To Achieve Big Wins for Terrestrial Conservation, Prioritize Protection of Ecoregions Closest to Meeting Targets

Abstract

•We develop a global strategy to prioritize ecoregions for protection over time•Our strategy could protect twice as many ecoregions as business as usual by 2030•Our strategy could halve the ecoregion protection gap compared with business as usual•Considering the ongoing race between conversion and protection of habitats is key To save species from extinction, conservation is racing to establish new protected areas (PAs) before natural habitats are lost. We thus need a strategy to efficiently allocate conservation resources toward PAs. This strategy also has to be global to meet the international targets for PAs set by the Convention on Biological Diversity (CBD). One key aspect of these targets is that all broad ecosystem types (called ecoregions) should have a minimum level of protection equal to an area target. Here, we show that simply prioritizing ecoregions that are the closest to meeting the CBD's area target for PAs performs almost four times better than the “business as usual” approach: under the current annual budget for PAs, up to 260 more ecoregions that meet their targets by 2030. Our work addresses the ongoing race between habitat conversion and habitat protection, a factor seldom accounted for in the PA literature despite real-world implications. Most of the terrestrial world is experiencing high rates of land conversion despite growth of the global protected area (PA) network. There is a need to assess whether the current global protection targets are achievable across all major ecosystem types and to identify those that need urgent protection. Using recent rates of habitat conversion and protection and the latest terrestrial ecoregion map, we show that if the same approach to PA establishment that has been undertaken over the past three decades continues, 558 of 748 ecoregions (ca. 75%) will not meet an aspirational 30% area protection target by 2030. A simple yet strategic acquisition plan that considers realistic futures around habitat loss and PA expansion could more than double the number of ecoregions adequately protected by 2030 given current funding constraints. These results highlight the importance of including explicit ecoregional representation targets within any new post-2020 global PA target. Most of the terrestrial world is experiencing high rates of land conversion despite growth of the global protected area (PA) network. There is a need to assess whether the current global protection targets are achievable across all major ecosystem types and to identify those that need urgent protection. Using recent rates of habitat conversion and protection and the latest terrestrial ecoregion map, we show that if the same approach to PA establishment that has been undertaken over the past three decades continues, 558 of 748 ecoregions (ca. 75%) will not meet an aspirational 30% area protection target by 2030. A simple yet strategic acquisition plan that considers realistic futures around habitat loss and PA expansion could more than double the number of ecoregions adequately protected by 2030 given current funding constraints. These results highlight the importance of including explicit ecoregional representation targets within any new post-2020 global PA target. The 2020 Strategic Plan for Biodiversity states that 17% of Earth's land area should be placed under protection and that protected areas (PAs) and other effective area-based conservation measures must represent the current diversity among habitats and species within their borders (Convention on Biological Diversity [CBD] Aichi Target 11).1Convention on Biological DiversityAichi Biodiversity Targets.2016https://www.cbd.int/sp/targets/Google Scholar Ecological representation is a central pillar of this target, recognizing that while it may not be possible to save everything on Earth, nations should strive to preserve a representative sample of all ecosystems and habitat types.2Olson D.M. Dinerstein E. The global 200: a representation approach to conserving the Earth's most biologically valuable ecoregions.Conserv. Biol. 1998; 12: 502-515Crossref Scopus (1030) Google Scholar As a consequence, ecological representation is reported by most nations and global institutions.3UNEP-WCMC and IUCNProtected Planet Report 2016.2016https://www.unep-wcmc.org/resources-and-data/protected-planet-report-2016Google Scholar At present, there are large gaps in the PA network such that many ecosystem types and species have little or no formal protection;4Venter O. Fuller R.A. Segan D.B. Carwardine J. Brooks T. Butchart S.H. Di Marco M. Iwamura T. Joseph L. O'Grady D. et al.Targeting global protected area expansion for imperiled biodiversity.PLoS Biol. 2014; 12: e1001891Crossref PubMed Scopus (307) Google Scholar this pattern holds at national5Powell G.V.N. Barborak J. Rodriguez M. Assessing representativeness of protected natural areas in Costa Rica for conserving biodiversity: a preliminary gap analysis.Biol. Conserv. 