakou et al., 2017a) is essential to improve understanding on marine ES flows. Similarly, ES flow assessment and mapping methods remain patchy and, in most cases, target specific ES. Within the marine environment, the methods applied for ES flow assessments are reproducible in limited cases (Verutes et al., 2017). Alternatively, they are tailored to specific case studies or use land-based adaptations of existing approaches (Austen et al., 2019).
Within the heterogeneous landscape of approaches, conceptualizations, data, quantification, and mapping methods, this systematic literature review aims to create an overview of all the currently available information on marine ES flow assessments. The ultimate goal is to concretize the commonly agreed-upon concepts, identify core points of discrepancy or conceptual and methodological gaps among existing studies, and potential ways forward in ES flow research. To achieve this we: 1. Review the scientific literature on marine ES flows and provide a catalog on the types of marine SES assessed and definitions, indicators, and classification systems used; 2. Review methods and tools used to assess marine ES flows; 3. Use this information to propose updated definitions and conceptualizations of marine ES flows; and 4. Use this information to propose ways forward for marine ES flow assessments.

Systematic literature review
We performed a systematic literature review following the PRISMA guidelines, using explicit eligibility criteria (Moher et al., 2009). Through the PRISMA statement, readers can identify reliable findings from which conclusions can be drawn. With the meta-analysis (a distinct feature found only in PRISMA), statistical methods can be used to summarize and analyze those findings from the included publications (Sarkis-Onofre et al., 2021). The PRISMA statement allows readers to identify author bias and systematically replicate the review methods using explicit reporting steps. The flow diagram in Fig. 1 provides an overview of the selection steps carried out on this review based on the PRISMA statement (Moher et al., 2009). The selection process was as follows: 1) identification of records; 2) title and abstract screening; 3) full-text screening based on eligibility criteria; 4) meta-analysis of included records. We examined only peer-reviewed articles since this process ensures that papers published in journals answer relevant research questions and raise meaningful conclusions (Kelly et al., 2014). Grey literature was excluded from this literature review. Despite its recognized added value due to the vast amount of information it provides, its inclusion poses a series of practical challenges (e.g., lack of information on whether these documents are reviewed, documents' length, difficulty in accurately assessing the content's relevance to the topic) (Mahood et al., 2014). During the initial part of the literature review process, the Scopus® search engine was used to search for peerreviewed articles based on the year of publication, language, author, etc (https://www.scopus.com/). We only considered Scopus® since it has the broadest coverage of any interdisciplinary abstract and citation database (Sullo, 2007), allowing researchers to search beyond the examined discipline. Only published articles until and including October 2020 (cutoff date 31/10/2020) were taken into consideration. Only publications in English were considered for the analysis.
The first step of the selection process was identifying records, including a distinct set of search terms in the title, keywords, or abstract. We searched papers that captured three major domains: socialecological systems, ecosystem services, and telecouplings within the marine and coastal environment. The search string that we used was: "social-ecological" or "socio-ecological" or "ecosystem service" and "flow*" or "telecoupling" or "distribution" or "trade" and "model*" or "map*" or "quantif*" and "marine" or "coastal". Since studies related to the telecouplings framework (Liu et al., 2013) did not appear in our results, we decided to add nine more records based on another search string, namely, "telecoupl*" and "marine" or "coast*". The next step in the selection process was the title and abstract screening. The title screening excluded publications that did not mention the terms "ecosystem service" or "social-ecological system". Studies that did not relate to the assessment of marine ecosystems were also excluded. During the abstract screening, articles were excluded if there were no keywords related to an assessment (mapping, quantification, valuation, and modeling) of marine ES flows or framed in some cases as the distribution of goods and benefits to beneficiaries. In many cases, ES flows were assessed, but studies did not explicitly refer to the ES flow term. For example, Wabnitz et al, (2018) quantified reef fish consumption as an indicator of the final distribution of ES to beneficiaries, but the paper did not mention the term ES flow. Such work entails valuable and relevant information, especially in marine ES literature, where ES flow and other information are assessed but often labeled differently (Liquete et al., 2013a). To ensure that such information is included in our review, we performed a full-text screening, and articles were included based on keywords such as the "delivery of ecosystem services" or "ecosystem service use" and "benefit distribution" or "benefit sharing", "spatial distribution", or "value distribution". In addition to these criteria, we excluded articles with no evidence to report an ES flow indicator. After the eligibility check, 54 publications related to marine ES flows were selected for the qualitative and quantitative synthesis, and analysis ( Fig. 1).

Literature analysis
The literature analysis aimed to extract information on marine ES flows, which would allow us to best quantify, model, and map ES flows within the marine environment. As ES flows are vectors, they always have a point of origin and an end-point, where the benefit is actually received. Therefore extracting information on flows required the collection of some metadata on the points of ES supply and benefit for a more comprehensive assessment. Within this review, we attempted to collect all that information from the selected publications.

Table 1
The main attributes of ES flows extracted from the literature review and their descriptions. In the last column, the classification of attributes extracted from the review process is also given. The symbol (-) in the classification of attributes corresponds to main features that did not follow a specific classification.

