Modelling coral reef biodiversity for prioritizing marine protected area in Madagascar and adjacent islands 

Modelling coral reef biodiversity for prioritizing marine protected area in Madagascar and adjacent islands | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Modelling coral reef biodiversity for prioritizing marine protected area in Madagascar and adjacent islands TIm McClanahan This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7932631/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Purpose To determine the distribution of biodiversity in coral reefs of Madagascar, Comoros, and Mayotte. Methods A Gradient Boosting Model (GBM) evaluating seventy environmental spatial databases predicted reef biodiversity field data to create spatial predictions in 2854 6-km 2 mapped reef cells in 2020 and 2050. Predicted biodiversity were compared to past provincial protected area prioritization activities and the current listing of marine national parks (MNP), Locally Managed Marine Areas (LMMAs), and the World Protected Area Database (WDPA). Results Twenty-one national high biodiversity priority cells were selected for Madagascar, 3 for Mayotte, and 8 for Comoros. Sixteen of the 32 selected high biodiversity locations were contained in 44 of the 102 possible listed WDPA protected areas. The east and coastal reefs south of Antongil Bay and offshore coral reefs islands were notably excluded from national but not LMMA designations. Madagascar’s west coast was better represented than the east coast in WDPA locations. Based on surface temperate predictions, coral cover declined in 55% and gained in 7%, while numbers of taxa declined in 72% but gained in 14% of the grid cells between 2020 and 2050. Conclusions Spatial cells with minor climate-induced changes or gains in coral reef cover and diversity attributes were broadly scattered among governance authorities. However, most locations with little climate change effects were in southwest Madagascar where overfishing is likely to undermine their climate refugia potential. Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction Expansion of protected areas requires making decisions focused on factors of which biodiversity and adaptation to climate change are centrally important (Wilson et al. 2020 ). However, many protected area prioritization studies are based on limited data and have failed to objectively articulate appropriate spatial and temporal scales in the patterns of biodiversity. This data limitation problem is particularly acute for the increasing numbers of local and small-scale management and protected area activities (McClanahan 2023 ). A common limitation of protected area prioritization is the limited number of observations and sampling relative to the number of taxa and the area of coverage. For example, large-scale taxonomic diversity maps are often biased by the patchy and haphazard nature of species occurrence records (Kusumoto et al. 2020 ). Thus, biodiversity distributions over large areas of poorly studied coasts are frequently based on extrapolations from presence/absence data in sparsely sampled cell grids (Jenkins and van Houtan 2016 ). To a better understanding the distribution of numbers of reef species along a poorly sampling coastline of Madagascar and its smaller island neighbors of Comoros and Mayotte, a biodiversity and climate change modelling approach was developed here. The purpose being to model coral reef cover and biodiversity in poorly known locations and thereby provide a gap analysis of biodiversity and marine protected areas for 2020. Madagascar and associated islands Madagascar and the surrounding nations of Comoros and Mayotte are islands known for their unique biodiversity (Cook et al. 2003; Parravicini et al. 2011). Yet, knowledge of marine biodiversity is fragmentary due to lack of widespread data collection throughout the region. Additionally, Madagascar is isolated geographically but at the center of ocean current connectedness (Crochelet et al. 2016 ; Maina et al. 2020 ). Biodiversity investigations of colonial French investigators undertaken between 1961 and 1972 were limited to the southwest regions around Toliara. These investigators focused on taxonomy and identified many marine taxa (> 6000) in the Grand Recif of Toliara (Cooke et al. 2003 ). Subsequent comparisons among Madagascar’s regions have been limited to three areas, namely the reefs near Toliara, the islands around Nosy Be, and the Masoala Peninsula (McClanahan et al. 2009 ; McClanahan and Jadot 2017 ; Samoilys et al. 2022 ; Randrianarivo et, al. 2022, 2023, 2024). These studies find the marine ecosystems are highly affected by human use, but observations are also often limited to accessible areas where human impacts are expected (McClanahan and Jadot 2017 ; Cowburn et al. 2018 ; Gough et al. 2020 ; Randrianarivo et al. 2022 ). Additionally, there is evidence for some spatial homogeneity of the reef communities that may result from the island and isolated nature compared to the more extensive African coastline (McClanahan and Jadot 2017 ; Samoilys et al. 2022 ; McClanahan et al. 2024 ). Thus, there is much to learn about Madagascar’s broader and finer spatial scale biodiversity patterns. Despite the poor knowledge, marine conservation planning and implementation is now rapidly underway with the increasing commitment to local scale community conservation (Gardner et al. 2018 ). Coral reefs contain the highest shallowest water diversity, so coral reef diversity should be a good proxy for overall numbers of species. Mapping of coral reefs showed that most of the ~ 14,000 km 2 of Madagascar’s reefs are found on the western and northern sides of the island (Burke et al. 2012 ). Madagascar has two ecoregions, namely the large Western and Northern Madagascar Ecoregion that contain the two nations of Comoros and Mayotte. The third, or Southeast Madagascar Ecoregions contains fewer coral reefs. For perspective, Madagascar’s coral reefs are ~ 20% of the total reef area in the western Indian Ocean marine province (McClanahan et al. 2024 ). Madagascar faces climate change predictions of rising temperatures, changing rainfall patterns, and more intense cyclones (UNEP Interactive Country Fiches; dicf.unep.ch). The south is expected to see the most significant warming, while the north may experience a decrease in rainfall. Climate changes are projected to increase droughts and impact agriculture and potentially increase reliance on marine resources. Coastal areas will experience sea-level rise and coastal erosion. Site specific studies of coral reefs show considerable variability arising from differences in temperature variability and fishing pressure (McClanahan et al. 2009 ; Randrianarivo et al. 2022 , 2024 ). The southwest, for example, has lower radiation, cooler temperatures, and rainfall than the north. Therefore, thermal stress characteristics and watershed runoff are reduced (Sully et al. 2022 ), but the southwest is also associated with cyclones and greater human dependence on marine resources (Bruggemann et al. 2012 ). The high thermal stress and rainfall in the north may increase watershed runoff but the land also produces more food, which can reduce human dependency and pressure on northern fisheries (Cinner et al. 2016 ; McClanahan and Jadot 2017 ). Additionally, warm water conditions can also acclimate corals to ongoing and future warming (McClanahan et al. 2011 ). Given the coastal variability in natural ecological processes and human dependence, the question remains how coral cover and species will be affected and distributed in the face of increasing climate change and human pressures. Past protected area prioritization, planning, and actions Past biodiversity conservation planning reports have identified four areas in Madagascar, namely the Masoala Peninsula, the Grand Recif of Toliara, the northwest regions of Antsiranana, and the extreme southern tip of the island. These selections were based on knowledge of the existence of whales, high coral diversity, and remote and potentially unique but unstudied fauna (Obura et al. 2012 ; van det Elst and Everett 2015 ). Yet, in most cases plans may reflect a bias dependent on the long history of study of marine faunal taxonomic work in Toliara, observations of whale migrations, bird nesting, and concerns to protect remote areas with few people. Therefore, how past priority and existing protected area selections will differ from environment-taxa modelling predictive approaches requires investigation. Conservation actions have developed rapidly during the past 20 years in these three countries (Gardner et al. 2018 ; UNEP-Nairobi Convention and WIOMSA 2021). Much of area designation has occurred since the 2003 IUCN World Parks Congress in Durban where the Malagasy government committed to increase the nation’s protected areas (Fig. 1 a). For example, since 2003, 20 national MPAs have been created along with > 200 Locally Managed Marine Areas (LMMAs) (UNEP-Nairobi Convention and WIOMSA 2021). These areas are a mixture of state, co-managed, and private control. Many areas rely on local management known as dina where local committees can set rules and manage resources in collaboration with other authorities (Parker et al. 2024 ; Zafimahatradraibe et al. 2024). In many cases, the biodiversity of these locations has not been reported or monitored. Mayotte and Comoros are under the authority of different national governments since 1974, and Mayotte became a French territory in March 2011. Comoros has one established Marine Park or Moheli gazetted in 2015 with strong island government community support (Fig. 1 b) (UNEP-Nairobi Convention and WIOMSA 2021). Three additional parks are in the process of being established, namely the Coelacanth MNP, the Mitsamiouli-Ndroudé MNP, and the Shisiwani MNP. The entire lagoon of Mayotte and its EEZ were classified by the French government in 2010 as a MNP covering 68,800 km 2 but only 1% was classified as highly protected. Nevertheless, it is one of the largest marine protected areas in the Indian Ocean. The 2021 UNEP report also identifies the small (0.6 km 2 ) Nature Reserve of M’bouzi established in 2007 prior to the larger designation. In contrast, the WDPA identifies 16 protected areas of which some are small coastal stretches and small offshore islands that may not manage resources beyond the shoreline. Methods Overview Several scientific modelling and software advances have made it increasingly possible to predict finer scale patterns and thereby map marine biodiversity based on environmental proxies (Pilowsky et al. 2022 ). These include: 1) moderate resolution mapping of habitats, such as coral reefs, at large scales, 2) large-scale underwater data collection and collaboration on important biodiversity proxy metrics (i.e. coral cover and taxa composition), 3) global satellite coverage of environmental variables that are proxies for influences on biodiversity, 4) statistical machine learning algorithms that can handle large amounts of complex data to make predictive models, and 5) sea surface water temperature predictions that are available from IPCC data time series ( https://www.ipcc.ch/data/ ). The convergence of these tools provides the possibility of developing models to predict biodiversity on modest scales and to test different predictions for different scenarios. Environmental information at scales of satellite and shipboard observations can potentially reduce or correct biases created by incomplete, sparse, and anecdotal information. The Ethics and Consent to Participate declarations are not applicable to this study. The research described below was undertaken to better understand the finer-scale variability in taxonomic richness in the coral reefs of Madagascar, Comoros, and Mayotte. Specifically, coral and fish richness, and a proxy of total species were mapped on the 6 km 2 scale where most cells lacked field census data. A provincial model that predicted the number of species based on ~ 2000 coral and fish census used empirical relationships with many environmental, demographic, and management variables (Fig. 2 ) (McClanahan et al. 2024 ). This was carried out with the machine learning algorithm described below. Empirical relationships with the environmental variables derived from the algorithms were used to predict numbers of taxa in all 3361 6-km 2 mapped coral reef cells of the study region. Because the African continent has higher biodiversity than Madagascar (McClanahan 2019 ; Samoilys et al. 2022 ), faunal province will differ from national biodiversity priorities. Therefore, to find priorities specific to Madagascar, the two ecoregions and 3 governance authorities were evaluated here separate from the western Indian Ocean province results. Additionally, temperature variables available from the IPCC temperature time series (RCP2.6 and RCP8.5) were used to predict coral cover and number of coral taxa in 2050 based on 2020 temperature-taxa associations. Biodiversity mapping framework The United Nations Environmental Program (UNEP) and International Union for the Conservation of Nature (IUCN) have compiled protected areas in various stages of planning and establishment. The World Database of Protected Area (WDPA) is available from the World Conservation Monitoring Center (WCMC) (protectedplanet.net). Shape files were downloaded and overlayed with the modelled high biodiversity priority areas. This allowed us to determine the current designations that corresponded to the modeled priorities. This database lists 60 protected areas for Madagascar, 26 for Mayotte, and 16 for Comoros. WDPA numbers are larger than the recent provincial compilation of the UNEP-Nairobi Convention report, as different criteria were used for inclusion. Many protected areas have been designated but the authority to manage them and their success is often unreported. Many may also be in a proposal or planning process and not