2000; 93: 35-41Crossref Scopus (99) Google Scholar,6Oldfield T.E.E. Smith R.J. Sr H. Leader-Williams N. A gap analysis of terrestrial protected areas in England and its implications for conservation policy.Biol. Conserv. 2004; 120: 303-309Crossref Scopus (103) Google Scholar and global7Rodrigues A.S. Andelman S.J. Bakarr M.I. Boitani L. Brooks T.M. Cowling R.M. Fishpool L.D. Da Fonseca G.A. Gaston K.J. Hoffmann M. et al.Effectiveness of the global protected area network in representing species diversity.Nature. 2004; 428: 640-643Crossref PubMed Scopus (971) Google Scholar scales. Ecoregions are the preferred unit when mapping ecosystems globally.1Convention on Biological DiversityAichi Biodiversity Targets.2016https://www.cbd.int/sp/targets/Google Scholar,3UNEP-WCMC and IUCNProtected Planet Report 2016.2016https://www.unep-wcmc.org/resources-and-data/protected-planet-report-2016Google Scholar Those that are not yet protected to the desired level (e.g., 17%) but could still meet the protection target (e.g., having <83% converted land) are faced with a race to establish new PAs before natural habitats are degraded.8Montesino Pouzols F. Toivonen T. Di Minin E. Kukkala A.S. Kullberg P. Kuustera J. Lehtomaki J. Tenkanen H. Verburg P.H. Moilanen A. Global protected area expansion is compromised by projected land-use and parochialism.Nature. 2014; 516: 383-386Crossref PubMed Scopus (214) Google Scholar, 9Watson J. Venter O. Mapping the continuum of humanity's footprint on land.One Earth. 2019; 1: 175-180Abstract Full Text Full Text PDF Scopus (11) Google Scholar, 10Jones K.R. Venter O. Fuller R.A. Allan J.R. Maxwell S.L. Negret P.J. Watson J.E.M. One-third of global protected land is under intense human pressure.Science. 2018; 360: 788-791Crossref PubMed Scopus (245) Google Scholar This race makes the strategic allocation of limited conservation funds a priority for achieiving global biodiversity targets. We developed a dynamic protection strategy that achieves maximum representation of ecoregions by 2030 while accounting for ongoing habitat conversion. We chose 2030 as a time frame because it is the current time horizon set for the Sustainable Development Goals (under which any future CBD PA target must be embedded). Given the uncertainty regarding the future of international protection targets, we used the current requirements of Aichi Target 11, i.e., to protect 17% of each global ecoregion up to 2020 and an aspirational 30% target by 2030, which is now being widely proposed by the conservation community.11Dinerstein E. Vynne C. Sala E. Joshi A.R. Fernando S. Lovejoy T.E. Mayorga J. Olson D. Asner G.P. Baillie J.E.M. et al.A global deal for nature: guiding principles, milestones, and targets.Sci. Adv. 2019; 5: eaaw2869Crossref PubMed Scopus (155) Google Scholar We tested several simple but robust PA expansion strategies to determine which method best achieves these goals. In some ecoregions, additional protection is needed to meet the representation goals,12IUCN and UNEP-WCMCThe World Database on Protected Areas.2016https://www.protectedplanet.net/Google Scholar but unprotected land may already be too modified to be suitable for conservation because it is unlikely to be successfully restored.13Mappin B. Chauvenet A.L.M. Adams V.M. Di Marco M. Beyer H.L. Venter O. Halpern B.S. Possingham H.P. Watson J.E.M. Restoration priorities to achieve the global protected area target.Conserv. Lett. 2019; 12: e12646Crossref Scopus (23) Google Scholar Incorporating land-conversion processes into our analysis reduces the amount of land available for protection and ultimately implies that some ecoregions will not have enough unconverted land remaining to meet the area protection target by 2030. We first established a “business as usual” (BAU) strategy for PA expansion within each ecoregion on the basis of the observed rates of land protection12IUCN and UNEP-WCMCThe World Database on Protected Areas.2016https://www.protectedplanet.net/Google Scholar and conversion between 1993 and 2009.14Venter O. Sanderson E.W. Magrach A. Allan J.R. Beher J. Jones K.R. Possingham H.P. Laurance W.F. Wood P. Fekete B.M. et al.Sixteen years of change in the global terrestrial human footprint and implications for biodiversity conservation.Nat. Commun. 