ES supply
Coastal and offshore habitat type supplying ES a An element of the sea or coast "that can be consistently defined spatially in the field to define the principal environments in which organisms live" (Bunce et al., 2008).
coastal, offshore, both Location of ES supply The area of "the hypothetical maximum yield of selected ES" ( Burkhard et al., 2014), irrespective of whether humans actually receive or benefit from them. We used the centroid coordinates of the ecosystem, country or region to represent this information. -

ES flow Definition of marine ES flow
The definition of marine ES flow as given by the authors. -

Classification of ES flow
It examines whether information on ES flows is categorized in particular groups and whether the author created this classification or it was adapted. We also extract the number of classes.
-ES flow indicator A measure or metric based on information that makes it possible for policy-makers to understand the condition, trends, and rate of change in ES flows (Maes et al., 2016). The spatial scale in which the flow of ES is located. It is expressed by the attributes of space and spatial extent.Temporal scale: The time scale at which the ES flow occurs or is examined. spatial scale: local, regional, global, undefined

ES flow distribution
It describes the spatial relationship between ES supply and beneficiary across three distinct spatial levels (local, regional, global). Matching flows are at the same scale with ES supply and extra-local flows are the ES flows that are not co-located with the supplying ecosystem (Drakou et al., 2017b) and have a regional or global extent.
matching flows: local to local, regional to regional, global. extra-local flows: local to regional and global, regional to global

Administrative scale of ES flow
The scale on which policies and their instruments operate (e.g., administrative units) that does not always match the scale of anthropogenic processes and their associated impact on ecosystems and biodiversity (Henle et al., 2014). It is sometimes referred to as the "level of jurisdiction" that the decision-making process influences. ES beneficiary Beneficiary d A stakeholder (individuals or groups of people) who benefits directly from a biological or physical resource, ES, institution, or social system, or someone who is, or may be affected positively by a public policy (Harrington et al., 2010). -

Type of ES benefit received
The direct and indirect outputs from ecosystems "that have been turned into products or experiences that are no longer functionally connected to the systems from which they were derived" (Potschin and Haines-Young, 2016).
ecological, social-cultural, economic, more than one, all of the above, undefined

Spatial scale of beneficiary
The spatial scale in which the beneficiary of the ES is located. local, regional, global, undefined a The characterization of the ecosystem type follows the author definition. When the studies were referring to marine ecosystems and services excluding the coastal zone, we included them in the offshore category. b We used a simplified-adapted version of the method classification provided by Santos-Martin et al. (2018) to systematically organize the reviewed methods. In two tiers, the first tier based on whether quantitative or qualitative methods are used and the second on specific mapping and modelling methods chosen (e.g., participatory mapping, modelling). c For the temporal scale; when the information was not provided, we considered the years that the project was active in a particular region. d If information on ES beneficiaries was not available, we reported the individuals or groups of people who receive, use or benefit from the ES flow as beneficiaries. e The term (sub-)continental refers to either entire continents or groups of countries within the same continent (e.g., EU).
For the selected literature, we collected information on three categories (Tables 1 and 2): i) general metadata of the publication (e.g., title, year); ii) information on ES flows assessed (e.g., definitions, ES flow indicators, indicator types, method, and scale of assessment); and iii) methodological attributes of ES flow indicators (e.g., data used, software or tools, result visualization types such as maps). The information extracted from the selected publications was categorized using existing classifications when available or by own developed classifications (feature description column in Tables 1 and 2).

Meta-analysis
To further analyze the data extracted from the selected publications and draw conclusions on the questions we have posed, we used descriptive statistics through what PRISMA calls meta-analysis (Moher et al., 2009). We combined the information quantitatively and qualitatively using graphs, tables, and maps. Flow and Sankey diagrams were used to identify which methods have been used to assess different habitat types (marine, coastal, or both) and also to which classes of ES flow indicators (provisioning, regulating, and cultural) they correspond. Such quantitative summaries of features allowed us to combine and present the outcomes of selected studies (Mann, 2016).
To get an overview of the geographical distribution of the reviewed studies, we generated a map indicating approximately their location. We only included studies that reported beneficiaries and those for which this information could be extracted (Table 1). To generate this map, we simplified the reviewed information by using the centroid of marine ES supply locations from the selected case studies and indicated the spatial magnitude of flows through symbols of different sizes. Finally, we used descriptive statistics to report and represent the data collection methods found in the literature analysis. We presented the results for the marine ES flow outputs using proportional statistics to identify the studies that were using maps and graphs, including their different types.

Additional records for ES flow assessment methods
To discuss more constructively the methods used to assess ES flows and not to restrict ourselves to those used solely for marine SES, we performed an additional literature search. We went through selected literature to detect the most commonly used methods. This information was then used to either discuss the transferability of established methodologies to the marine environment or highlight challenges in this transferability among diverse systems. For the supplementary search, we also used the Scopus® search engine. Our search focused only on the term "ecosystem service flow" and articles written in English and found before 31/10/2020. Through this search, we identified 88 records. The final selection was 42 articles. We did not perform the full literature review for these articles as with the core 54 articles but only extracted information on the assessment method.

Overview of the reviewed literature
We reviewed in total 54 publications on marine ES flows that matched the publication inclusion criteria. In the following sections, we use the term "flow indicators" to refer to the ES flow indicators (165 in total) detected in the reviewed publications. The first publication that matched the criteria was detected in 2009. Overall, publications referring to marine ES flows had an average frequency of appearance of about four papers per year and, after 2013, an average of more than six papers per year, which showed an increase in publications, especially the last couple of years (Fig. 2).
Most publications were either concentrated around Europe and the Mediterranean sea (30%) or on the southeast part of Africa around the Mozambique channel (19%) (Fig. 3). A lot of case studies were also found in North America (19%), South America (11%), and Asia (11%). Much fewer case studies were located in Oceania (4%) and Antarctica (2%). 6% of the publications were dealt with in global case studies. When it comes to the origin of ES flows (i.e., ES supply), most publications (47%) referred to ES supplied by the coastal zone (Fig. 5). 28% of the publications referred to ES supplied by offshore habitats and 25% jointly by coastal and offshore habitats. On the destination of flows (i.e., ES benefit), we attempted to extract information on the beneficiaries. This information was available in 70% of the examined case studies. In up to 55% of the reviewed publications, the authors assessed the type of ES benefit flowing to the beneficiaries. Analytically, in 7% of the total publications, all ES possible benefits were examined (i.e., ecological, socio-cultural, and economic); in 24%, more than one benefit type was assessed within the same study, while 15% looked only at economic and 9% only at socio-cultural benefits.