the post-implementation stage. Environmental layers data Environmental data compilations used several sources, which resulted in 70 spatially complete variables derived from a combination of satellite and shipboard measurements (see the list of data sources in McClanahan et al. 2024 ). Environmental data were a mixture of oceanographic data, such as photosynthetic active radiation (PAR), pH, calcite, dissolved oxygen, diffusion attenuation, salinity, net ocean primary productivity, chlorophyll-a variables, phytoplankton carbon, and wave height (Tyberghein et al. 2012 ; Yeager et al. 2017 ). Several water temperature or thermal stress metrics were calculated including sea surface temperature (SST) mean, median, range, standard deviation, skewness, kurtosis, rate of rise, and cumulative excess heat or degree-heating weeks (DHW). Several composite thermal and water quality stress metrics were included, such as the Global Stress Model, an indicator of thermal inputs (Maina et al. 2011 ), and a composite nutrient concentrations model (Andrello et al. 2022 ). Finally, estimates of reef connectivity calculations were used to estimate potential larval flow including measures of connectivity, net flow, indegree, outdegree, and retention for each cell in this region (Fontoura et al. 2022 ). Geographic variables included wilderness (> 4 hours travel time from human population), travel distance to people, shore, and ports, and market gravity or the number of people living on the shore or cities as divided by the square of the distance or travel time (Maire et al. 2016 ). Cells were assigned to four fisheries management categories including unrestricted fishing, restricted fishing, low compliance closures, and high compliance closures. These classifications were based on information in published literature, the experience of the observers, and discussions with knowledgeable observers (McClanahan et al. 2015 ). Fish census observers also recorded the depth and habitats of the sites as reef edge, reef crest, reef flat, or reef lagoon. Detailed methods and model results have been presented elsewhere and here the focus is on predictions of biodiversity and management implications for Madagascar (McClanahan et al. 2024 ). Field data collection The model uses the above environmental data to make predictions of coral taxa for 2020 based on environment empirical field data associations revealed by Gradient Boosting Model (GBM) or specifically the Boosted Regression Tree (BRT) software. Field data were collected in all three countries from several field trips. The Western Indian Data were collected between 1995 and 2020 and environmental data prior to the sampling was used in the BRT modelling process. The proxy for total number of coral reef taxa used a combination of coral and fish taxa sampled by the following methods. Coral sampling Corals were visually sampled in haphazardly placed quadrats of ~ 2 m 2 where all corals > 5-cm were identified and counted in ~ 15–20 replicates (McClanahan et al. 2007 ). Thus, the values used here were the total number of taxa in ~ 40 m 2 . Taxa identification was to the genus level, but Porites colonies were identified further as massive, branching, or Porites rus and Galaxea as either G. astreata or G. fascicularis . 1001 well distributed sites were sampled in the region (Fig. 2 ). The two observers with the most samples (N.A. Muthiga and T. McClanahan) were compared and found to have no significant differences (McClanahan et al. 2024 ). Fish sampling Two experienced observers (T. McClanahan and J. Wickel) counted fish in designated areas or belt transects of 500-m 2 (McClanahan 1994). Replicates undertaken close to each other were pooled or averaged dependent on the methods such that the final units were number of species per ~ 500 m 2 . The number of species for the 6 selected families known to be good proxies for total fish diversity (Acanthuridae, Chaetodontidae, Labridae, Pomacanthidae, Pomacentridae, and Scaridae) were extracted (Allen and Werner 2002 ). Each observer also estimated biomass in their transects as the sum weights of the individual species or families based on length estimates and known length-weight relationships. A total of 1201 transects were sampled throughout most of the nations and two ecoregions of the Madagascar (Fig. 2 ). Model spatial and temporal predictions Statistical machine learning algorithms or GBM are increasingly being used to make predictions with large and complex data sets. Specifically, the BRT algorithm was the specific GBM used here. BRT is a commonly used algorithm for evaluating complex environmental-ecological data and shown to be a top performer among machine learning options (Elith et al., 2008 ; Kuhn and Johnson 2013 ). Prior analyses indicate that BRT models are preferred because they are effective at handling nonlinear relationships, missing values in covariates, interactions between predictors, and have a high predictive performance (Kuhn and Johnson 2013 ). Once the taxa-environment relationships were established from the empirical census data, the model was used to predict biodiversity using the environmental data in all 3361 mapped cells for 2020 and for the projected 2050 temperature conditions. Predictions included the number of fish and coral taxa for the sampled areas, but the normalized average was used as a proxy for the total number of taxa in a cell and therefore the spatial cell’s proxy for biodiversity. Model predictions were tested for efficacy using a 70 − 30% training and testing procedure to determine the fits, which were R 2 of ~ 80% for the full data and ~ 45% for the training and testing data. Models require keeping some variables constant between spatial cells to make comparable predictions. For example, numbers of fish species are strongly correlated with fish biomass and change with water depth. Therefore, to make fair between-cell comparisons, biomass was held constant at 600 kg/ha and depth at 10 meters (McClanahan et al. 2024 ). Therefore, the maps are based on partial effects where local depth and fish biomass were held constant and the predictions are for these constants but where other environmental variables are from the various databases. The future state model selected those temperature variables common to the above model and RCP8.5 Business-as-Usual and RCP2.6 or the carbon emission reduction scenarios. These variables were used to make biodiversity predictions for the current scenario values in 2020 and future predictions in 2050. Specifically, the first BRT selection process selected 6 variables shared IPCC temperature data. These variables were the mean SST (CMIP does not give median SSTs), skewness, kurtosis, bimodality, and cumulative excess heat (degree-heating weeks = DHW). Therefore, no future predicted human demographic, or environmental variables were included in future forecasts, so the predictions are largely based on future temperatures and 2020 associations with other environmental and demographic data. Coral cover was taken from a previously machine learning model for 2020 and 2050 (McClanahan and Azali 2021 ). Data used in scatterplots were the changes in coral cover and number of taxa for all spatial cells for the years 2020 and 2050. Scatterplots present the differences in the values between the years (2050–2020) as a measure of change or resilience for the 3361 mapped cells. It should be appreciated that the sea surface temperature predictions for 2050 have reached the + 1.5 0 C above baseline by 2025, or 25 years earlier than predicted by the IPCC data time series (Hansen et al. 2023 ). Results Biodiversity maps for coral, fish, and the total diversity proxy are presented in the following figures (Fig. 3 ). Predictions for coral Model predictions for numbers of coral taxa indicate both broad and fine-scale patterns (Fig. 3 a). Predicted numbers of coral taxa in Madagascar were overall highest in the north and northwest of the Antsiranana and Mahajanga Provinces. This high diversity extends west to include Mayotte and Comoros. The Ankarea MNP (Mitsio Islands) was in the center of this biodiversity location. Diversity was predicted to be highest in the western or leeward side of Madagascar in the northern Antsiranana Province. Numbers of coral taxa was predicted to decline to the south into Mahajanga and Toliary Provinces. The border region between Antsiranana and Mahajanga Provinces in the Ambanja District had a cluster of lower diversity, but numbers increased to the south until midway down the Mahajanga province. The Ankivonjy MNP is located just north of the high and low predicted transition of coral diversity or just north of the Mahajanga provincial border. The Toliara Province in the southwest was predicted to have lower numbers of corals overall but had a high diversity location south of Velondriake and near the Soarkiake MNPs. Nevertheless, some high diversity reefs may not be contained in these parks. In eastern Madagascar and Antsiranana, the Masoala MNP in the Antalaha District was predicted to have lower coral diversity than the Ambodivahibe MNP in the northeast. Further south in the eastern Tomasina Province, high diversity was predicted from the northern border south to St. Anne Island. High diversity was predicted for the Nosy-Boraha and Soanierana Ivongo Districts. The southern part of Toliara and all Fianarantsoa Provinces were predicted to have fewer reefs and a low diversity of corals. Predictions for Comoros and Mayotte indicated high numbers of coral taxa throughout these islands. Calculations of the predictions of the average numbers of taxa suggest Mayotte had slightly higher numbers of coral taxa (25.5 ± 2.0 (SD)) taxa per 40 m 2 than the Comoros Islands (23.4 ± 1.7). Predictions for fish Predictions of numbers of fishes in Madagascar indicate some differences compared to coral diversity predictions (Fig. 3 b). The number of fish was highest in Antsiranana Province but with some clear onshore-offshore patterns or higher diversity in offshore reefs. Again, the offshore Ankarea MNP was the predicted center of this fish diversity. Fish diversity was high and less variable than corals in Ambanja District but declined further to the south and only high again in the Barren Islands of Mahajanga Province. Fish diversity in the northeast was predicted to be higher than in corals. Locations of high diversity were predicted for the entire Masoala Peninsula in the Antalaha District. Diversity predictions declined in the Antoginil Bay of the Maroantsetra District, due to poor water quality conditions. However, south of Nosy-Boraha and Soanierana Ivongo there was also a high predicted fish diversity that extended south to the Toamasina District. The southwest Toliara Province had low predicted fish diversity overall but had high predicted diversity in the north of the Province around the Velondriake and near the Soarkiake MNPs. The province had the same onshore-offshore diversity patterns observed elsewhere. The southern part of Toliara and all of the Fianarantsoa Provinces were predicted to have few reefs and a low diversity of fish except for an offshore island in Fianarantsoa Province. Predictions for numbers of fish species in Comoros and Mayotte indicated more spatial variability than corals, particularly for the onshore-offshore gradients. For example, high numbers of species were predicted on the outer reef of Mayotte, but numbers declined rapidly shoreward. Predictions for the west and south sides of Mayotte were to have more frequent species than the island's east and north sides. In Comoros, Mwali (Moheli) was predicted to have the most fish species, followed by Ngazida (Grande Comoros) and high spatial variability in species for Ndzuwani (Anjouan) island. The proposed Mitsamiouli Ndroudeb MNP in the north of Ngazida island was predicted to have high numbers of fish species. The proposed Coelacanth MNP on the southern tip of Ngazida has reefs only on the far southern end, which was also a location predicted to have high numbers of fish. Predictions for a mixture of moderate and high numbers of fish was made for locations within the proposed Shisiwani MNP located on the northwest of the Ndzuwani island. Destructive fishing is reported to be common in Ngazida and Ndzuwani and actual are expected to be lower than predicted numbers of species. Predictions for total numbers of species Predictions using the biodiversity proxy for the total number of marine taxa indicate similar patterns with both coral and fish (Fig. 3 c). Notably, there was high predicted total diversity from Ankarea west to Mayotte and Comoros. Other patterns align with high diversity predictions offshore in the northwest. The nearshore locations around Masoala and southern reefs extending into the northern Toamasina Districts were predicted to have high total diversity. Similarly, high diversity was predicted for the Velondriake and Soarkiake MNPs in northern Toliara and Morombe Districts. Comoros and Mayotte reefs were predicted to have high numbers of marine taxa. Mwali had the highest number of taxa followed by Ngazida. Predicted taxa in Mayotte were high on the outer reef and declined shoreward. Predictions for Ndzuwani were similar but patterns were more variable along the coast. Comparisons with past reports and protected area locations Summarizing the findings of the past reports and the number of criteria used in the biodiversity evaluations (3 of taxa and 4 of spatial scales) indicated that there were no locations in Madagascar that fit all 12 WIO provincial criteria (Table 1 ; Fig. 1 ). Therefore, by the model criteria, all selected areas were national and not WIO provincial priorities. Two past reports identified Grand Recif and Nosy Ve in Toliara Province and Helodrano in Antoginal Bay in Tomasina County. None of these were among the model’s high biodiversity selections. A large area that included Masoala was identified by the World Heritage