Governance Planning for Sustainable 1 Oceans in a Small Island State 2

11 Promoting the UN Sustainable Development Goals (SDGs) will require aligning government institutions 12 and must contend with the often siloed nature of institutions within organizations, making the 13 identification of cooperative institutional networks that promote SDGs a priority. We develop and apply 14 a method which combines SDG interlinkage analysis, which helps determine priorities and prerequisites 15 for SDG attainment, with the transition management framework, which aligns policy goals with 16 institutional designs and programs. Using Aruba as a model case study of a small island state with a 17 planning committee for SDG 14 and a current economic reliance on marine tourism, we show that 18 prioritizing increased benefits to SIDS from sustainable development of marine resources includine 19 tourism (SDG 14.7) provides the most direct co-benefits to other SDGs. When considering indirect co- 20 benefits, reducing marine pollution (SDG 14.1) emerged as an key supporting target to achieve other 21 important ocean targets. In order to support sustainable ocean development, we show that Aruba 22 depends on international support through mitigating climate change (SDG 13) and developing 23 international partnerships (SDG 17) as well as promoting sustainable economies (SDG 8), terrestrial 24 conservation (SDG 15), building strong institutions (SDG 16) and promoting sustainable consumption 25 and production practices (SDG 12) domestically. Using SDG interlinkages as a guide for institutional 26 cooperation, we find that the Aruban institutions with the most potential to coordinate action for 27 sustainable ocean development are those that coordinate economic, social, and international policy, 28 rather than institutions specifically focused on environmental policy. Our results provide insight for 29 sustainable development planning across small island states where ocean resources are key for 30 development priorities.


32
The UN Sustainable Development Goals (SDGs) were envisioned as interrelated, recognizing the deeply 33 connected world we live in and that a transition to a sustainable society requires complementary 34 dynamics across natural, social, economic, and governance domains (UN 2015). However, the 35 development of planning protocols for strategically achieving the SDGs is elusive, and an emerging 36 major research theme in sustainability science is determining appropriate governance structures to 37 achieve such multi-attribute goals in the face of complex systems (Rotmans et al. 2016;Singh 2020). A 38 governance system dedicated to sustainable development must be organized to act in an 39 interconnected way, regulating the specific linkages among and within domains to promote co-benefits 40 and mitigate trade-offs among SDGs. Here, we propose and implement a governance planning 41 framework to strategically align policy priorities and governance actors to achieve the SDGs. 42 Siloed policy prescriptions that fail to adopt integrated perspectives across social-ecological systems can 43 be ineffective or counterproductive , as sustainable development requires cross-scale 44 and, importantly, for operational planning, cross-institutional cooperation (Rotmans et al. 2016;45 Biermann et al. 2017). As an example of failing to integrate across social-ecological dimensions, policies 46 focused on protecting and growing natural capital can backfire if they enhance social inequalities and 47 ultimately undermine the legitimacy of institutions to resource users (Christie 2004). The importance of 48 social and governance considerations in effective, sustainable development projects is a particularly 49 important issue for the ocean and coastal systems where the top-down enforcement of large ocean 50 spaces can be capacity-limited and voluntary compliance is often essential (Gill et al. 2017). Conversely, 51 policies to decrease social inequity in resource-dependent communities can fail if policies do not 52 adequately account for resource supply and dynamics, such as when capacity-enhancing subsidies are 53 used to support fishing communities, and this contributes to long-term fisheries decline and collapse 54 (Cisneros-Montemayor et al. 2020). Though our comprehension and ability to represent the complexity 55 that underlies sustainability is increasing, our ability to translate this into effective policy planning and 56 implementation remains elusive. 57 Our planning method builds on and integrates the transition management framework and literature on 58 SDG interrelationships, two fields that are influential in sustainability studies but have thus far not been 59 integrated. Here, we focus on two scales within the transition management framework, i) strategic 60 scales -the priorities set at the level of values and visions, and ii) tactical scales -the institutions and 61 organizations mandated to achieve the visions (Loorbach 2007;Rotmans et al. 2016). The transition 62 management framework focuses on coordinating these multiple levels to increase the probability of 63 achieving desired outcomes and reduce the likelihood of misaligned and counterproductive results. The 64 SDG interrelationships research has been conducted across multiple countries and SDG areas, mainly 65 focusing on identifying synergies or trade-offs among SDGs, and the context in which they may occur. 