Definitions of marine ES flows
Within the reviewed literature, we detected a considerable variation in definitions of ES flows. We detected three major groups of ES flow assessments: i) those defining ES flows; ii) those referring to ES flows without clearly defining them within the ES concept, and; iii) those using ES flow indicators without naming them as such. i) The 23 publications in which marine ES flow was clearly defined and discerned from the rest of the ES components (e.g., Liquete et al., 2013bLiquete et al., , 2016. Most of those used an adapted definition from Syrbe & Walz (2012) to describe ES flows as "the spatial and temporal connections between a service providing area (SPA) and a service benefiting area (SBA) area through a service connecting area (SCA)" (e.g., Rova et al, 2018). Other definitions were those of, e.g., Owuor et al. (2017), who described ES flows as a set of ES currently consumed or used in a particular region. Liu et al. (2013) described ES flows as part of the telecouplings framework, i.e., as movements of material, energy, or information between the systems that are transferred as a result of actions taken by agents. ii) The 11 publications in which the concept of ES flow was not distinguished from ES supply and benefit (e.g., Arkema et al., 2015). In such cases, the term "ES flow" and "delivery of ES" was mentioned and quantified but was not clearly defined. iii) The 20 publications in which ES flows were looked into but were named differently (e.g., Wabnitz et al., 2018). In this case, the measured indicators reflect ES flows, similar to those mentioned in the previous two cases, but do not use the ES flow terminology. An example is Wabnitz et al., (2018), who assessed reef fish consumption, a commonly used ES flow indicator, but they did not label the indicator as ES flow.

Table 2
The extracted information on methodological attributes of marine ES flow indicators, their description, and the classification of the information.

Data collection methods
The method that has been used to collect the ES flow data

Classifications of ES flows
We observed that only 26% of the selected publications followed an established (4%) or developed their own classification (22%) of ES flows. Most of those using existing classifications followed Burkhard et al., (2014) (no flow, very low flow, low flow, medium, high flow, very high flow) and Egoh et al., (2008) (flow hotspots and coldspots). Most studies used a classification of ES flows which refers to low, medium, or high intensity of flows, for example, the range of values of access for recreational use (Ruskule et al., 2018). Others developed new classifications, such as Drakou et al., (2018), who defined three types of spatial marine ES flows, namely, one-to-one flows, open loop, and closed-loop, to account for the spatial dimension and direction of ES flow. Similarly, Ghermandi, (2015) classified marine ES flows as international, domestic, and local. Onat et al. (2018) classified coastal exposure flows in a 1 to 5 gradient from very low to very high vulnerability. Raya Rey and Huettmann, (2020) classified ES flows as intracoupling, pericoupling, and telecoupling.

ES flow indicators
Direct ES flow measurements were used for 35% of the indicators, such as seafood consumption (Lange and Jiddawi, 2009;Liquete et al., 2016;Miller et al., 2017), seafood harvest (Guerry et al., 2012), and amount of traded species (Garcia Rodrigues and Villasante, 2016;Rakotomahazo et al., 2019). Most authors used proxy indicators (46%) to assess ES flows. Representative examples of proxies were the number of companies involved to account for wildlife watching services (Outeiro et al., 2015) or the number of recreational float homes to assess the service of recreation (Guerry et al., 2012). A few of the assessed publications used composite indicators (19%), such as coastal protection ES flows (Liquete et al., 2013bGrizzetti et al., 2019) and the coastal vulnerability index (Mandle et al., 2017;Onat et al., 2018). A complete list of extracted ES flow indicators is presented in Appendix A2 ( Table 4).
The indicators detected were diverse in how they capture biophysical or socio-economic information or the combination of both. In particular, 49% of marine ES flow indicators derive socio-economic information, such as "fish consumption" , "visitor days" (Arkema et al., 2015), and "visitation and demand for tours" (Outeiro et al., 2015). Other marine ES flow indicators (25%) consist of mainly biophysical information, such as the "amount of phytoplankton biomass" (Deininger et al., 2016) or the "wetland area" (Butler et al., 2013). Lastly, 26% of the examined indicators integrated both biophysical and socio-economic aspects of SES, such as the "recreation opportunity spectrum"  or the "supply chain segment length" (Dvarskas, 2018). A pattern observed was that for most recreational ES (72%), marine ES flows were mainly described by socioeconomic indicators. For regulating ES, most reviewed indicators (40%) described biophysical elements of the SES. This was especially the case  For the rest of the case studies, circles represent "matching ES flows," i.e., the cases for which the location of supply coincides with the location at which the benefit is received. Different sizes and colors represent different scales (local, regional, global). Diamonds represent the "extra-local ES flows", i.e., the cases for which the location of supply differs from the location of benefit reception. The diamonds are placed in the supply location, and the size and color represent the different magnitude of flows.