report, and the windward or eastern ocean-exposed peninsula was identified by the model’s criteria. Masoala peninsula ranked high for the fish and total diversity metrics, but not by the coral criteria. The model and criteria selected new sites in the coastal reefs south of Antongil Bay or Ivontaka-Antanambe villages and those south to St Anne Island in Tomasina. Table 1 Selection of priority diversity locations selected by the biodiversity machine learning model predictions for Madagascar, Mayotte, and Comoros. Marine protected areas listed were determined from the joint project of IUCN and UNEPs – World Conservation Monitoring Center (WCMC) compilation or World Database of Protected Area (WDPA). Criteria are based on a maximum of 12 possible criteria described in McClanahan et al. ( 2024 a) and as the number of past reports where they are listed as priorities. The 12 criteria are top selections for 3 measures of biodiversity (coral, fish, and their combination) and four spatial scales (nation, ecoregion, province, and reef clustering). Ecoregion name Country Name World Heritage Marine Sites of Outstanding value SWIOFP Biodiversity hotspots Sum Criteria (McClanahan et ala. 2024a) Sum Criteria (Regional Reports) Marine Protected area Western and Northern Madagascar Madagascar Antisaranana - Masoala Antongil bay, Northeast Madagascar Antongil bay 2 2 Masoala Western and Northern Madagascar Madagascar Nosy Mitsio North and Northeast Madagascar, Ambodivahibe - Sahamalaza 8 1 Ankarea Western and Northern Madagascar Madagascar Baie Lotsaina - Ankazomalemy North and Northeast Madagascar, Ambodivahibe - Sahamalaza 6 1 Nosy Hara Western and Northern Madagascar Madagascar Nosy Sakatia/Nosy Be North and Northeast Madagascar, Ambodivahibe - Sahamalaza 6 1 Lokobe Western and Northern Madagascar Madagascar Ankazoberavina - Baie Androfiabe North and Northeast Madagascar, Ambodivahibe - Sahamalaza 6 1 Nosy Antsoha, Ampasindava, Ankivonjy Western and Northern Madagascar Madagascar Nosy Faly North and Northeast Madagascar, Ambodivahibe - Sahamalaza 2 1 Western and Northern Madagascar Madagascar Nosy Ambariovato North and Northeast Madagascar, Ambodivahibe - Sahamalaza 1 1 Southeast Madagascar Madagascar Southeast Madagascar Southern Madagascar (the deep south) 0 1 Western and Northern Madagascar Madagascar Anjiambe - Ampisikanana 6 0 Ambodivahibe, Analamerana, Loky manambato, Oronjia, Western and Northern Madagascar Madagascar Ambohotrabo - Tsiadamaba 4 0 Mangoky Ihotry wetland complex, Velondriake, Manjaboaka, Soriake Western and Northern Madagascar Madagascar Vohemar bay 4 0 Western and Northern Madagascar Madagascar Nosy Boraha - Toamasina 3 0 Southeast Madagascar Madagascar Nosy Faho 3 0 Western and Northern Madagascar Madagascar Maintirano 2 0 Western and Northern Madagascar Madagascar Ambatonjanahary - Anjiabe 2 0 Western and Northern Madagascar Madagascar Ivontaka - Anatanambe 2 0 Mananara nord, Seranambe, Vohitralanana, Ambodimangamaro Western and Northern Madagascar Madagascar Analalava - Mahabo 1 0 Southeast Madagascar Madagascar Nosy Dombala - Nosy Fonga 1 0 Southeast Madagascar Madagascar Ampanotoamaizina 1 0 Southeast Madagascar Madagascar Vohimasina 1 0 Southeast Madagascar Madagascar Manakara 1 0 Southeast Madagascar Madagascar Grande recif Grande Recif (11) 0 1 Southeast Madagascar Madagascar Nosy Ve Nosy Ve 0 1 Western and Northern Madagascar Mayotte Longoni bay - Tsingoni Comoros - Glorieuses crescent 8 1 Ilots M'Tzamboro, Mayotte, Pointes Et Ilots Du Nord, Baie De Dzoumogne-Longoni Western and Northern Madagascar Mayotte Tsingoni - Baie de Kani Comoros - Glorieuses crescent 7 1 Mayotte, Littoral De Sada-Chiconi, Littoral De Kani-Keli, N'Gouja Western and Northern Madagascar Mayotte Pamanzi bay - Baie de Kani Comoros - Glorieuses crescent 4 1 Ilots De La Passe, Ilots De Dembeni, Ilots De Bandrele, Mayotte, La Vasiere des Badamiers, Littoral De Mamoudzou, Littoral De Bandrele, Ilots Mbouzi, Pointes Et Plages De Saziley Et Charifou, Cratere De Petite Terre, Littoral De Dembeni, Vasiere Des Badamiers Western and Northern Madagascar Comoros, Mwali Island Fomboni Comoros - Glorieuses crescent Moheli (1) 10 2 Moheli - Zone de transition, Parc National de Moheli Western and Northern Madagascar Comoros, Mwali Island Nioumachoa Comoros - Glorieuses crescent Moheli (1) 10 2 Moheli - Reserve Mea, Moheli - Reserve Magnougni et dzaha 1, Moheli - Reserve Magnougni et dzaha 2, Moheli – Reserve Nioumachioua, Moheli - Reserve Ouenefou, Parc National de Moheli Western and Northern Madagascar Comoros, Ngazida Island Moroni- Singani Comoros - Glorieuses crescent 12 1 Parc National Coelacanthe Western and Northern Madagascar Comoros, Ngazida Island Niamaoui - Chomoni Comoros - Glorieuses crescent 12 1 Parc National Mitsamiouli Ndroude Western and Northern Madagascar Comoros, Ngazida Island Dimani - Mohoro Comoros - Glorieuses crescent 10 1 Western and Northern Madagascar Comoros, Ndzouani Island Djomani - Domoni Comoros - Glorieuses crescent 6 1 Western and Northern Madagascar Comoros, Ndzouani Island Anjouan SW Comoros - Glorieuses crescent 4 1 Parc National Shisiwani Western and Northern Madagascar Comoros, Ndzouani Island Anjouan N, NE Comoros - Glorieuses crescent 2 1 Of these 60 WDPA in Madagascar, 8 specific locations overlapped with the model’s 21 larger biodiversity hotspot locations. These WDPA specific names of the protected area are Masoala, Ankarea, Nosy Hara, Lokobe, Nosy Antshoa, Ampasindaya, Ankivonjy, Ambodivahibe, Analmerana, Lokymanambato, Oronjia, Mangoky, Ihbotry wetland complex, Velondriake, Manjaboaka, Soriake, Mananara nord, Seranambe, Vohitralanana, and Ambodimangamoro. Designations are complicated in that several smaller protected areas are contained in larger protected areas. For example, all WDPA protected areas in Mayotte are within the larger national protected area. Therefore, the 3 hotspot locations identified from the model, namely Longoni bay – Tsingoni, Tsingoni - Baie de Kani, and Pamanzi bay - Baie de Kani, are all contained in either the larger park or smaller protected areas. In Comoros, most of the coastal areas of the island of Mwali are in the Moheli park but this park contains several reserves. Four of these reserves are contained in the modelled biodiversity hotspot location named Nioumachoa. On Ngazida Island, the Coelacanth and Mitsamiouli Ndroude national parks are contained in the selected Moroni-Singani and Niamaoui = Chimoni hotspots, respectively. On the island of Ndzouani, the Shisiwani MNP is contained in the Anjouan southwest selected hotspot. Therefore, 3 of the 8 selected hotspots are not included among WDPA park designations. These are the Dimaini-Mohoro on Ngazida Island and the Diomani-Domoni and Ajouan northeast hotspots on Ndzouani. Therefore, of the 102 WDPA protected areas in these 3 countries, 44 were overlapping with half (16) of the models selected 32 biodiversity hotspots. In northern Antsiranana Province there was a large area selected by the criteria and the World Heritage report. Many of these reefs were included in the marine parks of Ambodivahibe, Nosy Hara, Ankarea, and Ankivonjy. Specifically, these included the biodiversity selected locations of Vohemar, Anjabe Ampisikanana, Nosy Mitsio, Nosy Sakatia, Nosy Ambariovato, Baie Androfiabe, and Analalava Mahabo. The smaller area of Analalave Mahabo in northern Mahajanga Province was selected for its predicted high numbers of coral taxa. However, numbers of fish species were predicted to be low, and so was the total diversity. Vohemar was selected for high numbers of fish taxa but predicted to have modest numbers of coral taxa. There were also a few offshore islands in the Toliara and southeastern Fianarantsoa Provinces that were selected for their predicted high numbers of fish species. These include the offshore coral reef islands of Nosy Dombala, Nosy Fonga, and Ampanotoamaisina south of Toamasina. Some coastal rocky reefs of Vohimasina and near Manakara in the south were also identified by the same criteria. The western coast included a small coral reef island ~ 15 km offshore from Maintirano contained in a RAMSAR designated area. Further south are the islands and the coastal fringing reefs from Andavakoaka south to Tsifota on the coast from the Ifaty forest. Many of these locations are included in Velondriake and Soarkiake MNPs. Climate change predictions Plots of the model’s predicted changes in coral community cover and number of taxa from 2020 to 2050 indicate widespread losses in cover and taxa (Fig. 4 ; Table 2 ). In the Business-as-Usual scenario (RCP 8.5) average coral cover was predicted to decline in Madagascar from 34.6 ± 11.1% (SD) in 2020 to 24.6 ± 9.9% in 2050. The number of taxa was also predicted to decline from 16.5 ± 1.9 to 15.6 ± 1.5 taxa per 40 m 2 . Therefore, most predictions lay in the bottom left quadrat where there were losses in coral cover and taxa. In fact, 72% of the reefs were predicted to have losses in both cover and taxa, while 7.3% were predicted to have gains in both coral and taxa. Mapped cells with gains were distributed among Mahajanga, Toliary, Toamasina and Antsiranana provinces. Table 2 Results of impacts on coral cover and number of taxa for two climate change scenarios using CMIP5.0 variables or (a) business as usual and (b) carbon emission reductions. Predicted coral cover (%) and number of taxa (per ~ 40m 2 ) in 2020 and 2050. The number of coral reef cells predicted to have the 4 combinations of gains and losses (2050 − 2020) of coral cover and number of taxa. Country 2020 Coral cover, % 2050 Coral cover, % 2020 Number of taxa 2050 Number of taxa Cover loss/ taxa loss, n cells (%) Cover loss/ taxa gain, n cells (% Cover gain/ taxa loss, n cells (%) Cover gain/ taxa gain, n cells (%) Business-as-usual scenario RCP8.5 Madagascar 34.6 (11.1) 24.6 (9.9) 16.5 (1.9) 15.6 (1.5) 1643 (72) 183 (8) 289 (12.7) 167 (7.3) Mayotte 47.1 (3) 44.8 (4.2) 25.5 (2) 23.1 (1.4) 214 (79.6) 0 55 (20.4) 0 Comoros 31.5 (5) 29.5 (5.5) 23.4 (1.7) 22.6 (1.6) 236 (99.2) 2 (0.8) 0 0 Carbon emission reduction scenario RCP2.6 Madagascar 34.6 (11.1) 31.4 (16) 16.5 (1.9) 15.7 (1.6) 1259 (55.2) 48 (2.1) 656 (28.7) 319 (14) Mayotte 47.1 (3) 56.7 (9) 25.5 (2) 24.7 (1.8) 17 (6.3) 0 212 (78.8) 40 (14.9) Comoros 31.5 (5) 20.7 (4.6) 23.4 (1.7) 22.3 (1.9) 226 (95) 12 (5) 0 0 In the carbon emission reduction scenario (RCP2.6) average coral cover in Madagascar was predicted to decline less but more variably from 34.6 ± 11.1% (SD) in 2020 to 31.4 ± 16.0% in 2050. The number of taxa was predicted to be smaller and more variable from 16.5 ± 1.9 to 15.7 ± 1.6 taxa per 40 m 2 . Still, predictions for most reefs fell in the bottom left quadrat but with fewer losses in coral and taxa than the Business-as-Usual scenario. For example, 55.2% of the reefs were predicted to lose both cover and taxa, while 14.0% predicted to experience gains in both cover and taxa. Reef cells are predicted to have gains in cover, but losses in species increased from 12.7% to 28.7% when comparing the two climate change scenarios. The carbon emission scenario RCP2.6 was predicted to have better outcomes for Mayotte, Toliary, and Mahajanga reefs. Mayotte reefs were predicted to have among the highest coral cover in 2020 at 47.1 ± 3.0% and to be reduced to 44.8 ± 4.2% in 2050 in the Business-as-Usual RCP8.5 scenario. Mayotte was also predicted to have a higher number of taxa than Madagascar at 25.5 ± 2.0 taxa per 40 m 2 and predicted to decline to 23.1 ± 1.4 in 2050 if carbon emissions are not reduced. If carbon emissions are reduced, the model predicts an increase in coral cover above the 2020 baseline in 2050 to 56.7 ± 9.0%. This will also produce a smaller loss of coral taxa to 24.7 ± 1.8 per 40 m 2 relative to the 25.5 ± 2.0 2020 baseline. Comoros reefs were predicted to have the lowest 2020 baseline coral cover of 31.5 ± 5.0% and to be reduced to 29.5 ± 5.5% in 2050 in the Business-as-Usual and 20.7 ± 4.6 in the carbon reduction scenario. Thus, the prediction is for a greater decline in cover with reduced emissions, which is the opposite of Mayotte. Comoros has a predicted intermediate number of coral taxa in 2020 at 23.4 ± 1.7 and small predicted to decline to 22.6 ± 1.6 per 40 m 2 in 2050 in the Business-as-Usual scenario. If carbon emissions are reduced, the predicted numbers of taxa in 2050 were unchanged in 2020 at 22.3 ± 1.9 taxa per 40 m 2 . Predictions are based on ocean temperature conditions and did not include other important human influences such as fishing and watershed erosion. Discussion The methods used here were able to map biodiversity on a large scale and identify several new conservation or protected area priorities for Madagascar. Moreover, these were identified by an objective measure based on multiple environmental variables, and at a broader and finer spatial scale than past efforts. Past efforts used presence-absence data (Allnut et al. 2012) or prioritized at the larger WIO faunal province delineation (McClanahan et al. 2024 ). The higher diversity on the African continent makes it important to independently establish national level priorities in Madagascar. Additionally, the 6.25 km 2 scale of analysis provided a much finer scale for decision making that may be more realistic in scaling the costs and efficacy of management. Notably, many recent conservation initiatives are small-scale community managed areas (Gardner et al. 2018 ). The outcomes for the reef coral and fish faunae environmental relationships showed a mixture of shared and divergent environmental influences on the taxa. For example, water temperature variability affected both groups, but fishes were more influenced by connectivity and larval retention, productivity, and fish biomass than corals. In contrast, physicochemical factors, such as waves, currents, dissolved oxygen, salinity, and chronic and acute temperature stresses most affected corals. Past prioritization decisions represent the historical