2016; 7: 12558Crossref PubMed Scopus (550) Google Scholar We then defined several realistic PA strategies where ecoregions are prioritized for protection according to characteristics such as the amount of land already protected and annual rates of conversion (see Table 1 for descriptions). These PA acquisition strategies used single-step myopic algorithms (i.e., with investment decisions made annually) and were tested for a wide range of budgets. We examined a “quick win” strategy whereby ecoregions that are the closest to the protection target (17% or 30% depending on the time step) are prioritized for further gazetting and an opposite “greatest need” strategy targeting ecoregions that are the furthest from the target and in need of the most investment. A realistic approach would be to focus on land that can be cheaply acquired,15Naidoo R. Balmford A. Ferraro P.J. Polasky S. Ricketts T.H. Rouget M. Integrating economic costs into conservation planning.Trends Ecol. Evol. 2006; 21: 681-687Abstract Full Text Full Text PDF PubMed Scopus (717) Google Scholar which we implemented in the “cheap land” strategy by prioritizing ecoregions with the smallest agricultural opportunity cost,16Naidoo R. Iwamura T. Global-scale mapping of economic benefits from agricultural lands: implications for conservation priorities.Biol. Conserv. 2007; 140: 40-49Crossref Scopus (143) Google Scholar or where we can afford the most land for the budget, as explored in the “quick and cheap” strategy. Because of the threat of rapid land conversion, decision makers might choose to focus on the ecoregions experiencing the highest rates of conversion (“most threatened” strategy), or they could focus on ecoregions that are closest to being too converted in terms of area (“last chance” strategy). These myopic strategies were also compared with a “random” strategy whereby ecoregions were selected randomly for land protection.Table 1Summary of the Acquisition Strategies between 2009 and 2030StrategyDescriptionAlgorithm Ranking for Regions below TargetBAUmaintain acquisition and conversion rates observed between 1993 and 2009noneRandom (null model)select ecoregions to protect randomlyrandomQuick winprioritize protection of ecoregions that are the closest to meeting target-level protectionaccording to amount of land needed to reach the target-level protection, from smallest to largestGreatest needprioritize protection of ecoregions that are the furthest from meeting target-level protectionaccording to amount of land needed to reach the target-level protection, from largest to smallestCheap landprioritize protection of ecoregions where buying land is the cheapest (i.e., smallest opportunity cost, estimated as potential revenue per year per hectare for the most profitable crop)according to the median cost of available cells in the ecoregion, from smallest to largestLast chanceprioritize protection of ecoregions that are the closest to being too converted to reach the targetaccording to amount of land left to reach conversion level, from smallest to largestMost threatenedprioritize ecoregions that are being converted the fastest to minimize the loss of area availableaccording to the rate of conversion, from largest to smallestQuick and cheapprioritize ecoregions that are the closest to meeting target-level protection and where buying land is the cheapestaccording to the amount of land needed to reach the target-level protection multiplied by median cost of available land, from smallest to largestBAU stands for “business as usual” and represents a strategy whereby observed acquisition and conversion rates remain the same as those observed between 1993 and 2009. Open table in a new tab BAU stands for “business as usual” and represents a strategy whereby observed acquisition and conversion rates remain the same as those observed between 1993 and 2009. By comparing seven alternative strategies with BAU (Table 1), our goal was to identify which PA strategy would maximize the number of ecoregions with at least 30% of their area protected by 2030. As a proxy for including the cost of buying available land for protection, we calculated the median opportunity cost of a km2 of available area in each ecoregion in 2009.16Naidoo R. Iwamura T. Global-scale mapping of economic benefits from agricultural lands: implications for conservation priorities.Biol. Conserv. 