Scales of marine ES flow -spatial, temporal, and administrative
We analyzed the spatial scale and extent of ES flows. In 57% of the reviewed publications, ES flow from ecosystems locally to reach local beneficiaries, a flow type we label as "local to local" (e.g., Europe - Ruiz-Frau et al., 2013;Miller et al., 2017;North America -Dvarskas, 2018). 11% of reviewed publications assessed regional to regional ES flows from larger in extent systems such as the Mediterranean sea, Mozambique channel sea, and Europe. Another set of publications assessed marine ES flows across different spatial levels, i.e., there was a difference between the point of supply and the point of benefit reception. Such local to regional and global flows were 20% of the case studies. These were assessed in ten locations and mainly supplied from South America (Ghermandi et al., 2018) and South-East Africa (Neugarten et al., 2016). Regional to global flows (7%) were supplied from locations in the Mozambique channel, WCPO region, and Europe. Lastly, 5% of the case studies had a global extent (two in coral reefs and one in fisheries).
The temporal scale of marine ES flows was specified in 72% of the case studies. Some publications considered a large temporal extent (e.g., Miller et al., 2017, who assessed ES flow within a range of 70 years), but the most common range observed was two years (24% of the case studies). We observed that the reviewed publications did not account for any temporal lags or differences in temporal occurrence between supply, flow, and benefit reception.
Regarding the administrative scale, the literature analysis showed that 36.3% of the ES flows were managed locally or nationally. In 5.5% of the publications, a global organization or institution (e.g., Commission for the Conservation of Antarctic Marine Living Resources -CCAMLR-in the Weddell Sea marine protected area as in Deininger et al., 2016) managed the ES flow, while the 21.2% were administered at the (sub)-continental level (e.g., European Union as in Ghermandi, 2015). For 37% of the reviewed publications, we could not collect any evidence on the level at which the decision-making processes took place. In addition, we detected mismatches in 4% of the publications, in which the systems were managed by authorities with a larger extent of power. For example, for coastal recreation, ES were also flowing regionally and globally, but they were managed by the town and country's planning division of the ministry of planning and development within Trinidad and Tobago (Ghermandi et al., 2018). Another example was seaweed farming which occurred and was managed locally, but the flows of benefits reached beneficiaries globally (Lange and Jiddawi, 2009).
When assessing the end-point of ES flow, i.e., the beneficiary side, we detected that flow beneficiaries were located at local levels (54%) in most publications. 32% of them were found at the regional scale (e.g., coastal protection in Zulian et al., 2014), while global beneficiaries (e.g., carbon sequestration benefits in Ghermandi et al., 2019) appeared much less in the literature (9%). In a few instances, the spatial scale in which they were located could not be explicitly defined (5%). Examples were the cases in which we could identify beneficiary types, such as recreational divers (Menzel et al., 2013), but we could not collect information on their origin.

Marine ES flow assessments methods and tools
Most publications used a combination of methods for assessing marine ES flows. We observed that after 2015, assessments at the regional and global level were more frequent, but the majority still examined flows locally. Combinations of different approaches were limited until 2013, but since more advanced geospatial tools became available, integrated and spatially explicit methods have appeared more often in the last five years (Fig. 4). For instance, after 2015, more studies used spatially explicit methods for assessing marine ES flows at the regional level. We also observed a similar pattern at the global level after 2017. Non-spatial or the combination of non-spatial and monetary methods appeared less throughout the years.
Most ES flow assessment methods were spatially explicit (54%). The second most usual type of analysis was the combination of spatial and monetary methods (28%) followed by non-spatial methods (11%). Nonspatial and monetary methods appeared much less in the literature (7%). In Appendix A1 (Table 3), we provide an overview of all assessment methods used in the reviewed literature. A Sankey diagram (Fig. 5) indicates the types of marine ES flows, the supplying ecosystem types, and the methods used to assess them. For example, in offshore habitat types, most authors used spatial methods to assess most ES flows (Leslie et al., 2009;Villa et al., 2011). Also, more than 40% of flows for provisioning ES were linked to these habitats in this case. Only 5% of the publications used non-spatial assessment methods, and 2% used a combination of non-spatial and monetary approaches to assess provisioning ES flows for offshore habitats. ES flows associated with both coastal and offshore habitats were analyzed using a variety of methods, including also monetary approaches (e.g., Lange and Jiddawi, 2009). We looked more into detail at the spatially explicit methods since the non-spatial methods were more limited. Approaches such as the spatial variance (Halpern et al., 2011), the hot spot analysis approach (Dvarskas, 2018), the analysis of site-specific visitor landings (Deininger et al., 2016), and the GIS-based index modeling (Onat et al., (2018) considered additional variables for the assessment including distribution patterns across space. Various InVEST models (Guerry et al., 2012;Mandle et al., 2017;Zhang et al., 2020) and the Ecosystem Services Mapping (ESTI-MAP) tool (Zulian et al., 2014) also appeared frequently (7%). More coordinated efforts to develop a systematic approach are driven by Villa et al., (2011) with the Service Path Attribution Network (SPAN) and Liu et al. (2013) with the telecouplings approach. The SPAN modeling method has been employed in marine systems to some extent due to its benefit-centric approach, but still, it has limited application. On the other hand, Liu et al. (2013) have established a comprehensive method, but its focus is explicitly on assessing "distant" flows. Only 6% of the publications used this type of approach in marine systems. Moreover, from Table 3, we observed that 13% of the publications used a value chain analysis or value transfer approach either through a meta-analytical method as presented in Ghermandi et al., (2018Ghermandi et al., ( , 2019 or a spatially explicit one (Camacho-Valdez et al., 2014;Drakou et al., 2018). Additionally, 5% of the publications used the Social Cost of Carbon (SCC) approach to map and quantify carbon flows (Drakou et al., 2017b;Ghermandi et al, 2019;Luisetti et al., 2019). 4% of the publications used a participatory method for the assessment.
Available tools and software for assessing ES flows included the ESTIMAP tool, which is applied using a set of different process-based models within a GIS environment. We noticed that 13% of the studies used modeling tools to assess marine ES flows, such as the coastal vulnerability model (e.g., Mandle et al., 2017;Onat et al., 2018), and 7% developed their own models with programming (e.g., Thiault et al., 2017;Rova et al., 2018). In about 30% of the publications, we could not find any specific information on the software or tool used. The rest of the studies with the software used are represented in Table 5 (Appendix A3).