focus on planning large-scale protected areas (Wells et al. 2016 ). This approach has, however, been augmented in Madagascar with the development of smaller- or community-scale proposals that now cover large areas of Madagascar (Rocliffe et al. 2014 ). These areas include many smaller LMMAs associated with many tropical habitats (UNEP-Nairobi Convention and WIOMSA 2021). The outcomes of both the 20 nationally gazetted protected areas and the > 200 LMMA for protecting biodiversity were a primary concern for assessing their relationship to underlying biodiversity patterns and potential for conservation impact. In many cases, high resource dependency and low funding for conservation has often undermined the effectiveness of larger nationally protected areas, where outcomes seldom differ from effective national fisheries restrictions management (McClanahan et al. 2015 ; Randrianarivo et al. 2022 ). Smaller community managed areas may have similar problems associated with their small size relative to the resource and areas needed to support many of the captured animals (McClanahan and Graham 2015 ). Government partnerships with NGOs and the private sector may help to overcome some of these historical problems. Community management has been in place for sufficient time to provoke several lessons (Gardner et al. 2020 ). Common problems include the difficulties of promoting participation and good governance, applying rules, resolving conflicts with outsiders, influencing fish prices such that benefits exceed fishing costs, long-term commitments to monitoring human use and key resources, and sustainable funding (Parker et al. 2024 ; Zafimahatradraibe et al. 2025 ). Some important solutions may include the importance of co-management rather than community management, outside support to negotiate conflicts among neighboring communities, a focus on managing natural resources rather than the more common focus on distributing and selling captured resources, increasing the production of fish to reduce poverty and social disparity, collaborative decision making, diversification and entrepreneurial sourcing of income, and monitoring feedbacks to resource users to improve adaptive management (McClanahan and Abunge 2019 ; Gardner et al. 2020 ). Lessons such as these provide some guidelines to consider among many policy and management options. Nevertheless, much remains to be learned about how to support biodiversity while reducing poverty over the long term. The finer spatial scales of the presented maps compared to past reports may be more useful for emerging smaller-scale governance and LMMA conservation priorities. Moreover, ecological services are among key current priorities and include fisheries production, shoreline protection, and local biodiversity conservation. Therefore, ecological functions provided by fishes and corals align well with the human services prioritization approach that drives LMMA management. Indeed, sustaining high and stable fish catches is a major social concern throughout Madagascar and Comoros (Le Manach et al. 2012 ; Cinner et al. 2016 ; McClanahan and Jadot 2017 ; Gough et al. 2020 ; Zeller et al. 2021 ; Ranaivomanana et al. 2023 ). Fish resources have been reported to be in better condition in Mayotte, associated with greater wealth and lower local natural resource dependency (McClanahan and Jadot 2017 ). Conservation priorities Variability in environmental associations with the two taxa resulted in variability in the selection of locations. For example, some locations were selected because of high predicted numbers of fish, such as those in eastern Madagascar including Anjabe, Ampisikanana, Vohemar, Ivantake, Masoala, Antanambe, Nosy Dombolo, Nosy Fonga, Ampanotoamizina, Vohimasina, and Manakara. Locations in the northwest were picked for high numbers of corals, such as Antalava-Mahabo. Most of the other western sites were selected based on both corals and fish or the proxy for total numbers of species. Locations included many of the areas recently included as national marine parks, specifically Ambodivahibe, Nosy Hara, Ankarea, Ankivonjy, Velondriake, and Soariake. The combined taxa or proxy approach used here provided thorough coverage that should make moderate to good predictions for where high numbers of species are located. Many of these locations are heavily influenced by other human watersheds and fishing impacts (Bruggemann et al. 2012 ; Maina et al. 2013 ). Many existing MNPs appeared to coincide with the high diversity predictions. Therefore, the machine learning model supports national decisions with a few exceptions. Among the exceptions are the Kirindy Mite MNP north or Velondriake MNP, which was predicted by the model to have low but variable diversity among its islands. Some notable omissions of the national park selections would be the reefs south of Antongil Bay or the Ivontaka-Antanambe and St Anne Island villages. These reefs are, however, an area of dense LMMAs that may provide some conservation benefits (UNEP-Nairobi Convention and WIOMSA 2021). Additionally, there were some offshore coral islands in the east and west predicted to have high diversity, namely Nosy Dombala, Nosy Fonga, and Ampanotoamaisina in the islands offshore in the east. In the west coast, islands from Maintirano to Barren Islands and Iles Eparses were also identified by the model and as a RAMSAR designation. Most notable is the high overlap of the model’s biodiversity predictions and MNPs on the west coast of Madagascar. Some of the regions priority areas may have been ignored due to past efforts that focused on the WIO provincial rather than the national scale priorities. Scale of delineation can be quite important in terms of the numbers, areas, and distribution of selected sites (Grace et al. 2022 ). Madagascar does not fare well in provincial delineation selections for several reasons, including poor coverage of species data, the isolated island biogeographic affect, and the dispersed nature of the coral reefs. For example, the coarse scales used by the World Heritage Report included all of Mayotte and Comoros as a priority whereas the selection method identified 3 specific sites in Mayotte and 8 in Comoros. World Heritage selected two large areas at the northwestern and southern tips or Madagascar whereas criteria used here identified 21 more specific locations in Madagascar (Table 1 ). SWIOFP identified 3 specific sites, but few of these selections were supported by this biodiversity analyses. SWIOFP used different criteria and therefore chose Antongil Bay for the hump-back whale breeding sanctuary, Nosy Ve for tropical breeding birds, and Grand Recif for coral diversity based on historical information that was based on extensive but spatially restricted sampling. Management priorities Fish biomass Designations such as MNR or LMMA do not necessarily provide evidence for protection or the status of the resources (Randrianarivo et al. 2023 ). Past work has shown that fish biomass is one of the single best indicators of resources status and biodiversity. Biomass is also expected to be manageable by fisheries regulations. An empirical field study of Madagascar and Mayotte reef fishes concluded that managing the biomass of fish is more important than site selection (McClanahan and Jadot 2017 ). However, the model used here held biomass constant so that it was possible to visualize and compare sites for the same biomass (600 kg/ha) and depth (10 m). This produces spatial variability driven more by non-biomass factors, which exposes the underlying biodiversity patterns expected without the current high human fishing pressure. This approach produces more environmentally driven spatial variability and helped with the identification of several prioritization sites mentioned above. Nevertheless, managing biomass on large scales is likely to be one of the simplest and most effective targeted approaches to protect biodiversity while maintaining optimal fisheries yields. By focusing on this goal, it is possible to indirectly protect fish biodiversity, maximize yields, and manage other ecological processes (McClanahan 2018 , 2022). Fish biomass is managed through the usual restrictions on access, fishing gear, times, locations, and capture choices (McClanahan and Abunge 2020 ). These preferences vary with fishers and communities, but people are more willing to agree to restrictions when local governance is effective. Some poorly known influential national attributes include fisheries governance, laws, compliance, and human dependence on fish resources. In general, managing fisheries for maximum sustained yields on large scales combined with closure areas approaching 30% of the nearshore is likely to achieve optimal management (Kerwath et al. 2013 ; McClanahan 2021 , 2022a ). An evaluation of fish biomass in the Northwest Madagascar Ecoregion found high variability but that 65% of the studied reefs had biomass greater than the recommended maximum sustained yield level of 600 kg/ha (McClanahan and Jadot 2017 ). Most of the lower biomass levels were close to urban centers or in the southwest where dry conditions make fishing one of the few animal food options (Bruggemann et al. 2012 ). Most reefs in Mayotte had biomass above this level but Comoros values are lower. Fish stock biomass is an important indicator of the success of tradeoffs between human needs and biodiversity conservation. Coral diversity Coral biodiversity was primarily driven by environmental forces of temperature stresses, currents, and waves. These environmental forces will change as climate warms. Climate predictions suggest further degradation of the reef environment by 2050 for both climate scenarios. After accounting for fish biomass and depth, the strongest predictors of the total biodiversity proxy were median SST and metrics of variability (excess heat, kurtosis or chronic SST variability, rate of SST rise, bimodality, and skewness). Many of these variables are frequently among the key factors that influence coral bleaching and cover. For example, the standard climate-coral impact models frequently include cumulative excess heat, and the rate of SST rise as the main stresses (McClanahan 2022b ). However, studies in the WIO province have shown that SST variability metrics of kurtosis, bimodality, and skewness had equal or greater influences on both biodiversity and coral cover (McClanahan and Azali 2021 ). This environmental complexity produces considerable spatial variability in the responses predicted for the two climate change scenarios. Both coral and fishes are affected by the historical patterns of acute and chronic stress and not just excess heat. In fact, modest amounts of excess heat promote corals acclimation to climate change (McClanahan and Azali 2021 ). Therefore, high environmental variability around Madagascar created the heterogeneous responses among the 2050 predictions. Heterogeneity is evident in the predicted losses of taxa, with both gains and losses of cover and biodiversity. Most reef cells were predicted by the Business-as-Usual scenario to have both losses of coral cover and taxa (55 to 72%). However, a smaller number of reefs may escape the consequences of climate change in the next 30 years if the carbon emission reductions are achieved and other human local disturbances reduced. The smaller group predicted to benefit from climate change was between 7 and 14% of reef cells. A modelling study of Madagascar provided evidence that deforestation of watersheds may be more detrimental to corals than climate change, even where rainfall is predicted to decline in some provinces (Maina et al. 2013 ). While southwest Madagascar may be a refuge when seen from climate impact perspective (Sully et al. 2022 ), the local fishing impacts are expected to undermine the climate refuge prediction (Bruggemann et al. 2012 ; Ranaivomanana et al. 2023 ). Model caveats Several technical advancements in conservation science were represented in the presented approach and outcomes (Pilowsky et al. 2022 ). Nevertheless, the availability of field data was core to the empirical predictions that allowed broader scale predictions. Fortunately, the collaborators shared methods that were comparable (McClanahan et al. 2007 ). Moreover, the machine learning model accounted for observer and biomass effects when making predictions. In the case of fishes, the machine learning model held biomass constant and above the saturation level required to make comparable predictions of numbers of species between reef cells. The flexibility and ability to control for many factors is a key strength of the machine learning and partial effects approach. Tests of model performance using cross-validation procedures indicated good predictive ability. Yet, when making predictions for many cells on large scales, there is the possibility of overfitting and missing important local conditions, especially below the 6.25-km 2 scale. Moreover, the model cannot account for local unmodelled variables, such as damaging fishing methods, rare cyclones, or point-source pollution. Nevertheless, the model represents a considerable advance in marine spatial modelling in a country lacking broad-scale and comparable historical data. Challenges remain to test predictions and account for human and other local factors not currently available. The outcomes will, however, depend on the metrics and values of the assessments, such as ecological functions of taxa, their various values, uniqueness, evolutionary relatedness, and threats (Brooks et al. 2006 ). The methods used here did not focus on some important conservation concerns, such as remoteness, large body sizes, species with connections to land, and rarity. Nevertheless, hard coral is a well sampled invertebrate and a taxon with many other dependent species. When corals are combined with fish, the diversity proxy includes multiple taxon known for high species richness. The positive