66 (Nilsson et  framework thus provides a structure to plan sustainable development governance, and SDG interlinkage 68 analysis can "map out" the operating space that a governance system will need to function in (Singh 69 2020). We specifically focus on interlinkage frameworks first trialed for the ocean, that emphasize 70 categorical differences in kinds of interlinkages, and rely on structured expert elicitation and literature 71 review (Singh et al. 2018). We used a transdisciplinary approach combining academic methodologies 72 and local knowledge holders from Aruban civil service and local nonprofits. The categories of the 73 interlinkage framework differtiate where relationships among SDG targets are co-benefits or trade-offs, 74 where a target is a pre-requisite for another or if it is optional for another, and where a relationship 75 holds regardless of context or not.In this study, we identify the SDG targets that government agencies 76 are responsible for and devise collaborative institutional networks to regulate and manage the critical 77 areas that promote or hinder specific SDG Ocean targets. The resulting network represents a new 78 governance system organized around prioritized SDGs and their interconnections. 79 We develop this planning method for sustainable development planning in Aruba, a Small Island 80 Developing State (SIDS) prioritized within SDG 10 focussing on equality across countries and within SDG 81 14 focusing on sustainable marine development. Additionally, Aruba has established a government 82 commission (SDG Commission of Aruba) to develop guidance towards achieving the SDGs in the country 83 by forming partnerships across government, non-governmental organizations, and private industry. 84 Around 99% of Aruba's total territory is ocean, which is central to Aruban culture and generates 90% of 85 economic activity through coastal tourism (Vaslet and Renoux 2016 Ocean targets (besides SDG 14.6). Experts indicated that governance context (e.g., policy 112 implementation) was the most prominent factor regulating whether context-dependent co-benefits 113 were realized ( Figure S1). 114 indicates the number of relationships originating from or receiving relationships. The origin of a 117 relationship between SDG targets are indented, and the receiving end of the relationship extends out 118 further. Different colors represent different kinds of relationships, and darker shades represent greater 119 agreement among experts. SDGs are ordered by the number of relationships received by SDG Ocean 120 targets, with the SDG with the highest number of receiving co-benefits at the top of the figure and 121 following SDGs ordered clockwise from there. Only relationships with at least 2/3 agreement are shown. 122 123 Though increased economic benefits to SIDS (SDG 14.7) was determined to be the most important SDG 124 Ocean target producing co-benefits to other SDGs, through IO models we determined that reducing 125 marine pollution (SDG 14.1) contributes the most towards SDG 14.7 co-benefits among the SDG Ocean 126 targets, considering interdependencies among SDG 14 targets (Table S1). We also found that reducing 127 marine pollution is important in contributing to co-benefits from marine protection (SDG 14.5) and 128 restoration (SDG 14.2) (Table S1). In particular, reducing marine pollution is the most important 129 prerequisite for producing co-benefits through marine restoration (SDG 14.2), reduing acidification 130 impacts (SDG 14.3), and marine protection (SDG 14.5) (Table S2). Proper governance context (e.g., the 131 implementation of policy) was considered as the most prominent factor in regulating whether a co-132 benefit/optional/context-dependent relationship was realized ( Figure S1). 133 Prioritizing SDGs for the Oceans institutions (SDG 16), and sustainable consumption and production practices (SDG 12) also provide 152 many co-benefits for achieving ocean targets. Less prominent (in terms of the number of co-benefits) 153 were sustainable cities and communities (SDG 11), resilient infrastructure (SDG 9), clean energy systems 154 (SDG 7), and clean water and sanitation (SDG 6). Experts also identified the top two co-beneficial SDGs 155 (climate action and international partnerships) as the most essential prerequisites across the SDG Ocean 156 targets that contribute the most benefits across SDGs (14.7 -sustainable marine development, 14.1 -157 reducing marine pollution, 14.2 -marine restoration, and 14.5 -marine protection). 