Land-based ES flow assessment methods
From the analysis of additional ES flow studies that did not explicitly refer to the marine systems (42), we observed that spatially explicit approaches appeared more frequently in the literature in comparison with the marine ES flow assessments. We found that 90% of authors used  Estimates refer to approximate calculations performed by using pre-processed data; databases refer to accessible data within existing databases either specific to the project or collected upon request; models refer to collecting data from modeling outputs; workshops refer to data collected as an output of an expert working group or a participatory process; others refer to data collected from crowdsourcing methods, etc. spatially explicit methods to assess land-based ES flows, and 24% of them were combined with a monetary approach. Some of these integrative methods include the spatially explicit value transfer method (Troy and Wilson, 2006) and the spatial subsidy approach . Representative examples of spatial methods we found were the ES flow matrix approach (Vrebos et al., 2015), the surface water supply model within the SPAN modeling approach (Li et al., 2019), and the cumulative viewshed models in Egarter Vigl et al., (2017). Examples of more analytical methods for land-based ES flow assessments were the Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) models used to map dust particle trajectories to assess the wind erosion prevention service in China (Xiao et al., 2017;Xu et al., 2018). Another commonly used approach was the bayesian models to analyze sediment retention, vegetation and soil carbon storage, and open space proximity ES (Zank et al., 2016).

ES flow input data
We observed considerable heterogeneity in the input data and collection methods for marine ES flows. Aerial or satellite images and remote sensing techniques were mostly used to gather data in 20.2% of the cases (i.e., to determine land use land cover classes in coastal regions). In 19.5% of the reviewed publications, data were collected through interviews or surveys. In 18.9% of the marine ES flows assessed, estimates (i.e., approximate calculations of seafood biomass) were used to collect data. Databases (referring to available data in existing databases either particular to the project or found upon request) were used in 16.5% of the cases (Fig. 6). In 14.5% of the case studies, a modeling approach was used. For example, to extract data on the total catch and the trophic level of catch, Liquete et al., (2016) used the Ecopath with Ecosim model from Piroddi et al., (2015). Data collection workshops, which usually require stakeholder involvement, were less frequent (5.4%). These methods are mainly used when the focus is on local communities and the marine ES they manage. Other data collection methods such as crowdsourcing or statistics (i.e., official government or organization statistics for tourism arrivals as in Lange, 2015;Ghermandi et al., 2019) appeared to be also limited (3.4%) for ES flow assessments. Less frequent data collection methods were the direct field observations (1.7%).

ES flow assessment outputs
In 85% of the reviewed publications with a visual output, marine ES flow assessment outputs were represented through visualizations in the form of a map or a combination of a map with a table or a graph, and only 15% of the cases were graph or table representations. From the observed outputs, we could group them in two broad categories: those that visually showed the flow as a vector with a start (location of ES origin) and end-point (location of ES benefit) and those that represented flows as quantities without a spatial attribute. Only a small part (1.8%) of the reviewed studies had no visual output.
Among the spatially explicit outputs, 4.3% of the maps represented flows as arrows from one location to another in a 2-D map, while most of them (95.7%) presented this information in the form of a density map. Representative examples of density maps were the study of Liquete et al., (2016) that visualized the coastal protection flow, and the analysis of Camacho-Valdez et al, (2014), which used aggregate raster maps to show the total ES value flows. Dvarskas, (2018) used a network of links between nodes (without direction) to represent seafood and recreation flow along a service supply chain. Output maps from Drakou et al., (2018) showed different types of seafood flows depicted as arrows with net benefits flowing among twenty countries and five continents.
In the case in which flows were depicted as quantities, various graph types were used to visualize the assessment outputs. Many studies combined graphs and maps, enhancing the level of information represented (32%). For example, the study of Ruiz-Frau et al., (2013) used maps to show the distribution of recreational activities (sea-kayaking) and a scatterplot to present the relationship between the kayaking route popularity and the presence of wildlife in the region. Similarly, Canu et al., (2015) used scatterplots and maps to visualize the carbon sequestration fluxes in the Mediterranean sea. Some of the ES flows were represented through a combination of graphs (10%), and a great proportion was depicted through scatterplots (15%). Spider charts (6%) and bar plots (4%) were used less frequently.

Discussion & conclusions
Understanding, defining, and accurately measuring ES flows can improve the decision-making process to manage marine ecosystems effectively. Current management of human-ocean interactions is predominantly sectoral, focusing on managing specific sectors of coastal or offshore human activities, such as fisheries or tourism (Winther et al., Fig. 7. The marine ES flow definition spectrum. Dark boxes represent ES flow indicators which depending on their relative contributions of biophysical and socio-economic information, are placed in different parts of the spectrum. For example, indicator 1seafood catch, is calculated using biophysical data on seafood biomass. Therefore it is placed within the "biophysical" side of the spectrum. Indicator 2 -catch per unit effort is calculated by integrating information on fish caught during a fishing trip with the number of fishing hours. Such ES flow indicators combine biophysical with social data, and therefore they are placed within a more central location in the spectrum. 2020). Policy guidelines are usually designed at a global, regional, and national level, while actual management mostly happens nationally or sub-nationally. Yet human-ocean interactions occur and have impacts that span beyond anthropocentric boundaries, making the management of this reciprocal relation difficult to capture (Drakou et al., 2017b). Recognizing, understanding, and describing human-ocean links as flows of ES can shed light on this relationship. At the same time, its accurate assessment and quantification is the first step towards the inclusion of this information within decision-making processes. In this review, we captured the current status of marine ES flow assessments, highlighted points of ambiguity and heterogeneity within recent peer-reviewed literature, and discovered areas of conceptual and methodological convergence that can serve as a solid basis for future research. A limitation of the study is that only peer-reviewed publications written in English were included, which might exclude valuable grey literature publications (Mahood et al., 2014).