correlation found for these two groups indicated that these taxa should often correlate with a more diverse group of unsampled taxa, such as species associated with seagrass and mangroves (McClanahan et al. 2024 ). The modest fit for the coral-fish correlation suggests that the two faunal groups combined should provide a general but not necessarily accurate proxy for total richness. Past prioritization efforts often share selection choices or biases in terms of selected taxa and human impacts. Many of these selections were justifiable based on immediate threats and needs of rare, space-requiring, and sensitive species. However, the model used here made a first estimate of broad-scale total diversity at finer scales than previous efforts. Past selections have disproportionally focused on large-bodied species with broad distributions and threats. Some other species, such as the southern humpback whales ( Megaptera novaeangliae ) and nesting birds have broad ranges, and the identified sites focused on their nesting locations, habitats, invasive species, and reproductive needs (Feare et al. 2007 ; Le Corre et al. 2015 ; Amaral et al. 2016 ). However, the large migrations of these species suggest population viability may require more than sites for reproduction. Management across these species’ migratory ranges needs to be combined with the protection of specific breeding locations. Nevertheless, mating and nesting sites can be critical for many species and are therefore a good attribute for making conservation decisions. While visible and charismatic species may promote ecotourism, tourism often faces economic limits and instability in countries with volatile politics, such as Madagascar and Comoros (Spash 2021 ). Therefore, providing smaller-scale information that focuses on species with important ecological services should be useful for decision makers that need to balance food production with healthy and diverse ecosystems. The biodiversity proxy method will fail to identify locations with specific species attractions, such as breeding birds and marine mammals. It should, however, be more objective for other subtidal taxa and suggests that Grand Recif was not more diverse than other locations. Grand Recif received much attention from taxonomists, and this may have created the perception of high diversity but also represents high sampling effort. Nevertheless, without comparing Grand Recif with other locations using the same sampling effort, the conclusions can be challenged. Field studies comparing corals for similar levels of sampling between regions concluded there were smaller differences among provinces (Randrianarivo et al. 2022 ). Other studies suggest that the northwest Nosy Be-Ankarea locations had the highest number of coral taxa due to rare taxa often found at depth (McClanahan et al. 2009 , 2011 ). Randrianarivo et al. ( 2022 ) found that fishing and the abundance of herbivorous fishes was a more important influence on coral diversity than region. The model makes fair comparisons by holding factors constant to avoid these confounding problems. Thus, the model predictions better represent a baseline than an empirical post-human impact evaluation. Conclusions The above approach provided a more spatially refined view of biodiversity than previous national, global, and regional efforts (Allnut et al. 2012; Jenkins and Van Houtan 2016 ). The focus on Madagascar ensures the ranking of sites is relevant to this island region and not diminished by potentially higher diversity locations, such as the African coastline or Coral Triangle (Kusumoto et al. 2020 ; McClanahan et al. 2024 ). Previous efforts have used the presence/absence of key taxa, limited sampling, and charismatic and visible taxa rather than the smaller-scale environmental conditions that create niches that support species and affect their distributions (van der Elst and Everett 2016). Environmental modelling provides an alternative broad-scale approach by accounting for environmental variability and ecological processes at smaller scales. These methods were better able to identify potential patterns of causation and small to modest-sized area-based conservation planning and management needs. Given more testing, the methods should eventually replace or augment past and current practices of mapping diversity from sparse or selective information. The method requires empirical data, so models are expected to improve as these data increase in scale and resolution. Mapped cell-level predictions showed that smaller scale environmental factors were among the strongest influences on predictions. Therefore, failure to account for this variability is likely to produce weaknesses in past recommendations based on single species and coarser scale biogeographic factors. Clearly, ground-truthing and evaluating environmental and important demographic influences will be an important next step in the prediction and prioritization process. However, the current model provided a selection of 32 locations of modest size in Madagascar, Mayotte, and Comoros that should be among those considered for focused management (Table 1 ). Fortunately, a high percentage of these sites have already either been included in national or LMMA management systems. The challenge now is to ensure that selected delineations receive appropriate and effective management to secure their biodiversity. The pressures of climate change, increasing cyclones, and the continual demand for fish and forest resources will continue to challenge the need to resolve nature and human need tradeoffs. Addressing climate change will require participation towards carbon emission reductions goals. This requires political action to promote emission reductions nationally and globally. Reforestation that addresses watershed management and carbon emissions is also another high priority solution (Maina et al. 2013 ). Research in fisheries is showing there are win-win outcomes for biodiversity and fisheries yields when managing catch for maximum sustained yields (McClanahan 2018 , 2022a ). Therefore, keeping fish stocks near or above optimal levels (600 kg/ha) is a practical goal to get yield and biodiversity benefits. Decisions to act on the recommended priorities will require strengthening governance institutions and commitments at multiple scales to achieve adaptation goals. Declarations Author Contribution T.R.M. concepualized the study, raised the funds, collected the field data, supervised the analysis of data and production of tables and figures, and wrote and edited the manuscript Data Availability Links to environmental data are available in cited data source publications (McClanahan et al. 2024). Access to field data collected in the study sites requires a formal request to the author. References Allen GR, Werner TB (2002) Coral reef fish assessment in the ‘coral triangle’ of southeastern Asia. Environ Biol Fish 65:209–214 Allnutt TF, McClanahan TR, Andréfouët S, Baker M, Lagabrielle E, McClennen C, Rakotomanjaka AJ, Tianarisoa TF, Watson R, Kremen C (2012) Comparison of marine spatial planning methods in Madagascar demonstrates value of alternative approaches. PLoS ONE 7:e28969 Amaral AR, Loo J, Jaris H, Olavarria C, Thiele D, Ensor P, Aguayo A, Rosenbaum HC (2016) Population genetic structure among feeding aggregations of humpback whales in the Southern Ocean. Mar Biol 163:1–13 Andrello M, Darling ES, Wenger A, Suárez-Castro AF, Gelfand S, Ahmadia GN (2022) A global map of human pressures on tropical coral reefs. Conserv Lett 15:e12858 Brooks TM, Mittermeier RA, Da Fonseca GAB, Gerlach J, Hoffmann M, Lamoreux JF, Mittermeier CG, Pilgrim JD, Rodrigues ASL (2006) Global biodiversity conservation priorities. Science 313:58–61 Bruggemann JH, Rodier M, Guillaume MM, Andréfouët S, Arfi R, Cinner JE, Pichon M, Ramahatratra F, Rasoamanendrika F, Zinke J, McClanahan TR (2012) Wicked social–ecological problems forcing unprecedented change on the latitudinal margins of coral reefs: The case of southwest Madagascar. Ecol Soc 17 Burke L, Reytar K, Spalding M, Perry A (2012) Reefs at risk revisited in the coral triangle 978-1-56973-791-0. World-Fish Center, International Coral Reef Action Network, Washington, D. C Cinner JE, Huchery C, MacNeil MA, Graham NA, McClanahan TR, Maina J, Maire E, Kittinger JN, Hicks CC, Mora C, Allison EH, D’Agata S, Hoey A, Feary DA, Crowder L, Williams ID, Kulbicki M, Vigliola L, Wantiez L, Edgar G, Stuart-Smith RD, Sandin SA, Green AL, Hardt MJ, Beger M, Friedlander A, Campbell SJ, Holmes KE, Wilson SK, Brokovich E, Brooks AJ, Cruz-Motta JJ, Booth DJ, Chabanet P, Gough C, Tupper M, Ferse SCA, R. U., Sumaila, Mouillot D (2016) Bright spots among the world’s coral reefs. Nature 535:416–419 Cooke A, Lutjeharms JRE, Vasseur P (2003) Marine and coastal ecosystems. In: Goodman SM, Benstead JP (eds) The Natural History of Madagascar. University of Chicago Press, pp 329–474 Cowburn B, Samoilys M, Obura D (2018) The current status of coral reefs and their vulnerability to climate change and multiple human stresses in the Comoros Archipelago, Western Indian Ocean. Mar Pollut Bull 133:956–969 Crochelet E, Roberts J, Lagabrielle E, Obura D, Petit M, Chabanet P (2016) A model-based assessment of reef larvae dispersal in the Western Indian Ocean reveals regional connectivity patterns—Potential implications for conservation policies. Reg Stud Mar Sci 7:159–167 Elith J, Leathwick JR, Hastie T (2008) A working guide to boosted regression trees. J Anim Ecol 77:802–813 Feare CJ, Jaquemet S, Le Corre M (2007) An inventory of Sooty Terns (Sterna fuscata) in the western Indian Ocean with special reference to threats and trends. Ostrich-Journal Afr Ornithol 78:423–434 Fontoura L, D’agata S, Gamoyo M, Barneche DR, Luiz OJ, Madin EMP, Eggertsen L, Maina JM (2022) Protecting connectivity promotes successful biodiversity and fisheries conservation. Science 375:336–340 Gardner CJ, Cripps G, Day LP, Dewar K, Gough C, Peabody S, Tahindraza G, Harris A (2020) A decade and a half of learning from Madagascar's first locally managed marine area. Conserv Sci Pract 2:e298 Gardner CJ, Nicoll ME, Birkinshaw C, Harris A, Lewis RE, Rakotomalala D, Ratsifandrihamanana AN (2018) The rapid expansion of Madagascar’s protected area system. Biol Conserv 220:29–36 Gough CL, A. KM, Dewar BJ, Godley E, Zafindranosy, Broderick AC (2020) Evidence of overfishing in small-scale fisheries in Madagascar. Front Mar Sci 7:317 Grace MK, Akçakaya HR, Bennett EL, Boyle MJW, Hilton-Taylor C, Hoffmann M, Money D, Prohaska A, Young R, Young R, and B. L (2022) The Impact of Spatial Delineation on the Assessment of Species Recovery Outcomes. Diversity 14:742 Hansen JE, Sato M, Simons L, Nazarenko LS, Sangha I, Kharecha P, Zachos JC, von Schuckmann K, Loeb NG, Osman MB (2023) Global warming in the pipeline. Oxf Open Clim Change 3:kgad008 Jenkins CN, Van Houtan KS (2016) Global and regional priorities for marine biodiversity protection. Biol Conserv 204:333–339 Kerwath SE, Winker H, Götz A, Attwood CG (2013) Marine protected area improves yield without disadvantaging fishers. Nat Commun 4:2347 Kuhn M, Johnson K (2013) Applied predictive modeling. Springer, New York Kusumoto B, Costello MJ, Kubota Y, Shiono T, Wei CL, Yasuhara M, Chao A (2020) Global distribution of coral diversity: Biodiversity knowledge gradients related to spatial resolution. Ecol Res 35:315–326 Le Corre M, Danckwerts DK, Ringler D, Bastien M, Orlowski S, Rubio CM, Pinaud D, Micol T (2015) Seabird recovery and vegetation dynamics after Norway rat eradication at Tromelin Island, western Indian Ocean. Biol Conserv 185:85–94 Le Manach F, Gough C, Harris A, Humber F, Harper S, Zeller D (2012) Unreported fishing, hungry people and political turmoil: The recipe for a food security crisis in Madagascar? Mar policy 36:218–225 Maina J, De Moel H, Zinke J, Madin J, McClanahan T, Vermaat JE (2013) Human deforestation outweighs future climate change impacts of sedimentation on coral reefs. Nat Commun 4:1–7 Maina J, McClanahan TR, Venus V, Ateweberhan M, Madin J (2011) Global gradients of coral exposure to environmental stresses and implications for local management. PLoS ONE 6:e23064 Maina JM, Gamoyo M, Adams VM, D'agata S, Bosire J, Francis J, Waruinge D (2020) Aligning marine spatial conservation priorities with functional connectivity across maritime jurisdictions. Conserv Sci Pract 2:e156 Maire E, Cinner J, Velez L, Huchery C, Mora C, Dagata S, Vigliola L, Wantiez L, Kulbicki M, Mouillot D (2016) How accessible are coral reefs to people? A global assessment based on travel time. Ecol Lett 19:351–360 McClanahan TR (2018) Multicriteria estimate of coral reef fishery sustainability. Fish Fish 19:807–820 McClanahan TR (2021) Marine reserve more sustainable than gear restriction in maintaining long-term coral reef fisheries yields. Mar policy 128:104478 McClanahan TR (2022a) Fisheries yields and species declines in coral reefs. Environ Res Lett 17:044023 McClanahan TR (2022b) Coral responses to climate change exposure. Environ Res Lett 17:073001 McClanahan TR (2023) Local heterogeneity of coral reef diversity and environmental stress provides opportunities for small-scale conservation. Divers Distrib 29:1324–1340 McClanahan TR, Abunge CA (2019) Conservation needs exposed by variability in common-pool governance principles. Conserv Biol 33:917–929 McClanahan TR, Abunge CA (2020) Perceptions of governance effectiveness and fisheries restriction options in a climate refugia. Biol Conserv 246:108585 McClanahan TR, Ateweberhan M, Graham NAJ, Wilson SK, Sebastián CR, Guillaume MM, Bruggemann JH (2007) Western Indian Ocean coral communities: Bleaching responses and susceptibility to extinction. Mar Ecol Prog Ser 337:1–13 McClanahan TR, Ateweberhan M, Omukoto J, Pearson L (2009) Recent seawater temperature histories, status, and predictions for Madagascar's coral reefs. Mar Ecol Prog Ser 380:117–128 McClanahan TR, Azali MK (2021) Environmental variability