2007; 140: 40-49Crossref Scopus (143) Google Scholar According to this measure, the median annual protection budget (i.e., the budget available for buying land) between 1993 and 2009 was more than US$114 million. In our future projections, we therefore tested a range of annual budgets varying between $1 million and $160 million annually. After removing those with no cost or human footprint data, we were left with 748 ecoregions in our analysis. Between 1993 and 2009, 24.8% (n = 185) of these were not being converted and 15.5% (n = 116) were not being protected. In 2009, just before the 2010–2020 Aichi Targets were established, 247 ecoregions had ≥17% of their area under protection (33.0%) and 226 were ≥83% converted (30.2%); in these latter ecoregions, the 17% target (or anything higher) is unattainable without restoration. In addition, 29 ecoregions (ca. 3.9%) were so heavily converted that they had no available land for protection. The BAU scenario performed well in predicting the number of ecoregions with 17% or more PA in year 2016 (293 ecoregions predicted versus 279 observed with an 85.1% match between the two sets). When projecting the BAU to 2030, we predicted that 190 ecoregions would be at least 30% protected by 2030 but that 321 ecoregions would be too converted to meet this target. This would leave 237 ecoregions without 30% protection by 2030 but with sufficient land to hypothetically reach that target (Figure 1A). The strategy that led the most ecoregions to reach 30% protection by 2030 depended on the annual budget available for PAs. All PA expansion strategies performed better than BAU for the average observed annual budget (∼$114 million) and yielded between 56 and 260 additional adequately protected ecoregions than BAU in 2030 (Figures S1 and S2). Moreover, there was always an acquisition strategy that performed better than BAU in 2030. For the smallest annual budget ($1 million), only the “cheapest” strategy outperformed BAU. However, for as low as $10 million annually (less than 10% of the current budget), the “cheap land,” “quick win,” and “last chance” strategies respectively yielded 137, 71, and 16 additional adequately protected ecoregions in 2030 (Figure 2). With smaller budgets, the “cheap land” strategy performed best for meeting the 2030 target, but with budgets between $85 and $130 million annually, the “quick win” strategy performed best. For the highest budgets, all acquisition strategies performed nearly as well as each other, yielding between 269 and 279 more ecoregions adequately protected than BAU. These results were unaffected by the assumptions made for the rate of conversion of ecoregions (Figure S3). Under the current observed budget, the “quick win” strategy achieved the protection of almost all available ecoregions at the desired level (Figures 1A and 1B). The timing of protection, however, varied between ecoregions (Figure 1C). The rate of accumulation of ecoregions reaching adequate protection over time also varied among strategies and budgets (Figures S1 and S2; Tables S1 and S2). As the target changed from 17% to 30% after 2020 in our simulations, there was a reduction in number of ecoregions meeting the target. For example, BAU yielded 309 ecoregions with a 17% target but yielded 190 in 2030 with a 30% target. However, even in 2020, at least one acquisition scenario always outperformed the BAU strategy regardless of the budget. The “quick and cheap” strategy performed best up to 2020 for the smaller budgets and equally as well as the “quick win” strategy for budgets > $100 million (Table S1). However, when the target increased to 30%, the best acquisition strategy became either the “cheap land” or the “quick win” strategy. For larger budgets, most strategies performed very well. In contrast, the strategy of prioritizing ecoregions that are the furthest from the 17% target (“greatest need” strategy) performed the worst across