Re-thinking ES flows
Within the reviewed literature, we discovered a plurality in definitions and conceptualizations of ES flows. Many define flows as "the benefits people actually receive" (Layke, 2009) or as "ES consumed" (e. g., reef fish consumption as in Wabnitz et al., 2018), giving the notion a strong anthropocentric view. Others give a stronger emphasis on the biophysical component, identifying them as "ecological functions that contribute to human well-being" (e.g., Liquete et al., 2013b), or even as "the different types of ecosystem goods and services generated by stocks of natural capital" (Ghermandi et al., 2019). Within the reviewed ES flow conceptualizations, a core element leading to the observed abovementioned variation was that biophysical, social, and economic attributes comprise ES flows in variable analogies. The detected heterogeneity is reflected already in the conceptualization of ES as a "boundary object," defined, established, and operating across disciplines (Abson et al., 2014;Ainscough et al., 2019;Lundgren, 2021). Most "boundary object" concepts, e.g., resilience, which are established in the integration of different disciplines, share similar challenges (Brand and Jax, 2007). Under the need to recognize these plural typologies and emerging challenges, we define ES flows as a social-ecological spectrum through which ES stream in a non-linear manner from the point of supply to the points where the benefits are received (Fig. 7). We choose the term spectrum to indicate "a condition not limited to a single entity, but a set of elements that can vary, without steps, across a continuum" (definition adapted from The American Heritage Dictionary of the English Language 4th Ed. Hought, 2004).
Within that definition spectrum, we explore the placing of marine ES flow indicators, i.e., the measurable quantity of ES flow. Overall, there is an expressed uncertainty in the literature on how we quantify ES flows with indicators (Hattam et al., 2015a). According to our findings, this uncertainty is potentially one reason why proxy indicators and not direct measurements have been used so frequently. We detect studies that generated ES flow indicators measuring economic values (Canu et al., 2015) or others that spatially accounted for ES flows as ES vectors (Raya Rey et al., 2017). Other studies assessed the full ES distribution process (Dvarskas, 2018), and others focused solely on the actual flow, without addressing its point of origin (supply) or destination (benefit) (Ghermandi et al., 2019). Especially in marine SES, for which flows also happen through the trade of, e.g., fish or seafood biomass, in many cases, studies focus on the supply of flow without addressing the reception of the final benefit, looking at, e.g., fish catch (Miller et al., 2017), generating confusion of where these benefits actually flow to. The missing information on the beneficiaries is crucial in these cases, as they are the ones who drive the demand for the use of additional marine resources (Bagstad et al., 2014). Given that, this valuable plurality of approaches might also have a downside when it comes to the use and operationalization of this concept to inform decision-making, as it might miss essential elements that need to be taken into account when managing marine resources. To be effective to management, indicators should be specific, scalable, and transferable, among other criteria (Hattam et al., 2015a;Oudenhoven et al., 2018).
Given the detected gaps and convergence points in the literature, we propose the following ES flow attributes to act as the minimum required information when describing, quantifying, and mapping these flows: i) The relative contributions of social, economic, and biophysical attributes in ES flows as these are reflected within the ES flow "spectrum" within an ES flow indicator metadata sheet. This allows to detect in which part of the ES spectrum ES flow indicator falls, and therefore better understand the meaning, data behind, and interpretation of such an indicator. To achieve this, ES flow indicators, especially the composite ones, should contain information on the data used for their construction along with associated metadata. Such information also improves transparency on quantification means for such an indicator (Godar et al., 2016).
ii) The supplying ecosystem and the beneficiaries. As flows could be conceptually considered as a vector, it is essential to ensure that the origin (ES supply) and the flows destination (ES benefit) are specified when they are measured. Our review indicated that in most cases, especially the benefit of the flow, is rarely specified. This might induce confusion in sharing this information, as it leaves ES flow as an abstract and generic concept, which is sometimes hard to conceptualize, especially when there is a need to use this information for decision-making. In ES terms, this practically indicates that the supply and benefit side of ES should be clearly specified when referring to ES flows to more accurately indicate the flow trajectory.
iii) The directions and branches of ES flows. Understanding ES flows requires understanding the trajectory from the point of origin (ES supply) to the intermediate and final destination (benefit). Especially marine ES, which usually flow across large areas, regions, and continents Bennett et al., 2021), can flow directly or indirectly to many different beneficiaries. Understanding, measuring, and considering this information when quantifying and mapping ES flows is essential, as it allows for a more inclusive and transparent assessment. Similar to the cartographic concept of branching origin-destination data (Boyandin, 2013), ES flows can be depicted as a collection of flows of entities between geographic locations, with their own attributes. Capturing such information allows for spatially explicit quantification of ES flows. This information is crucial as it can unveil information on the distribution of benefits to different beneficiaries, with different powerinterest relations, as a means to highlight social inequities (Österblom et al., 2020). iv) When relevant, the spatial and temporal scales in and across which ES flows occur. This is especially valid for the case of scale mismatches, in which ES benefits flow, e.g., from local to global level, as is the case for many regulating ES . Similarly with traded ES, flows might occur from local to regional level, information which is relevant especially when quantification and mapping methods need to be used and adapted to cross spatial scales. Information on temporal scales could be relevant, especially when there are mismatches between the time of supply until the time of the benefit reception (Guerra et al., 2017). This can be reflected in the data used to assess and map these ES flows and, most notably, in the ES assessment outputs (see also Section 3.5.2 in this manuscript).