and threshold model’s predictions for coral reefs. Front Mar Sci 8:1774 McClanahan TR, Friedlander AM, Wickel J, Graham NAJ, Bruggemann HJ, Guillaume MMM, Chabanet P, Porter S, Schleyer MH, Azali MK, Muthiga NA (2024) Testing for concordance between predicted species richness, past prioritization, and marine protected area designations in the western Indian Ocean. Conservation Biology:e14256 McClanahan TR, Graham NAJ (2015) Marine reserve recovery rates towards a baseline are slower for reef fish community life histories than biomass. Proceedings of the Royal Society B: Biological Sciences 282:20151938 McClanahan TR, Graham NAJ, MacNeil MA, Cinner JE (2015) Biomass-based targets and the management of multispecies coral reef fisheries. Conserv Biol 29:409–417 McClanahan TR, Jadot C (2017) Managing coral reef fish community biomass is a priority for biodiversity conservation in Madagascar. Mar Ecol Prog Ser 580:169–190 McClanahan TR, Maina JM, Muthiga NA (2011) Associations between climate stress and coral reef diversity in the Western Indian Ocean. Glob Change Biol 17:2023–2032 Obura DO, Church JE, Gabrié C (2012) Assessing Marine World Heritage from an Ecosystem Perspective: The Western Indian Ocean. World Heritage Centre, United Nations Education, Science and Cultural Organization (UNESCO), Paris, France Parker QM, Longosoa HT, Long S, Jones PJS (2024) A longitudinal governance analysis of a locally managed marine area: Ankobohobo wetland small-scale mud crab fishery. Madagascar Mar policy 163:106138 Parravicini V, Villéger S, McClanahan TR, Arias-González JE, Bellwood DR, Belmaker J, Chabanet P, Floeter SR, Friedlander AM, Guilhaumon F, Vigliola L, Kulbicki M, Mouillot D (2014) Global mismatch between species richness and vulnerability of reef fish assemblages. Ecol Lett 17:1101–1110 Pilowsky JA, Colwell RK, Rahbek C, Fordham DA (2022) Process-explicit models reveal the structure and dynamics of biodiversity patterns. Sci Adv 8:eabj2271 Ranaivomanana HS, Jaquemet S, Ponton D, Behivoke F, Randriatsara RM, Mahafina J, Léopold M (2023) Intense pressure on small and juvenile coral reef fishes threatens fishery production in Madagascar. Fish Manag Ecol 30:494–506 Randrianarivo M, Botosoamananto RL, Guilhaumon F, Penin L, Todinanahary G, Adjeroud M (2023) Effects of Madagascar marine reserves on juvenile and adult coral abundance, and the implication for population regulation. Mar Environ Res :106080 Randrianarivo M, Gasimandova LM, Tsilavonarivo J, Razakandrainy A, Philippe J, Guilhaumon F, Botosoamananto RL, Penin L, Todinanahary G, Adjeroud M (2024) Assessing recovery potential of coral reefs in Madagascar and the effects of Marine Protected Areas. Regional Studies in Marine Science:103710. Randrianarivo M, Guilhaumon F, Tsilavonarivo J, Razakandrainy A, Philippe J, Botosoamananto RL, Penin L, Todinanahary G, Adjeroud M (2022) A contemporary baseline of Madagascar’s coral assemblages: Reefs with high coral diversity, abundance, and function associated with marine protected areas. PLoS ONE 17:e0275017 Rocliffe S, Peabody S, Samoilys M, Hawkins JP (2014) Towards a network of locally managed marine areas (LMMAs) in the Western Indian Ocean. PLoS ONE 9:e103000 Samoilys M, Alvarez-Filip L, Myers R, Chabanet P (2022) Diversity of coral reef fishes in the western Indian Ocean: Implications for Conservation. Diversity 14:102 Smallhorn-West P, Cohen PJ, Kukiti E, Rabbitt S, Sori F, Tapala S, Warren R, Boso D, Foale S (2022) Ten years of dynamic co-management of a multi-species reef fishery. Coral Reefs 41:1449–1464 Spash CL (2021) The economy’as if people mattered: revisiting critiques of economic growth in a time of crisis. Globalizations 18:1087–1104 Sully S, Hodgson G, van Woesik R (2022) Present and future bright and dark spots for coral reefs through climate change. Glob Change Biol 28:4509–4522 Tyberghein L, Verbruggen H, Pauly K, Troupin C, Mineur F (2012) and O. De Clerck. Bio-ORACLE: A global environmental dataset for marine species distribution modelling. Global Ecology and Biogeography 21:272–281 Van der Elst R, Everett B (2015) Offshore fisheries of the Southwest Indian Ocean: Their status and the impact on vulnerable species. Oceanographic Research Institute (ORI) and the Western Indian Ocean Marine Sciences Association (WIOMSA), Durban Wells S, Ray GC, Gjerde KM, White AT, Muthiga N, Bezaury JE, Creel BD, Causey J, McCormick-Ray R, Salm S, Gubbay, Kelleher G (2016) Building the future of MPAs–lessons from history. Aquatic Conservation: Marine and Freshwater Ecosystems 26:101–125 Wilson K, L. DP, Tittensor B, Worm, Lotze HK (2020) Incorporating climate change adaptation into marine protected area planning. Glob Change Biol :1–17 Convention UNEP-N and WIOMSA (2021) Western Indian Ocean Marine Protected Areas Outlook: Towards achievement of the Global Biodiversity Framework Targets. UNEP and WIOMSA, Nairobi, Kenya Yeager LA, Marchand P, Gill DA, Baum JK, McPherson JM (2017) Marine socio-environmental covariates: Queryable global layers of environmental and anthropogenic variables for marine ecosystem studies. Ecology 98:1976–1976 Zafimahatradraibe JA, Ranaivomanana LNJ, Rakotomahazo C, Randriamanantsoa B, Hartmann AC, Todinanahary GGB (2025) Authority, capacity, and power to govern: Three marine protected areas co-managed by resource users and non-governmental organizations. Mar policy 177:106647 Zeller D, Vianna G, Ansell M, Coulter A, Derrick B, Greer K, Noël SL, Palomares MLD, Zhu A, Pauly D (2021) Fishing effort and associated catch per unit effort for small-scale fisheries in the Mozambique Channel region: 1950–2016. Front Mar Sci 8:1048 Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7932631","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":543843084,"identity":"16e762e0-909b-47f3-8ccc-621a10abfde6","order_by":0,"name":"TIm McClanahan","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA1klEQVRIiWNgGAWjYLCCB0BswMDA+IAo1TwgIgGihdmAZC1sEkRpsRc7/PBDQs0deXP2s8eqeWrqGPilj1/Ab4t0mrFEwrFnhjt78tJu8xw7zCDZl1NAQEsOG0MC22HGDQdyzG7nNhxgMDjDk0CEln+H7Tecf2NWnNtQR6SWxLbDiRtu5Jgx5zYwA7WwH8Cv5TbQL4l9h5M33HhjLP3n2GEeyR4evDoY2GcnP/zw4dth2w3ncww/zqipk+PnYX+AXw+GtUBEXJSi2EyiLaNgFIyCUTDcAQDzRkTzMccHsgAAAABJRU5ErkJggg==","orcid":"","institution":"Wildlife Conservation Society","correspondingAuthor":true,"prefix":"","firstName":"TIm","middleName":"","lastName":"McClanahan","suffix":""}],"badges":[],"createdAt":"2025-10-23 13:23:39","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7932631/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7932631/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":96296485,"identity":"4d6cddc9-72ec-4c1b-8919-9256dd2cef38","added_by":"auto","created_at":"2025-11-19 13:38:57","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":13307507,"visible":true,"origin":"","legend":"","description":"","filename":"MadagascarBiodiversity2.0.docx","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/0b33cc54c1b75780dd3b33f7.docx"},{"id":96296482,"identity":"6c90096a-32e1-44b1-9a13-5503242fbc0d","added_by":"auto","created_at":"2025-11-19 13:38:57","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3678,"visible":true,"origin":"","legend":"","description":"","filename":"0b64a7eed4834c8cae8041a6762620d1.json","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/252d40b84939aebc87dd1d96.json"},{"id":96365669,"identity":"f39db115-8ca2-4d3f-8214-d0addfd3598a","added_by":"auto","created_at":"2025-11-20 10:10:40","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":196617,"visible":true,"origin":"","legend":"","description":"","filename":"0b64a7eed4834c8cae8041a6762620d11enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/bf6e53e8bc4ba88396962e15.xml"},{"id":96364639,"identity":"c6592d93-7da3-4691-a8e9-599189c168f0","added_by":"auto","created_at":"2025-11-20 10:09:29","extension":"emf","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":8700572,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.emf","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/4151e85db1d36c273b1e9728.emf"},{"id":96364150,"identity":"c8dffc98-021b-4524-9e7f-1d8cdf3d6dfb","added_by":"auto","created_at":"2025-11-20 10:08:59","extension":"emf","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2129404,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.emf","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/8e050c6752aa4b53d266357d.emf"},{"id":96296489,"identity":"0b01a6e3-35d8-4a07-8eca-f6cbb8b2b06f","added_by":"auto","created_at":"2025-11-19 13:38:58","extension":"emf","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4793568,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.emf","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/24b2075911bc493ad155d78b.emf"},{"id":96365584,"identity":"ccf5c121-3915-42e5-b373-88f7a7fba4ba","added_by":"auto","created_at":"2025-11-20 10:10:33","extension":"emf","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4116732,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.emf","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/109ade48f7212d720bc767e4.emf"},{"id":96296483,"identity":"0ca2d9be-048b-494b-b67e-514a16861fb4","added_by":"auto","created_at":"2025-11-19 13:38:57","extension":"emf","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4178964,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.emf","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/8a6c62010fb7fe79097f3aae.emf"},{"id":96296469,"identity":"7bea613d-aeb4-44c3-89c3-2b51a980afdb","added_by":"auto","created_at":"2025-11-19 13:38:56","extension":"emf","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3972516,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.emf","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/8ad10b23ee68ccd24c3740dc.emf"},{"id":96296470,"identity":"fbcdedef-f22a-47b0-98b7-0daac891deb4","added_by":"auto","created_at":"2025-11-19 13:38:56","extension":"emf","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2830796,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.emf","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/71eab674ee80104178e0a926.emf"},{"id":96296481,"identity":"099bfc89-3dcc-406a-8a2d-743db552bc9f","added_by":"auto","created_at":"2025-11-19 13:38:57","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":48076,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/4b906d1e5fab8506bfcd4293.png"},{"id":96296476,"identity":"573bac6c-65da-48d3-9152-029b1d46d24c","added_by":"auto","created_at":"2025-11-19 13:38:56","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":28816,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/9bcb7e4d72f10f75263b2eae.png"},{"id":96296487,"identity":"6910cd89-9dda-4085-a5f6-526f3f6bafc1","added_by":"auto","created_at":"2025-11-19 13:38:58","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":27377,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/b8d49ff0c2c8283e26238ee1.png"},{"id":96364844,"identity":"71663659-8435-435a-8d35-422e1e20e83f","added_by":"auto","created_at":"2025-11-20 10:09:43","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":22873,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/d9c200603e6cdda29ba0ca9f.png"},{"id":96296480,"identity":"0ab71b79-925d-460c-b4e1-d90afaf9fc65","added_by":"auto","created_at":"2025-11-19 13:38:57","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":23565,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/4875a55922bec754667069e5.png"},{"id":96296490,"identity":"c5622021-744c-426d-83f8-71066f9ae7e8","added_by":"auto","created_at":"2025-11-19 13:38:58","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":23414,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/1e084b8e4aacff1964033c60.png"},{"id":96296478,"identity":"718e636f-eb15-4074-9e07-9bd708bcbbbb","added_by":"auto","created_at":"2025-11-19 13:38:57","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":21834,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/b41c95feb4f28c13705e031d.png"},{"id":96296477,"identity":"8af4ef5d-a208-40ac-b58f-73fef1d2355d","added_by":"auto","created_at":"2025-11-19 13:38:56","extension":"xml","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":194796,"visible":true,"origin":"","legend":"","description":"","filename":"0b64a7eed4834c8cae8041a6762620d11structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/3a0d13c9d122c507ae0fe349.xml"},{"id":96296475,"identity":"a20e6eb5-1ddd-4c44-b4a5-5b4a25a783f1","added_by":"auto","created_at":"2025-11-19 13:38:56","extension":"html","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":202626,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/dd984bca3357a5bec9005f8b.html"},{"id":96296472,"identity":"460224fb-8afa-4da9-8263-a82b1303584c","added_by":"auto","created_at":"2025-11-19 13:38:56","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":246299,"visible":true,"origin":"","legend":"\u003cp\u003eDistribution of World Protected Areas Database (WDPA) protected areas and our methods biodiversity hotspot priorities in (a) Madagascar and (b) Mayotte and Comoros Islands. Areas selected in past SWIOFP, and Marine Heritage Reports are also mapped. The Mayotte MNP that covers the full island and extensive offshore areas is not shown.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/d157f2b3d3bf572b6e09e418.png"},{"id":96296471,"identity":"7ddfc54e-6db6-441a-b58b-102c95604121","added_by":"auto","created_at":"2025-11-19 13:38:56","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":41981,"visible":true,"origin":"","legend":"\u003cp\u003eStudy locations sampled for corals and fish in the study region to calibrate the machine learning predictive model.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/25a1f1986becead16946b789.png"},{"id":96296467,"identity":"78c7db3d-4d99-4dea-8cff-7b2b54b3d8c8","added_by":"auto","created_at":"2025-11-19 13:38:55","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":129887,"visible":true,"origin":"","legend":"\u003cp\u003eDistribution of the number of (a) coral per 40-m\u003csup\u003e2\u003c/sup\u003e, (b) fish per 500-m\u003csup\u003e2\u003c/sup\u003e, and (c) a proxy for total number of taxa.