most budgets up to 2020 (Table S2). Between 2021 and 2030, the worst-performing strategies were mainly “greatest need” and “quick and cheap.” Focusing on ecoregions with the greatest need might seem like a more equitable strategy because those ecoregions that have historically received the least protection are prioritized and, under a different objective, e.g., maximizing the amount of land protected across all ecoregions or achieving equitable representation,17Kuempel C.D. Chauvenet A.L.M. Possingham H.P. Equitable representation of ecoregions is slowly improving despite strategic planning shortfalls.Conserv. Lett. 2016; 9: 422-428Crossref Scopus (17) Google Scholar might perform better. In addition to maximizing the number of ecoregions that are 30% protected in 2030, we measured ecoregion representation by using two quantitative metrics designed to assess equity in representation, namely protection equality18Chauvenet A.L.M. Kuempel C.D. McGowan J. Beger M. Possingham H.P. Methods for calculating Protection Equality for conservation planning.PLoS One. 2017; 12: e0171591Crossref PubMed Scopus (16) Google Scholar and gaps in protection (protection gap),19Sutcliffe P.R. Klein C.J. Pitcher C.R. Possingham H.P. The effectiveness of marine reserve systems constructed using different surrogates of biodiversity.Conserv. Biol. 2015; 29: 657-667Crossref PubMed Scopus (30) Google Scholar over the entire PA network in 2030. The first metric looks at the overall evenness of protection across ecoregions, and the second looks at the average gap between how well an ecoregion is protected and the target (e.g., 17%). We found that all acquisition strategies contributed to improving ecoregion representation in the global PA network between 2009 and 2030 (Figure 3), although at different rates. Protection equality increased the most and protection gap decreased the most under the “cheap land” scenario up to a budget of $85 million, followed by “quick win” for larger budgets. These results support our findings that “cheap land” and “quick win” strategies perform the best and that the latter is best for the current protection budget. One caveat to this analysis is that we did not investigate the spatial configuration of the land available for protection with regard to fragmentation (i.e., how big would the resulting PAs be?) or the biodiversity present within a PA (i.e., are endemic species protected when we meet the ecological representation target?). Our aim was not to identify specific areas for protection but rather to show that there are quick and cost-effective ways to meet the ecological representation components of the CBD Aichi Target 11 as well as future targets. Detailed spatial plans are required for priority regions in order to maximize the gains within these regions. A second caveat was that we did not consider the amount of human pressure within individual PAs. Jones and colleagues found that up to one-third of the global PA estate is converted.10Jones K.R. Venter O. Fuller R.A. Allan J.R. Maxwell S.L. Negret P.J. Watson J.E.M. One-third of global protected land is under intense human pressure.Science. 2018; 360: 788-791Crossref PubMed Scopus (245) Google Scholar We followed a similar analysis that assessed PA coverage and habitat loss13Mappin B. Chauvenet A.L.M. Adams V.M. Di Marco M. Beyer H.L. Venter O. Halpern B.S. Possingham H.P. Watson J.E.M. Restoration priorities to achieve the global protected area target.Conserv. Lett. 2019; 12: e12646Crossref Scopus (23) Google Scholar,20Watson J.E.M. Jones K.R. Fuller R.A. Di Marco M. Segan D.B. Butchart S.H.M. Allan J.R. McDonald-Madden E. Venter O. Persistent disparities between recent rates of habitat conversion and protection and implications for future global conservation targets.Conserv. Lett. 2016; 9: 413-421Crossref Scopus (73) Google Scholar and considered that it would be impossible to assess the state of the PA based on the degree of human modification. This is because some PAs could have been designated in poor condition but could be rapidly improving through on-ground management, and we would misclassify these areas. Given that even converted PAs are nationally designated as “protected,” they are contributing the goals of a PA because they should (in theory at least) be stopping threats from increasing and ensuring that restoration occurs. It would be impossible to tease out which PAs are