The importance of scale
In many cases, existing assessments focus mainly on one spatial or temporal scale (e.g., Camacho-Valdez et al., 2013). Yet, the benefits society receives from the ocean span across the globe, either due to natural processes (e.g., species migration or flow of water masses) or through anthropogenic contributions (e.g., migration, trade) (Barbier, 2017). This has been vastly recognized in the research done through the telecouplings framework (Liu et al., 2013) and interregional ES flows . When it comes to dedicated work on marine ES, the research is much more limited to the extra-local framework, focusing on fisheries or regulating ES (e.g., Drakou et al., 2017b;Inácio et al., 2020) or, in rare cases, recreational ES .
The temporal scale is rarely examined in detail, which might hinder our efforts towards monitoring ES flow changes in time. The few cases which referred to temporal scale within marine ES flows mainly focused on particular snapshots in time (Butler et al., 2013;Gacutan et al., 2019;Havinga et al., 2020). Overall within the scientific literature of marine ES, the time has been assessed mainly with changes of ES supply (e.g., changes in fish biomass) or changes in habitat structures over time (Hattam et al., 2015b). Yet when it comes to understanding the temporality of flow, research is much more limited. ES might flow to the final beneficiary across varying time ranges within marine systems, e.g., from fish catch to processing, canning, and final consumption to a few hours when it comes to recreation. Yet, capturing this information proved to be challenging.
In only a few cases, we could trace information related to the administrative scale (defined as the scale at which the decisions are taken, from Henle et al., 2014). In the marine and coastal realm, ES flows occur across a range of scales and are usually different from the scale of supplying ecosystem or the scale at which the system is managed, generating mismatches. For instance, within the European Union, the Common Fisheries Policy (Jensen, 1999) is a policy direction aimed at managing fish stocks around the Union towards their more sustainable exploitation. While the EU is in charge of managing fish stocks, the actual flow of ES from fisheries happens both at the local and national levels within the Member States. Information on mismatches particularly related to the administrative scale appears to be limited in our results.
Recognizing and addressing scale plurality within marine ES flows can improve the credibility and accuracy of the information generated for decision-makers (Raudsepp-Hearne and Peterson, 2016). This information becomes crucial for analyzing ES flow patterns since scale variations can change the set of ES provided by seascapes in space and time. Indeed, more studies have examined the different spatial, and administrative scales of ES flows in the past years. Even so, a spatially explicit assessment of the interactions at different scales remains challenging (Geijzendorffer and Roche, 2014;Bennett and Chaplin-Kramer, 2016). This becomes explicitly demanding for complex interactions in regulating ES and flow paths in provisioning and cultural ES. For example, in the assessment of the carbon sequestration, the ES benefits are shared among a larger group of beneficiaries across scales, which might not determine the level of ES provision, but largely depend on it. Some countries might need to cover the costs of supplying ES, and others solely enjoy the benefits from ES flows. Examining marine ES flows across scales helps us address these potential equity concerns associated with benefit distribution and costs of providing or receiving ES .

Diversity in quantification methods
The assessment and quantification of ES flow are mostly centered around land-based approaches which are adapted to the marine system. Typically, an assessment involves the characterization of sending and receiving systems, identifying and quantifying relevant ES flow indicators, and evaluating uncertainty and data gaps (Koellner et al., 2019). In marine and particularly offshore systems, we find very few examples to involve all these steps.
The definition or conceptualization of ES flow practically indicated the corresponding method for its quantification and assessment. What struck us was the limited spatially explicit approaches, especially when it comes to distant, interregional, or extra-local flows, as they are called. The value of mapping and visualization of ES has been highlighted by many (Bagstad et al., 2013;Liquete et al., 2013a;Drakou et al., 2017a), but implementing this when it comes to describing ES flows has its limitations and even more when it comes to marine systems. Only a small percentage (4.3%) of the reviewed studies depict flow maps as origin-destination arrows. This contrasts with how ES flows are visualized from terrestrial ecosystems to beneficiaries, where depicting flows with arrows is a more common practice (Gondwe et al., 2011;Palomo et al., 2013;Bagstad et al., 2019). Also, the level of overlapping between methods applied in marine and non-marine SES was minimal. Additionally, an integrated and consistent approach similar to the work of Tonini and Liu, (2017) which could contain a spatially explicit toolset to map marine ES flows, is missing. We come to the point to argue that mapping in the traditional, static, 2-dimensional manner is no longer enough, and we might need to think of alternative ways of doing so. Visualizing ES flows has happened in disciplinary areas outside ES. Methods such as input-output tables (Kitzes, 2013) or arc diagrams (Wattenberg, 2002) might more comprehensively describe ES flows in a straightforward way. Certainly, more effort is required to understand how these different approaches can be combined towards a cohesive ES flow assessment, especially in marine systems where the lack of such an approach is still prominent.
Besides the method availability, a factor also hindering the ES flow assessment in marine systems is the lack of data for the actual flow of ES. Some authors report data incompleteness for particular ES , but these data gaps and uncertainties are addressed inadequately. Land use land cover data can be used to assess some coastal ES flows, but this information is not useful for offshore ES. Other data types, including databases and estimates, are used more frequently, especially at the local or national level, where this information is largely available (Drakou et al., 2017b) through, e.g., local and national level observatories (e.g., FFA, 2016). On a global scale, some recent efforts (Global Fishing Watch, 2017) show a large potential to identify the actual flow path of provisioning services such as seafood. The distribution path of such services to beneficiaries through intermediate activities, including transportation routes, etc, can be used for detailed accounting and visualization of such flows locally or across scales. Regarding regulating flows, models and databases will likely remain the main source of data collection. On the other hand, depicting recreational flows largely depends on the availability of spatially explicit movement data to account for the path of beneficiary groups to marine ecosystems.