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/9089f36f20aec0367ceb854c.png"},{"id":96296468,"identity":"2e9c8880-abd8-4427-bb5c-e35837579a36","added_by":"auto","created_at":"2025-11-19 13:38:55","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":148339,"visible":true,"origin":"","legend":"\u003cp\u003eScatterplot of the changes in coral cover and numbers of taxa in Madagascar Provinces, Comoros Islands Njazida (previously known as Grande Comoros), Ndzuwani (Anjuan), and Mwali (Moheli), and Mayotte French Territories between 2020 and 2050 based on IPCC temperature time series variables.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/eb298cab0bf5e45240decda7.png"},{"id":96453157,"identity":"8089a000-9b61-4f46-9a6a-ec6e4f2c0025","added_by":"auto","created_at":"2025-11-21 09:58:34","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1845559,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7932631/v1/b93ce798-a57a-4f29-88ba-10c30126a4fd.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Modelling coral reef biodiversity for prioritizing marine protected area in Madagascar and adjacent islands ","fulltext":[{"header":"Introduction","content":"\u003cp\u003eExpansion of protected areas requires making decisions focused on factors of which biodiversity and adaptation to climate change are centrally important (Wilson et al. \u003cspan citationid=\"CR61\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). However, many protected area prioritization studies are based on limited data and have failed to objectively articulate appropriate spatial and temporal scales in the patterns of biodiversity. This data limitation problem is particularly acute for the increasing numbers of local and small-scale management and protected area activities (McClanahan \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). A common limitation of protected area prioritization is the limited number of observations and sampling relative to the number of taxa and the area of coverage. For example, large-scale taxonomic diversity maps are often biased by the patchy and haphazard nature of species occurrence records (Kusumoto et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Thus, biodiversity distributions over large areas of poorly studied coasts are frequently based on extrapolations from presence/absence data in sparsely sampled cell grids (Jenkins and van Houtan \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). To a better understanding the distribution of numbers of reef species along a poorly sampling coastline of Madagascar and its smaller island neighbors of Comoros and Mayotte, a biodiversity and climate change modelling approach was developed here. The purpose being to model coral reef cover and biodiversity in poorly known locations and thereby provide a gap analysis of biodiversity and marine protected areas for 2020.\u003c/p\u003e\n\u003ch3\u003eMadagascar and associated islands\u003c/h3\u003e\n\u003cp\u003eMadagascar and the surrounding nations of Comoros and Mayotte are islands known for their unique biodiversity (Cook et al. 2003; Parravicini et al. 2011). Yet, knowledge of marine biodiversity is fragmentary due to lack of widespread data collection throughout the region. Additionally, Madagascar is isolated geographically but at the center of ocean current connectedness (Crochelet et al. \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Maina et al. \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Biodiversity investigations of colonial French investigators undertaken between 1961 and 1972 were limited to the southwest regions around Toliara. These investigators focused on taxonomy and identified many marine taxa (\u0026gt;\u0026thinsp;6000) in the Grand Recif of Toliara (Cooke et al. \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2003\u003c/span\u003e). Subsequent comparisons among Madagascar\u0026rsquo;s regions have been limited to three areas, namely the reefs near Toliara, the islands around Nosy Be, and the Masoala Peninsula (McClanahan et al. \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; McClanahan and Jadot \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Samoilys et al. \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Randrianarivo et, al. 2022, 2023, 2024). These studies find the marine ecosystems are highly affected by human use, but observations are also often limited to accessible areas where human impacts are expected (McClanahan and Jadot \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Cowburn et al. \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Gough et al. \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Randrianarivo et al. \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Additionally, there is evidence for some spatial homogeneity of the reef communities that may result from the island and isolated nature compared to the more extensive African coastline (McClanahan and Jadot \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Samoilys et al. \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; McClanahan et al. \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Thus, there is much to learn about Madagascar\u0026rsquo;s broader and finer spatial scale biodiversity patterns. Despite the poor knowledge, marine conservation planning and implementation is now rapidly underway with the increasing commitment to local scale community conservation (Gardner et al. \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2018\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eCoral reefs contain the highest shallowest water diversity, so coral reef diversity should be a good proxy for overall numbers of species. Mapping of coral reefs showed that most of the ~\u0026thinsp;14,000 km\u003csup\u003e2\u003c/sup\u003e of Madagascar\u0026rsquo;s reefs are found on the western and northern sides of the island (Burke et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). Madagascar has two ecoregions, namely the large Western and Northern Madagascar Ecoregion that contain the two nations of Comoros and Mayotte. The third, or Southeast Madagascar Ecoregions contains fewer coral reefs. For perspective, Madagascar\u0026rsquo;s coral reefs are ~\u0026thinsp;20% of the total reef area in the western Indian Ocean marine province (McClanahan et al. \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eMadagascar faces climate change predictions of rising temperatures, changing rainfall patterns, and more intense cyclones (UNEP Interactive Country Fiches; dicf.unep.ch). The south is expected to see the most significant warming, while the north may experience a decrease in rainfall. Climate changes are projected to increase droughts and impact agriculture and potentially increase reliance on marine resources. Coastal areas will experience sea-level rise and coastal erosion. Site specific studies of coral reefs show considerable variability arising from differences in temperature variability and fishing pressure (McClanahan et al. \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2009\u003c/span\u003e; Randrianarivo et al. \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2022\u003c/span\u003e, \u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The southwest, for example, has lower radiation, cooler temperatures, and rainfall than the north. Therefore, thermal stress characteristics and watershed runoff are reduced (Sully et al. \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), but the southwest is also associated with cyclones and greater human dependence on marine resources (Bruggemann et al. \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). The high thermal stress and rainfall in the north may increase watershed runoff but the land also produces more food, which can reduce human dependency and pressure on northern fisheries (Cinner et al. \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; McClanahan and Jadot \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Additionally, warm water conditions can also acclimate corals to ongoing and future warming (McClanahan et al. \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). Given the coastal variability in natural ecological processes and human dependence, the question remains how coral cover and species will be affected and distributed in the face of increasing climate change and human pressures.\u003c/p\u003e\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003ePast protected area prioritization, planning, and actions\u003c/h2\u003e\u003cp\u003ePast biodiversity conservation planning reports have identified four areas in Madagascar, namely the Masoala Peninsula, the Grand Recif of Toliara, the northwest regions of Antsiranana, and the extreme southern tip of the island. These selections were based on knowledge of the existence of whales, high coral diversity, and remote and potentially unique but unstudied fauna (Obura et al. \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; van det Elst and Everett \u003cspan citationid=\"CR59\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Yet, in most cases plans may reflect a bias dependent on the long history of study of marine faunal taxonomic work in Toliara, observations of whale migrations, bird nesting, and concerns to protect remote areas with few people. Therefore, how past priority and existing protected area selections will differ from environment-taxa modelling predictive approaches requires investigation.\u003c/p\u003e\u003cp\u003eConservation actions have developed rapidly during the past 20 years in these three countries (Gardner et al. \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; UNEP-Nairobi Convention and WIOMSA 2021). Much of area designation has occurred since the 2003 IUCN World Parks Congress in Durban where the Malagasy government committed to increase the nation\u0026rsquo;s protected areas (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). For example, since 2003, 20 national MPAs have been created along with \u0026gt;\u0026thinsp;200 Locally Managed Marine Areas (LMMAs) (UNEP-Nairobi Convention and WIOMSA 2021). These areas are a mixture of state, co-managed, and private control. Many areas rely on local management known as \u003cem\u003edina\u003c/em\u003e where local committees can set rules and manage resources in collaboration with other authorities (Parker et al. \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Zafimahatradraibe et al. 2024). In many cases, the biodiversity of these locations has not been reported or monitored.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eMayotte and Comoros are under the authority of different national governments since 1974, and Mayotte became a French territory in March 2011. Comoros has one established Marine Park or Moheli gazetted in 2015 with strong island government community support (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb) (UNEP-Nairobi Convention and WIOMSA 2021). Three additional parks are in the process of being established, namely the Coelacanth MNP, the Mitsamiouli-Ndroud\u0026eacute; MNP, and the Shisiwani MNP. The entire lagoon of Mayotte and its EEZ were classified by the French government in 2010 as a MNP covering 68,800 km\u003csup\u003e2\u003c/sup\u003e but only 1% was classified as highly protected. Nevertheless, it is one of the largest marine protected areas in the Indian Ocean. The 2021 UNEP report also identifies the small (0.6 km\u003csup\u003e2\u003c/sup\u003e) Nature Reserve of M\u0026rsquo;bouzi established in 2007 prior to the larger designation. In contrast, the WDPA identifies 16 protected areas of which some are small coastal stretches and small offshore islands that may not manage resources beyond the shoreline.\u003c/p\u003e\u003c/div\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003eOverview\u003c/h2\u003e\u003cp\u003eSeveral scientific modelling and software advances have made it increasingly possible to predict finer scale patterns and thereby map marine biodiversity based on environmental proxies (Pilowsky et al. \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). These include: 1) moderate resolution mapping of habitats, such as coral reefs, at large scales, 2) large-scale underwater data collection and collaboration on important biodiversity proxy metrics (i.e. coral cover and taxa composition), 3) global satellite coverage of environmental variables that are proxies for influences on biodiversity, 4) statistical machine learning algorithms that can handle large amounts of complex data to make predictive models, and 5) sea surface water temperature predictions that are available from IPCC data time series (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.ipcc.ch/data/\u003c/span\u003e\u003cspan address=\"https://www.ipcc.ch/data/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e). The convergence of these tools provides the possibility of developing models to predict biodiversity on modest scales and to test different predictions for different scenarios. Environmental information at scales of satellite and shipboard observations can potentially reduce or correct biases created by incomplete, sparse, and anecdotal information. The Ethics and Consent to Participate declarations are not applicable to this study.