158 Ensuring sustainable consumption and production practices (SDG 12) and achieving decent jobs and 159 economic growth (SDG 8) have the largest number of prerequisite co-beneficial relationships with all 160 SDG Ocean targets. Sustainable cities and communities (SDG 11), conserving life on land (SDG 15), 161 international partnerships (SDG 17), sustainable infrastructure (SDG 9), clean energy (SDG 7), and clean 162 water and sanitation (SDG 6) also provided some co-benefit/prerequisite/context-independent 163 relationships with SDG Oceans targets. In particular, SDG Oceans targets are dependent on Aruban 164 economies developing resource efficiencies (SDG 8.4), promoting sustainable tourism (SDG 8.9), 165 reducing waste generation through reduction, recycling, waste prevention and reuse (SDG 12.5). While 166 no targets among SDG 16 (peace, justice, and strong institutions) were considered to be prerequisite for 167 SDG Ocean targets by a supermajority of experts, there was strong agreement among a supermajority 168 (agreement score 0.71) that achieving policy coherence (SDG 17.14) was a prerequisite condition for 169 reducing marine pollution and restoring marine habitats, and high agreement (agreement score 170 between 0.5 and 0.66) that it is a prerequisite condition for all other SDG Ocean targets. 171 Considering only co-benefit/optional/context-dependent relationships, international climate action 172 (SDG 13), international partnerships (SDG 17), peace, justice, and strong institutions (SDG 16), and conserving life on land (SDG 15) provided the greatest number of relationships with SDG ocean targets. 174 Other SDG Ocean targets, jobs, and economy (SDG 8), and clean water and sanitation (SDG 6) also 175 provided context-dependent co-benefits with SDG Ocean targets. Experts indicated that governance 176 context (e.g., policy implementation) was the most prominent factor regulating whether context-177 dependent co-benefits were realized ( Figure S2). 178 Agreed on by a supermajority of experts, only two SDGs produced tradeoff/optional/context-dependent 179 relationships with SDG Oceans targets: jobs and economy (SDG 8) and reducing inequalities (SDG 10). 180 Sustaining per capita economic growth (SDG 8.1) and progressively achieving income growth of the 181 bottom 40% of the population above national averages (SDG 10.1) were the two SDG targets with 182 potential tradeoffs with minimizing ocean pollution (SDG 14.1), marine restoration (SDG 14.2), 183 mitigating ocean acidification impacts (SDG 14.3), and effectively protecting marine areas (SDG 14.5). As 184 with co-benefits, experts indicated that the governance context was the most prominent factor 185 regulating whether tradeoffs could be avoided ( Figure S2). In a scenario where Aruban institution structure is guided by direct regulation of SDG ocean targets (no 196 SDG relationships guide design), ten agencies must coordinate (Figure 4). The Directorate of Nature and 197 Environment (DNE) is directly responsible for helping to regulate all SDG Ocean targets and is also 198 connected to the largest number of other institutions (9) also responsible for regulating SDG ocean 199 targets. Using a battery of network centrality measures to calculate the most important institution in 200 this scenario (assuming agency importance to be determined by the most connected agency), we find 201 that all the centrality measures indicate that the DNE is the most important institution to coordinate 202 achievement of the SDG Ocean targets (see Figure S4 and Table S7). 203 In a scenario where Aruban institution structure is guided by considerations of prerequisite relationships 204 where SDG Ocean targets require other SDG targets, 34 Aruban agencies must coordinate ( Figure 4). 205 While the DNE is the only Aruban agency directly responsible for all the SDG Ocean targets in this 206 scenario, the Social and Economic Council (SEC) is directly responsible for the largest number of SDG 207 targets that are prerequisites for the SDG Ocean targets (6 SDG targets that are prerequisites). 208 Additionally, in this scenario, the SEC is connected to the largest number of other institutions (20) to 209 collaboratively regulate progress on all SDG targets needed to achieve SDG Ocean targets. Assuming 210 agency importance to be determined by the most connected agency, all the centrality measures indicate 211 that the SEC is the most important institution to coordinate the achievement of the SDG Ocean targets 212 (see Figure S5 and Table S8). 213 If the institutional structure is instead determined by considerations of all SDG relationships, including 214 gaining from all co-beneficial relationships and avoiding the potential of tradeoffs, 66 agencies must 215 coordinate ( Figure 4). Similar to the last scenario, while the DNE is directly responsible for all SDG Ocean 216 targets, the SEC is responsible for the largest number of SDG targets that affect SDG Ocean targets (13 217 SDG targets), and centrality measures again indicate the SEC as the most important institution to 218 coordinate achievement of the SDG Ocean targets (coordinating 42 other agencies, see Figure S6 and 219 Table S9). 220 is not part of the direct management scenario). 228

