Marine protected areas (MPAs) are spatial management tools used to protect and restore marine ecosystems, with the purpose to promote and maintain ocean ecosystem services (Allison et al., 1998; Lubchenco et al., 2003; Halpern et al., 2010). Different approaches have been used to design MPAs depending on the political and economic context of each region. Traditionally, MPAs have been implemented as a result of governmental or local community decisions (Gleason et al., 2010). However, there are recent examples of planning processes to design MPAs networks that are ecologically connected and managed as a system (Airamé et al., 2003; Fernandes et al., 2005; Lowry et al., 2009; Saarman et al., 2013; Hopkins et al., 2016).

Regional MPA network design processes are more common nowadays (e.g., Australia, California, USA, United Kingdom, among others) but mostly limited within the political boundaries of a country. However, the oceans are dynamic environments connected by water currents and animal movements; thus, neighboring countries share oceanographic processes that connect marine populations and ecosystems (Carr et al., 2003; Wilkinson et al., 2004; Torres-Moye et al., 2013). In these shared coastal environments, the actions of one country may affect the other, e.g., through the establishment of conservation measures, pollution control, or the introduction of invasive species (Mazor et al., 2013). Therefore, protection of marine ecosystems requires management at an eco-regional scale (units large enough to encompass ecological or life history processes) through the establishment of transboundary MPAs networks (Sandwith and Besançon, 2005; IUCN World Commission on Protected Areas, 2008; Guerreiro et al., 2010; Torres Moye, 2012; Giakoumi et al., 2013; Jessen et al., 2016).

Southern California (henceforth “California”), USA, and the northwest of Baja California (henceforth “Baja California”), Mexico, share the Ensenadian eco-region that extends from Santa Monica Bay, USA, to Punta Eugenia, Mexico (Blanchette et al., 2008). This transboundary eco-region represents a biogeographic area connected by the southern end of the California Current system (Di Lorenzo, 2003), a highly productive region of the northwestern Pacific Ocean (Dailey et al., 1993; Castro, 2010). Even though these two countries share the same eco-region, their social, economic, and management context differ across the border. While the south coast of California is one of the most populated regions of the world, Baja California holds a lower population density and mostly concentrated at the north, in the coasts of Tijuana and Ensenada. As for fisheries, in California, commercial fishing permits are granted to individual fishers or vessels in contrast to the concessions or permits for fisher groups or cooperatives in Baja California (Enríquez-Andrade et al., 2007). In both regions, recreational fisheries are open access by individual permits, but of greater economic importance in California. In the context of marine conservation, California recently implemented a network of MPAs, following the Marine Life Protection Act (MLPA), based on scientific principles and stakeholder involvement (Gleason et al., 2013; Kirlin et al., 2013; Saarman et al., 2013). However, Baja California still lacks a network of MPAs or a systematic conservation planning (Margules and Pressey, 2000) effort and has less biophysical information of its coastal systems compared to California.

Many considerations are used to identify areas for conservation in a spatial prioritization exercise (Moilanen et al., 2009). The design of an ecologically functional and socio-economically viable network of MPAs requires spatial information that captures the ecological and social complexity of the coastal systems (Naidoo et al., 2006; Klein et al., 2008a; Ban and Klein, 2009). Yet, very few regions around the world have detailed spatial information (Ban et al., 2009; Giakoumi et al., 2011, 2013). In many cases the distribution of habitats is used as a surrogate for species distribution, a cost-effective method to identify conservation priority areas (Ward et al., 1999; Roberts et al., 2001). On the other hand, fishing effort and/or commercial landings are used to quantify the socio-economic costs from the implementation of a network of MPAs (Naidoo et al., 2006; Klein et al., 2008a,b; Ban and Klein, 2009; Giakoumi et al., 2013; Mazor et al., 2013). Both ecological and socio-economic information are essential to identify conservation priority areas that minimize the possible socio-economic conflicts. Recently, other aspects have been included for the design of a network of MPAs. As an example, cumulative human impact maps (e.g., Halpern et al., 2008, 2009; Selkoe et al., 2009; Ban et al., 2010; Micheli et al., 2013) have been used to prioritize habitats in good condition (Klein et al., 2013), and management surrogates (Balmford et al., 2004; Naidoo et al., 2006) to quantify the enforcement costs (Davis et al., 2015).