indeed actually protected and which are only “paper parks.” In addition, because of data restrictions for the human footprint (only available to 2009), we chose to use only PA data to 2009 (instead of 2019) and base our comparison with BAU on modeled data rather than partially available data from 2009 onward. This could have introduced additional uncertainty into our predictions. However, the performance of protection models (r2 ranging between 0.38 and 0.99 with an average of 0.79) and of the resulting BAU strategy in 2016 (85.1% match in ecoregions predicted and observed to be 17% protected) indicated that our model performed well. Finally, the dataset used for calculating total area protected in each ecoregion and deriving the annual rate of protection does not take into account areas that might have been degazetted between 1993 and 2009 (PA downgrading, downsizing, and degazettement [PADDD]).21Mascia M.B. Pailler S. Protected area downgrading, downsizing, and degazettement (PADDD) and its conservation implications.Conserv. Lett. 2011; 4: 9-20Crossref Scopus (246) Google Scholar As a result, the rate of protection used in our analysis could have been inflated for some ecoregions. Because PADDD events are mostly rare and infrequent, we believe that the modeled rates of protection would not have been significantly affected by this omission. Questions about co-benefits (e.g., species representation or carbon storage) and optimal spatial configuration could be investigated at the implementation stage at a national scale. For example, spatial conservation planning tools, such as Marxan22Ball I.R. Possingham H.P. Watts M. Marxan and relatives: software for spatial conservation prioritisation.in: Wilson K.A. Moilanen A. Possingham H. Spatial Conservation Prioritisation: Quantitative Methods and Computational Tools. Oxford University Press, 2009: 185-195Google Scholar or Zonation,23Moilanen A. Kujala H. Leathwick J.R. The Zonation framework and software for conservation prioritization.in: Wilson K.A. Moilanen A. Possingham H. Spatial Conservation Prioritisation: Quantitative Methods and Computational Tools. Oxford University Press, 2009: 196-210Google Scholar can help identify suitable areas that meet multiple requirements including ecological representation and other biodiversity targets within priority ecoregions. To achieve the goal of a globally ecologically representative PA estate as outlined in the 2020 CBD Aichi Targets and beyond, a global strategy is needed for the protection of broad ecosystem types. The prioritization presented here implies a “global” top-down approach to planning future protection most relevant to organizations that work across many countries. We acknowledge, however, the difficulty associated with this approach given that conservation funding and decision making are usually at the national scale24Waldron A. Mooers A.O. Miller D.C. Nibbelink N. Redding D. Kuhn T.S. Roberts J.T. Gittleman J.L. Targeting global conservation funding to limit immediate biodiversity declines.Proc. Natl. Acad. Sci. U S A. 2013; 110: 12144-12148Crossref PubMed Scopus (274) Google Scholar and ecoregions are often shared between countries and continents.25Dinerstein E. Olson D. Joshi A. Vynne C. Burgess N.D. Wikramanayake E. Hahn N. Palminteri S. Hedao P. Noss R. et al.An ecoregion-based approach to protecting half the terrestrial realm.Bioscience. 2017; 67: 534-545Crossref PubMed Scopus (415) Google Scholar,26Olson D.M. Dinerstein E. Wikramanayake E.D. Burgess N.D. Powell G.V.N. Underwood E.C. D'Amico J.A. Itoua I. Strand H.E. Morrison J.C. Terrestrial Ecoregions of the World: a New Map of Life on Earth A new global map of terrestrial ecoregions provides an innovative tool for conserving biodiversity.BioScience. 2001; 51: 933-938Crossref Scopus (4486) Google Scholar However, international policy is a powerful platform from which to guide all levels of decision making, and this analysis shows that signatory nations to the CBD would be more effective in achieving positive biodiversity outcomes by embracing a prioritization schedule. This type of cooperation between nations was core in the Rio Principles,27United NationsRio Declaration on Environment and Development.1992https://www.un.org/en/development/desa/population/migration/generalassembly/docs/globalcompact/A_CONF.151_26_Vol.I_Declaration.pdfGoogle Scholar which laid the foundation for the CBD.