229
Although we find that sustainable marine use (SDG 14.7) directly contributes the most co-benefits to 230 achieving the SDGs overall in Aruba, considering indirect and cascading contributions shows reducing 231 pollution (SDG 14.1), restoring marine ecosystems (SDG 14.2), and marine protection (SDG 14.3) are also 232 important in providing diverse co-benefits across the SDGs. However, we also determine that these 233 same SDG Ocean targets receive the most co-benefits from achieving other SDGs. Importantly, these 234 SDG Ocean targets are most dependent (as determined by assessing prerequisite relationships) on 235 achieving other SDGs (including consumption and production systems and economic transformation, 236 SDGs 12 and 8, respectively) being realized. In consequence, we found that the most crucial Aruban 237 institution for coordinating regulations to achieve sustainable oceans was not an environmental agency 238 but a socioeconomic agency (the Social and Economic Council). Therefore, while our investigation into 239 the cascading roles of SDG Ocean targets show that environmentally focused targets underpin some of 240 the more economic goals -and in some ways support the frameworks for "environment-based" 241 sustainable development (Griggs et  Governance Planning in Small Island States 247 We found that the non-ocean SDGs with the highest number of co-benefits with SDG Ocean targets are: 248 climate action (13), international cooperation (11), peace, justice, and strong institutions (10), land 249 conservation (8), decent work and economic growth (8), and sustainable consumption and consumption 250 (6). These results showcase how important global cooperation is for Aruba to achieve ocean sustainable 251 development, given the scale of some key drivers of ocean environmental sustainability and industries 252 and the relative ability of small islands to mitigate their impacts. Aruban efforts to increase ocean 253 sustainability may significantly benefit by increased engagement in international diplomacy for climate 254 mitigation and international capacity development and technology transfer to Aruba (Keohane and  255 Victor 2016 whether these relationships would be tradeoffs or not. In particular, they pointed to where investment 270 was directed (whether primary, secondary, or tertiary economic sectors were invested in for economic 271 and income growth), whether policies enforcing waste reduction, recycling, and cleaner production 272 practices were followed, and whether cleaner consumption practices could be encouraged and 273 followed. Given that Aruba has seen significant economic benefits from oil and gas refining in the near 274 past, as well as the construction of desalinization processing, Aruba may choose to reinvest economic 275 opportunities in these industries. The existing infrastructure and immediate economic promises (income 276 and employment) of these industries may provide too important for Aruba to avoid in the future. 277 Though tradeoffs between economic growth and sustainable ocean development may have the 278 potential to be avoided does not mean they are easy to avoid, and Aruba may have to accept the 279 compromises and make decisions on which SDG target is more important. Scientific analysis has an 280 important role to play here in helping inform evidence-based policy decisions. 281 Designing Governance Institutions to Maximize the Potential of SDG Relationships and social agencies can play central roles to ensure ocean sustainability. Designing an integrative and 307 coherent policy for ocean sustainability will require an explicit consideration of which institutions have 308 responsibilities across the suite of sustainable ocean targets, and which institutions are most centrally 309 collaborative across relevant institutions to collaboratively achieve sustainability goals. 310 The methodology in this study directly addresses the imperative need for institutional and program 311 integration as we increasingly recognize the need for cross-scale and multidisciplinary development 312 goals. This method may eventually require a re-imagining of institutional purviews and relationships but, 313 given historical institutional architectures and inertia, in practice, this implies in the short-term an 314 increased awareness of the implications of progress within one institutions' mandate on the outcomes 315 of another' (Loorbach 2010 be an aspirational as well as an operational set of guidelines, but the latter will require specific and 368 evidence-based connections between sustainability principles and governance planning to create 369 governance systems to achieve these goals. 370

371