In this study, we present a novel approach for the design of a network of MPAs in a transboundary context using multiple considerations. Our prioritization exercise aims to use the same design criteria as in California to design a regionally feasible network of MPAs for Baja California that is ecologically linked with California's network. This study shows how, even in a region with limited spatial information, the available conservation principles, mapping, and planning tools can still be used to generate spatial conservation plans.

Globally, we are in a race to protect 10% of the marine ecosystems (Secretariat of the Convention on Biological Diversity, 2014) through “ecologically representative” MPAs. However, most countries are slowly reaching that target. Among other obstacles are the scarcity of spatial information (e.g., Giakoumi et al., 2013), and/or the lack of initiatives that allow nations to take decisions based on science and the minimization of socio-economic conflicts. Also, in many cases existing MPAs have inadequate designs to work as functional networks (Spalding and Hale, 2016), or were not established as eco-regional transboundary networks of MPAs (e.g., Guerreiro et al., 2010). In this study we assembled available ecological and socio-economic information and collected new data in order to identify priority zones for a future network of MPAs in Baja California, linked with the one in southern California. To achieve this we used the MLPA scientific criteria and the features that best represent the social and management context of our region. This study builds upon the existing literature on marine conservation planning in an eco-regional and transboundary context (e.g., Giakoumi et al., 2013; Mazor et al., 2013, 2014), as well as the importance of using multiple considerations in a prioritization exercise (e.g., Klein et al., 2013).

All of the MPA networks from the different scenarios met the habitat representation criteria (to conserve a target percentage of the marine habitats), and achieved a relatively high percentage of the other MLPA guidelines. In all the scenarios, including the enforcement scenario that incorporates all the considerations (Figure 4), the northernmost MPA of Baja California fulfilled the recommended size and distance from the southern MPAs of California (combined South La Jolla SMCA & SMR) and also included the 9 continent habitats; therefore, these networks could protect shared species, populations, and communities, and facilitate connectivity between both regions. A future transboundary network could be ecologically and economically beneficial for the Ensenadian eco-region and it is acknowledged that such an approach could promote the cooperation and the exchange of information and technology between both countries (UNEP-WCMC, 2008; McCallum et al., 2015).

At a regional level we were able to design networks that integrated multiple considerations (scenarios 2–4) and represent low opportunity costs for fishers and aquaculture investors. The scenario that incorporates all the ecological, social and management considerations (scenario 4) would represent an increase of only 3.8% in the socio-economic cost compared to planning without including habitat condition and enforcement capacity (scenario 2). This result supports previous recommendations to minimize conflicts by incorporating opportunity costs in marine conservation planning (Stewart and Possingham, 2005; Klein et al., 2008a, 2013; Ban et al., 2009; Mazor et al., 2014), with the aim of designing viable networks of MPAs. Planning based on multiple considerations also helped identifying priority conservation areas, since less optimal areas were found, and those that did achieve all the objectives were prioritized. This finding contributes to the existing knowledge showing that not only the costs are important in the selection of priority areas (Naidoo et al., 2006; Ban et al., 2009; Mazor et al., 2014), but also adding multiple considerations favors the selection of more priority areas.

Human population proved to be a relevant factor for the spatial arrangement of the priority planning units. The north zone, which has the highest population density, registered a higher opportunity cost and a lower habitat condition, and as a result less priority planning units emerged there, compared to the central and south zones. In contrast to Klein et al. (2013), the spatial arrangement of the priority planning units radically changed when we incorporated habitat condition data. This may be explained by the fact that in Baja California there is a higher contrast in population density between the three zones, and highlights the importance of including habitat condition in prioritization exercises for areas with demographic differences.

When the enforcement capacity was included (scenario 4), the spatial arrangement radically changed, and a more even representation was achieved. Because the north has a better enforcement capacity than the south, more priority areas emerged in this zone. These results show an indirect way to estimate enforcement costs in areas managed through fishing concessions or permits and can be applied or adapted to other similar regions; however, since enforcement capacity was an important factor, it becomes relevant the need to improve its estimation.