Conclusions
This review summarizes the scientific literature regarding marine ES flows, including definitions, indicators, assessment methods, and outputs. This is especially important for future ES research, as within the Anthropocene era and in a period in which natural elements are traded globally, research needs to move beyond understanding the point of ES origin or benefit receival exclusively. Understanding the patterns of ES flow is essential for a sustainable and equitable ocean future. Understanding the patterns of ES flow not as uni-directional linear processes but as complex networks with branching links across different locations and actors will shift the way we understand and manage marine ES. Drivers of ES flow do not only stem from the demand of the final beneficiaries. Intermediate actors such as trans-shipment companies (Miller et al., 2018;Seto et al., 2020) or other trans-national corporations (Österblom et al., 2015;Folke et al., 2020;Virdin et al., 2021) also play a major role while receiving intermediate benefits in the ES distribution process .
Below we summarize the core findings of this review and give recommendations for future work.
• We detected a vast plurality in definitions of marine ES flows. Similar to other concepts like resilience or connectivity, this plurality is Janus-faced (Brand and Jax, 2007;Bormpoudakis and Tzanopoulos, 2019) since both opportunities and challenges arise with it. On the one hand, the multitude of ES flow definitions raises options for different aspects of "ES flow" to be investigated, addressed and explored, across a range of ES and scales. This allows for the concept to be "validated" and better defined and could lead, in the long term, to a better and more grounded establishment of the concept. The downside of this plurality is that this large diversity might act as a barrier to its operationalization since the diversity of outputs generated might make it hard to find a coherent way to use this information within the decision-making process. • To account for the plurality in marine ES flow definitions and indicators, we propose an updated definition, which acknowledges the difference in relative contributions of social and biophysical elements across the ES flow "spectrum". habitat, or country. Assessments across locations or different spatial scales are rarer in marine SES. This is especially prominent within cultural and regulating ES, for which identification of the exact area of supply, direction, or beneficiary area is less straightforward. We argue that looking at marine ES flows across spatial scales or different locations is an essential next step in which marine ES research needs to focus, especially when it comes to generating appropriate and coherent methods for quantification and mapping. • 72% of marine ES flows are assessed at the coastal zone (either solely coastal or coastal and offshore together), while the rest 28% of the studies, address offshore habitats. A multitude of reasons could be behind this finding, such as the better knowledge of coastal zone compared to offshore habitats (Miller et al., 2017) or even proximity to where people mostly are (Menzel et al., 2013). Nonetheless, the focus on the offshore and benthic habitats should be shifted, especially in marine ES flows, as maritime activities such as transportation, fishing, offshore wind energy exploitation, and exploitation of oil and minerals occur offshore. Understanding the use of offshore habitats as a place where marine ES flows occur is crucial for informed and sustainable management of these ecosystems and the associated human activities. • Methods for marine ES flow mapping are very diverse depending on the scale of assessment, the system being looked at, or the focal ES. This diversity in approaches and outputs generated is underpinned by the lack of coherent conceptualization and framing of the concept. • Data availability proved to be a core limitation of existing marine ES flow assessments. With the increased availability of geospatial data, including distribution paths of ES to beneficiaries, intermediate activities, and type of benefits received, we can achieve greater accuracy for detailed quantification and visualization. • The marine ES flow and the use of ES from beneficiaries is a dynamic process across time. Yet, data used rarely capture the temporal dimension and might hinder research efforts to detect ES flow variations. We, therefore, urge for more temporal data to be collected and included within existing assessments.
We foresee that the information generated in this review will contribute directly towards enhancing marine but also terrestrial ES flow research, as it attempted to clarify existing concepts through an inclusive re-organization of existing research and theory. This review could spur more research towards better quantifying and mapping marine ES flows. We think that future research should move outside the disciplinary boundaries of the ES community and integrate with other disciplines, especially when it comes to methods applied. For instance, tools from industrial ecology used for material flow analysis (Brunner and Rechberger, 2004) might be great starting points for better quantification while directing towards new types of data that need to be collected. We suggest strong collaborations with cartographers and geovisualization experts to generate novel mapping approaches for ES flows, which span beyond the traditional static approaches but move into the fields of interactivity and dynamic flow mapping. A more coherent sharing of outcomes can lead to a better understanding of the ES flow dynamics and might shed light on known issues that have not been previously quantified and mapped. A great example of this is already the summarizing map of our research (Fig. 3), which already indicates that among the marine ES flows studied, the ES supplied from emerging economies tend to flow across the globe. In contrast, those flowing from the rest of the world seem to supply benefits to the areas where the supply comes from. This is a coarse observation and should be used as a conversation starter to understand these patterns better. Understanding the uses of marine resources through the prism of ES flows allows for more integrated and better-informed approaches towards managing and equitably sharing such resources. With those in mind, we believe that improved knowledge on marine ES flows can also support the implementation of policy responses and actions as part of the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services (IPBES) agenda, to achieve sustainable governance and management of seascapes, oceans, and marine systems.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Table 3
Overview of the marine ES flow assessment methods used within the reviewed literature. Methods are classified into categories (spatial, non-spatial, spatial and monetary, non-spatial and monetary) and sub-categories (mapping, participatory, models, quantification) based on their spatial or non-spatial characteristics.

Method categories
Sub-categories Marine ES flow assessment methods