Overview 372 This study follows three steps along the planning structure of the transition management framework. 373 First, we undertook an expert elicitation process to prioritize SDG Ocean targets based on each target's 374 contribution to other SDG targets, including direct, indirect, and cascading effects. Second, we 375 determine interrelationships between all other SDGs and SDG Ocean targets, paying particular attention 376 to SDG targets deemed necessary to achieve SDG 14 targets. This information effectively outlines the 377 strategic policy arena according to the transition management framework (Singh 2020) and indicates the 378 scope of social-ecological relationships that a governance system must be built around. Finally, we 379 identify the SDG areas that different Aruban government agencies are responsible for regulating action 380 towards and identify scenarios of institutional networks that are informed by SDG relationships. These 381 scenarios connect the strategic and tactical scales within the transition management framework (Singh 382 2020). 383 as a key challenge, and no capacity-enhancing subsidies are provided to fishers (in compliance with SDG 399 14.6). Aruba has a terrestrial national park that extends from its rugged north-eastern coast to the only 400

Aruba and the SDGs
Ramsar site on the south-western coast. Since 2019 Aruba also has four multi-use protected areas, but 401 these protected areas do not extend into the ocean (SDG 14.5). Though marine tourism has such high 402 economic value, it currently is not necessarily sustainable (part of the focus of SDG 14.7) as tourism in 403 Aruba focuses on warm weather and clean, sandy white beaches instead of a healthy marine ecosystem. 404

405
A workshop was convened to 1) prioritize SDG 14 targets based on maximizing the production of co-406 beneficial relationships across all other SDG targets; and 2) determine the SDG targets that promote co-407 beneficial relationships with ocean targets, while also identifying SDG targets that can act as tradeoffs 408 with ocean targets. While the first objective was set to determine ocean priorities, the second was to 409 understand the SDG support structure needed to ensure that ocean priorities can be met. Determining 410 the structure of Aruban institutions required to support ocean SDG priorities relies on this latter 411 objective being completed. 412 The workshop was held over ten days, with dedicated sessions on the relationships and effects of 413 progress on the SDG Ocean's targets to other SDG targets and vice versa. of having experts in multiple groups is that high agreement across experts is more robust, as there is 488 greater independence among the expert responses, akin to increasing the degrees of freedom in a 489 statistical design. Once all the experts provided their assessments, their answers were compiled to 490 generate maps of expert variation in responses. 491 Experts were asked to provide SDG target relationships, as well as indicate -whenever they showed an 492 optional/context-dependent relationship -the contextual element that regulated the relationship. 493 Experts were instructed to report whether the relationship was dependent on ecological factors 494 (defined as non-human biotic and abiotic conditions), economic factors (defined as the financial, market, 495 income, and labor conditions), social factors (defined as issues related to social norms, demographics, 496 and non-monetary social conditions), and governance factors (defined as institutions, policy, law, and 497 decision-making bodies). 498

499
Once all expert responses were collected, they were compiled and coded through a winner-takes-all 500 system of classification (except when "neutral" relationships were most prevalent), with the level of 501 agreement quantified. For example, if out of 20 experts, 15 thought a relationship was co-502 benefit/optional/context-dependent, while 3 of the other five thought the relationship was co-503 benefit/optional/context-independent. The remaining two thought the relationship was co-504 benefit/prerequisite/context-independent. The relationship was coded as co-benefit/optional/context-505 dependent, with an agreed level of 0.75 (15/20). Similarly, if out of 20 experts, five experts thought a 506 relationship was co-benefit/optional/context-dependent, two thought the relationship was co-507 benefit/optional/context-independent. The rest felt the relationship was neutral. The link was coded as 508 co-benefit/optional/context-dependent, with agreement level 0.25 (5/20). 509 To avoid the inclusion of spurious non-neutral relationships or non-neutral relationships with greater 510 expert disagreement than agreement, we set a threshold of agreement from which to continue our 511 analysis. We chose a supermajority of expert agreement (2/3 agreement) as a threshold to ensure that 512 our analysis focused only on those relationships with little disagreement. Once we determined our final 513 set of non-neutral relationships, we determined priority areas for both SDG ocean targets that are most 514 cross-cutting for all other SDGs as well as SDGs that are most related to the SDG ocean targets. 