This study represents the first effort to generate and integrate all the available and necessary information for the systematic conservation planning of the MPRC1 in Baja California. We illustrated a practical method to map marine habitats in data-poor areas and contribute with a novel approach to quantify the opportunity and management costs in areas managed through fishing concessions or permits. However, as in most marine spatial conservation exercises, we were faced with some limitations. Although we used the same four scientific guidelines from the MLPA, other more specific design criteria were not included (e.g., minimum habitat area set). Also, the available historical Google Earth images were more abundant in the north zone making the fine marine habitat mapping more challenging for the center and south zones. And even though we lessened these constraints employing other sources of information and direct field observations, we could not map some habitats such as deep rocky reefs and seagrass beds. On the other hand, for the opportunity costs we did not include other important fisheries resources such a lobster and finfish; however, our approach for the combined opportunity cost allowed us to consider the three most important activities in the region. In previous studies, it has been found that planning based on a single sector produces higher costs for other sectors, and therefore less efficient prioritization results (Mazor et al., 2014). Since monetary data were not available for each sector, we used an indirect measure to quantify their relative importance and assumed that the commercial fishing is a source of greater economic income followed by aquaculture and sport fishing. Despite some limitations, this is one of the few marine spatial planning exercises that include multiple measures of cost (e.g., Giakoumi et al., 2013; Mazor et al., 2013).

For the habitat condition classification, we could only use 13 of the 25 human impacts from the California Current study (Halpern et al., 2009). The lack of information of some local human impacts may overestimate the habitat condition in some sites (e.g., Torres-Moye and Escofet, 2014); nevertheless, this represents one of the few works that accounts for marine habitat condition. Finally, for the enforcement capacity we used an indirect measure which might not represent the real capacity since this can vary depending on the fishermen's organization and their available technology (e.g., radars and surveillance fast boats). However, our approach combines three measures compared to other similar works that are generally based in one metric (distance) (Ban et al., 2009; Davis et al., 2015).

With an adaptive ecosystem-based management approach (Katsanevakis et al., 2011), we recognize the need to continue generating spatial data for Baja California. This exercise identified some biophysical and socio-economic information gaps that could be used to inform governmental or academic priorities for future studies (programs designed to collect these data). Any further information which might improve our biophysical or socio-economical characterization could vary the spatial arrangement of the priority areas, and better support decision making in the region.

In this study we used the scientific experience of California for the design of MPAs; however, policies related to marine conservation and fisheries management differ across the border. Therefore, a future marine conservation planning process in Baja California might be different than that used in California. Many fishermen communities completely depend on the coastal marine resources, particularly those living south of Ensenada in sparsely populated areas. Restricting fishing activities in these communities would have negative consequences for their economic and social development. However, nowadays fishermen in Baja California are aware that their resources are limited and that they face new challenges such as climate change. Some local fisherman communities are working together with NGOs and scientists to co-manage their resources (Álvarez et al., 2015) by establishing community marine reserves or restocking marine populations (Micheli et al., 2012). Recently (December 2016), the Pacific islands from the MPRC1 were included as part of a Biosphere Reserve, representing the first legally protected marine areas in this region. This trend opens a window of opportunity to initiate a dialogue between local fishing communities, NGOs, government institutions, and the scientific community toward the design of a network of MPAs that is ecologically functional and beneficial for local communities.

In this work we show how, despite the constraints of a data-poor area, we were able to identify priority areas for conservation, which are ecologically linkable to an existing network of MPAs and socio-economically viable in a regional context. This exercise contributes to advancing global initiatives to design transboundary networks of MPAs managed in an eco-regional scale (Guerreiro et al., 2010; Mackelworth, 2012, 2016; Rosen and Olsson, 2013; Jessen et al., 2016). We hope that this work will lay the foundation to guide a future process of marine conservation planning in Baja California, including the collaboration between the north of Baja California and the south of California, with the aim to jointly coordinate initiatives to understand, manage, and conserve shared resources.

NA and GT designed the initial project with the final review of GS, GM, and FM. NA and GT acquired data. The analysis of the data and the interpretation was conducted by NA. NA drafted the work and was revised by all the authors. All authors approve the version to be published and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.