515 Quantifying the SDG ocean targets in terms of their contribution across other SDGs included an 516 additional step because we assessed the SDG ocean targets against each other, and therefore could 517 assess direct and secondary indirect relationships across SDGs. To calculate the total contribution of 518 achieving the SDG ocean targets across all other SDGs, we adopted Input-Output (IO) models. This 519 method is ordinarily used to estimate the contribution of specific economic sectors to the economy as a 520 whole by linking the production of each sector (or in this case, SDG target) to the consumption of others 521 (Leontief 1951). In this way, for example, the ripple effects of some industries can be particularly 522 important for an economy when their production is an essential input for other industries that may 523 themselves be important for still other industries. (For example, steel production used as input into ship 524 construction that is required for the shipping and trade industries). We adapt this method to calculate 525 the relative co-beneficial productive importance of each SDG ocean target, accounting for all ripple 526 effects stemming from interconnections among SDGs. We calculate the Leontief inverse using the 527 formula 528 where x is the relative co-beneficial productive importance of each SDG ocean target, accounting for the 529 sum of ripple effects from all other SDG ocean targets, I is the identity matrix, A is the matrix of 530 intermediate outputs (i.e., the proportion of SDG Ocean co-benefits from achieving a given SDG Ocean 531 target that leads to further co-benefits across the SDGs), and d is the total output (i.e., overall SDG 532 target benefits). Calculating the importance of interlinked SDG ocean targets was done for all co-533 beneficial relationships, for only co-benefit/prerequisite relationships, and only co-534 benefit/optional/context-dependent relationships. Co-benefit/prerequisite relationships are arguably 535 the most important, as other SDG targets cannot be achieved without the achievement of the specified 536 SDG ocean target. Co-benefit/optional/context-dependent relationships are potential co-benefits that 537 are realized if other conditions are met. 538 Quantifying the relationships of other SDGs to the SDG ocean targets were more straightforward, as we 539 could not consider their interaction/indirect contributions to the ocean targets, because we did not look 540 at how all other SDGs interacted with each other. We, therefore, summed the number of the different 541 kinds of co-beneficial and tradeoff relationships with the SDG ocean targets. 542 Once all SDG relationships were quantified, data summaries were prepared and sent out to the original 543 experts for vetting. This stage of elicitation was carried out over email. Experts were sent files with 544 graphics summarizing relationships and captions describing trends. Experts were asked to provide 545 feedback (particularly if they did not agree with some findings) or suggestions for describing prominent 546 results. During the vetting period, no experts identified disagreement with the findings, and some 547 provided extra context to describe findings. After vetting, we compiled our final dataset of SDG 548 relationships. SDG relationships were graphically represented in circos plots (using the R package 549 circlize, Gu 2014), a multivariate network graphing technique used often in genomics research to 550 organize nodes in nested structures (in our case nesting SDG targets within SDGs) and represent all links 551 between nodes. 552 All optional/context-dependent relationships, as determined by individual experts, were categorized as 553 dependent on environmental, social, economic, or governance dimensions. We tallied up all instances of 554 these considerations and determined what factor regulates context-dependent relationships. We 555 plotted the results using Sankey diagrams, using the R package SanKey (Csárdi and Weiner 2017). 556

557
To determine the structure of government institutions informed by SDG interconnections to promote 558 sustainable oceans, we first categorized the Aruban government agencies based on the SDG area(s) they 559 are responsible for. To do this, first, we reviewed the websites for each government agency (grouped 560 under five distinct government ministries) and classified them as contributing to individual SDG targets 561 across all SDG goals. We organized the institutions based on the description of responsibilities, as stated 562 on the website for each institution. We did not include the SDG Commission of Aruba in this analysis 563 because they have no regulatory authority over the SDG areas but instead are responsible for 564 connecting with business and non-governmental organizations to promote the SDGs. This list was sent 565 to the experts from the earlier workshop (who collectively work in, or have considerable experience or 566 familiarity with, all the Aruban ministries), to vet the classification for accuracy. Vetting was done over 567 email, specifically asking experts if our classification system captured the role of Aruban institutions in 568 practice (Singh et al. 2018). Over two iterations, our database of Aruban institutions was refined and 569 finalized. 570 Because we were interested in building institutional structures organized by SDG relationships, we 571 created interaction matrices of institutions regulating SDG targets that have connections with the SDG 572 ocean targets (in that direction). We considered three scenarios of institutional arrangement: a situation 573 where only direct institutional regulation was considered (so no SDG relationships were taken into 574 account), a condition where co-benefit/prerequisite relationships were considered (as they are needed 575 to achieve the ocean SDG targets), and a case where all SDG relationships were considered. The case 576 where only direct institutional regulation was considered most strongly resembles the current situation. 577 The prerequisite situation models an institutional structure minimally needed to ensure the 578 achievement of the SDG ocean targets. Finally, the situation with all SDG relationships models an 579 institutional arrangement that will provide the highest potential to achieve the SDG ocean targets by 580 capitalizing on co-benefits (both through promoting context-independent co-benefits and implementing 581 policy to realize the potential of context-dependent co-benefits) and mitigating tradeoffs. 582 In every situation, we modeled an ideal situation where all institutions that help regulate a specific SDG 583 target are in communication with each other. This assumption may not be realistic, but we are 584 interested in how SDG interlinkages change institutional design rather than assessing existing 585 institutional collaboration. From the results, we determine the institutions most connected with SDG 586 targets and most-connected with other institutions. The first indicates a measure of how important the 587 institution is as a regulator for ocean sustainability across targets, and the second suggests a measure of 588 how important that institution is as a collaborating entity, ensuring consistent policy planning across 589 institutions. On top of these metrics, we use a battery of measures of network centrality to determine 590 the most crucial institution based on network structure. To select the centrality measures, we first use 591 principal components analysis (Husson et al. 2017) and t-Distributed Stochastic Neighbor Embedding 592 analysis (Van Der Maaten 2014) to determine the centrality measures that are most informative given 593 the institutional network structure (see Figure S4). We use the CINNA package in R to identify the proper 594 centrality measures (Ashtinani 2019). We use the resulting four centrality measures to establish the 595 most important institutions, and compare these results with our simple counts presented above. 596 Institutional networks were developed in the R package igraph (Csardi and Nepusz 2006). where all co-benefits and tradeoffs between SDG targets are considered.      Table S7. Centrality measures for the importance of Aruban institutions in regulating progress on the various SDG Ocean targets, in a scenario where relationships between SDG targets are not considered. The Social and Economic Council (SEC, institution code I104) is not included in this scenario. The topological coefficient is a relative measure for the extent to which a node shares nodes with other nodes, so low values here indicate that an institution is connected with other institutions that are not otherwise connected. We interpret that as suggesting that institutions with low topological coefficient scores are more important for coordinating activities across institutions that are otherwise not connected with the broader institutional system. The average distance is a measure of how far, on average, a node is from other nodes, so a lower number indicates a more central node.  Table S9. Centrality measures for the importance of Aruban institutions in regulating progress on the various SDG Ocean targets, in a scenario where all co-benefits and tradeoffs between SDG targets are considered. The topological coefficient is a relative measure for the extent to which a node shares nodes with other nodes, so low values here indicate that an institution is connected with other institutions that are not otherwise connected. We interpret that as suggesting that institutions with low topological coefficient scores are more important for coordinating activities across institutions that are otherwise not connected with the broader institutional system.