\u003c/p\u003e\u003cp\u003eThe research described below was undertaken to better understand the finer-scale variability in taxonomic richness in the coral reefs of Madagascar, Comoros, and Mayotte. Specifically, coral and fish richness, and a proxy of total species were mapped on the 6 km\u003csup\u003e2\u003c/sup\u003e scale where most cells lacked field census data. A provincial model that predicted the number of species based on ~\u0026thinsp;2000 coral and fish census used empirical relationships with many environmental, demographic, and management variables (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) (McClanahan et al. \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). This was carried out with the machine learning algorithm described below. Empirical relationships with the environmental variables derived from the algorithms were used to predict numbers of taxa in all 3361 6-km\u003csup\u003e2\u003c/sup\u003e mapped coral reef cells of the study region. Because the African continent has higher biodiversity than Madagascar (McClanahan \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Samoilys et al. \u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), faunal province will differ from national biodiversity priorities. Therefore, to find priorities specific to Madagascar, the two ecoregions and 3 governance authorities were evaluated here separate from the western Indian Ocean province results. Additionally, temperature variables available from the IPCC temperature time series (RCP2.6 and RCP8.5) were used to predict coral cover and number of coral taxa in 2050 based on 2020 temperature-taxa associations.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eBiodiversity mapping framework\u003c/h3\u003e\n\u003cp\u003eThe United Nations Environmental Program (UNEP) and International Union for the Conservation of Nature (IUCN) have compiled protected areas in various stages of planning and establishment. The World Database of Protected Area (WDPA) is available from the World Conservation Monitoring Center (WCMC) (protectedplanet.net). Shape files were downloaded and overlayed with the modelled high biodiversity priority areas. This allowed us to determine the current designations that corresponded to the modeled priorities. This database lists 60 protected areas for Madagascar, 26 for Mayotte, and 16 for Comoros. WDPA numbers are larger than the recent provincial compilation of the UNEP-Nairobi Convention report, as different criteria were used for inclusion. Many protected areas have been designated but the authority to manage them and their success is often unreported. Many may also be in a proposal or planning process and not the post-implementation stage.\u003c/p\u003e\n\u003ch3\u003eEnvironmental layers data\u003c/h3\u003e\n\u003cp\u003eEnvironmental data compilations used several sources, which resulted in 70 spatially complete variables derived from a combination of satellite and shipboard measurements (see the list of data sources in McClanahan et al. \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Environmental data were a mixture of oceanographic data, such as photosynthetic active radiation (PAR), pH, calcite, dissolved oxygen, diffusion attenuation, salinity, net ocean primary productivity, chlorophyll-a variables, phytoplankton carbon, and wave height (Tyberghein et al. \u003cspan citationid=\"CR58\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Yeager et al. \u003cspan citationid=\"CR63\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Several water temperature or thermal stress metrics were calculated including sea surface temperature (SST) mean, median, range, standard deviation, skewness, kurtosis, rate of rise, and cumulative excess heat or degree-heating weeks (DHW). Several composite thermal and water quality stress metrics were included, such as the Global Stress Model, an indicator of thermal inputs (Maina et al. \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2011\u003c/span\u003e), and a composite nutrient concentrations model (Andrello et al. \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Finally, estimates of reef connectivity calculations were used to estimate potential larval flow including measures of connectivity, net flow, indegree, outdegree, and retention for each cell in this region (Fontoura et al. \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eGeographic variables included wilderness (\u0026gt;\u0026thinsp;4 hours travel time from human population), travel distance to people, shore, and ports, and market gravity or the number of people living on the shore or cities as divided by the square of the distance or travel time (Maire et al. \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). Cells were assigned to four fisheries management categories including unrestricted fishing, restricted fishing, low compliance closures, and high compliance closures. These classifications were based on information in published literature, the experience of the observers, and discussions with knowledgeable observers (McClanahan et al. \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Fish census observers also recorded the depth and habitats of the sites as reef edge, reef crest, reef flat, or reef lagoon. Detailed methods and model results have been presented elsewhere and here the focus is on predictions of biodiversity and management implications for Madagascar (McClanahan et al. \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003eField data collection\u003c/h2\u003e\u003cp\u003eThe model uses the above environmental data to make predictions of coral taxa for 2020 based on environment empirical field data associations revealed by Gradient Boosting Model (GBM) or specifically the Boosted Regression Tree (BRT) software. Field data were collected in all three countries from several field trips. The Western Indian Data were collected between 1995 and 2020 and environmental data prior to the sampling was used in the BRT modelling process. The proxy for total number of coral reef taxa used a combination of coral and fish taxa sampled by the following methods.\u003c/p\u003e\u003c/div\u003e\n\u003ch3\u003eCoral sampling\u003c/h3\u003e\n\u003cp\u003eCorals were visually sampled in haphazardly placed quadrats of ~\u0026thinsp;2 m\u003csup\u003e2\u003c/sup\u003e where all corals\u0026thinsp;\u0026gt;\u0026thinsp;5-cm were identified and counted in ~\u0026thinsp;15\u0026ndash;20 replicates (McClanahan et al. \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2007\u003c/span\u003e). Thus, the values used here were the total number of taxa in ~\u0026thinsp;40 m\u003csup\u003e2\u003c/sup\u003e. Taxa identification was to the genus level, but \u003cem\u003ePorites\u003c/em\u003e colonies were identified further as massive, branching, or \u003cem\u003ePorites rus\u003c/em\u003e and \u003cem\u003eGalaxea\u003c/em\u003e as either \u003cem\u003eG. astreata\u003c/em\u003e or \u003cem\u003eG. fascicularis\u003c/em\u003e. 1001 well distributed sites were sampled in the region (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The two observers with the most samples (N.A. Muthiga and T. McClanahan) were compared and found to have no significant differences (McClanahan et al. \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2024\u003c/span\u003e).\u003c/p\u003e\n\u003ch3\u003eFish sampling\u003c/h3\u003e\n\u003cp\u003eTwo experienced observers (T. McClanahan and J. Wickel) counted fish in designated areas or belt transects of 500-m\u003csup\u003e2\u003c/sup\u003e (McClanahan 1994). Replicates undertaken close to each other were pooled or averaged dependent on the methods such that the final units were number of species per ~\u0026thinsp;500 m\u003csup\u003e2\u003c/sup\u003e. The number of species for the 6 selected families known to be good proxies for total fish diversity (Acanthuridae, Chaetodontidae, Labridae, Pomacanthidae, Pomacentridae, and Scaridae) were extracted (Allen and Werner \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2002\u003c/span\u003e). Each observer also estimated biomass in their transects as the sum weights of the individual species or families based on length estimates and known length-weight relationships. A total of 1201 transects were sampled throughout most of the nations and two ecoregions of the Madagascar (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e).\u003c/p\u003e\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\u003ch2\u003eModel spatial and temporal predictions\u003c/h2\u003e\u003cp\u003eStatistical machine learning algorithms or GBM are increasingly being used to make predictions with large and complex data sets. Specifically, the BRT algorithm was the specific GBM used here. BRT is a commonly used algorithm for evaluating complex environmental-ecological data and shown to be a top performer among machine learning options (Elith et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2008\u003c/span\u003e; Kuhn and Johnson \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Prior analyses indicate that BRT models are preferred because they are effective at handling nonlinear relationships, missing values in covariates, interactions between predictors, and have a high predictive performance (Kuhn and Johnson \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Once the taxa-environment relationships were established from the empirical census data, the model was used to predict biodiversity using the environmental data in all 3361 mapped cells for 2020 and for the projected 2050 temperature conditions.\u003c/p\u003e\u003cp\u003ePredictions included the number of fish and coral taxa for the sampled areas, but the normalized average was used as a proxy for the total number of taxa in a cell and therefore the spatial cell\u0026rsquo;s proxy for biodiversity. Model predictions were tested for efficacy using a 70\u0026thinsp;\u0026minus;\u0026thinsp;30% training and testing procedure to determine the fits, which were R\u003csup\u003e2\u003c/sup\u003e of ~\u0026thinsp;80% for the full data and ~\u0026thinsp;45% for the training and testing data. Models require keeping some variables constant between spatial cells to make comparable predictions. For example, numbers of fish species are strongly correlated with fish biomass and change with water depth. Therefore, to make fair between-cell comparisons, biomass was held constant at 600 kg/ha and depth at 10 meters (McClanahan et al. \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Therefore, the maps are based on partial effects where local depth and fish biomass were held constant and the predictions are for these constants but where other environmental variables are from the various databases.\u003c/p\u003e\u003cp\u003eThe future state model selected those temperature variables common to the above model and RCP8.5 Business-as-Usual and RCP2.6 or the carbon emission reduction scenarios. These variables were used to make biodiversity predictions for the current scenario values in 2020 and future predictions in 2050. Specifically, the first BRT selection process selected 6 variables shared IPCC temperature data. These variables were the mean SST (CMIP does not give median SSTs), skewness, kurtosis, bimodality, and cumulative excess heat (degree-heating weeks\u0026thinsp;=\u0026thinsp;DHW). Therefore, no future predicted human demographic, or environmental variables were included in future forecasts, so the predictions are largely based on future temperatures and 2020 associations with other environmental and demographic data. Coral cover was taken from a previously machine learning model for 2020 and 2050 (McClanahan and Azali \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Data used in scatterplots were the changes in coral cover and number of taxa for all spatial cells for the years 2020 and 2050. Scatterplots present the differences in the values between the years (2050\u0026ndash;2020) as a measure of change or resilience for the 3361 mapped cells. It should be appreciated that the sea surface temperature predictions for 2050 have reached the +\u0026thinsp;1.5\u003csup\u003e0\u003c/sup\u003eC above baseline by 2025, or 25 years earlier than predicted by the IPCC data time series (Hansen et al. \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e\u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003eBiodiversity maps for coral, fish, and the total diversity proxy are presented in the following figures (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003ePredictions for coral\u003c/h2\u003e\u003cp\u003eModel predictions for numbers of coral taxa indicate both broad and fine-scale patterns (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea). Predicted numbers of coral taxa in Madagascar were overall highest in the north and northwest of the Antsiranana and Mahajanga Provinces. This high diversity extends west to include Mayotte and Comoros. The Ankarea MNP (Mitsio Islands) was in the center of this biodiversity location. Diversity was predicted to be highest in the western or leeward side of Madagascar in the northern Antsiranana Province. Numbers of coral taxa was predicted to decline to the south into Mahajanga and Toliary Provinces. The border region between Antsiranana and Mahajanga Provinces in the Ambanja District had a cluster of lower diversity, but numbers increased to the south until midway down the Mahajanga province. The Ankivonjy MNP is located just north of the high and low predicted transition of coral diversity or just north of the Mahajanga provincial border. The Toliara Province in the southwest was predicted to have lower numbers of corals overall but had a high diversity location south of Velondriake and near the Soarkiake MNPs. Nevertheless, some high diversity reefs may not be contained in these parks.\u003c/p\u003e\u003cp\u003eIn eastern Madagascar and Antsiranana, the Masoala MNP in the Antalaha District was predicted to have lower coral diversity than the Ambodivahibe MNP in the northeast. Further south in the eastern Tomasina Province, high diversity was predicted from the northern border south to St. Anne Island. High diversity was predicted for the Nosy-Boraha and Soanierana Ivongo Districts. The southern part of Toliara and all Fianarantsoa Provinces were predicted to have fewer reefs and a low diversity of corals. Predictions for Comoros and Mayotte indicated high numbers of coral taxa throughout these islands. Calculations of the predictions of the average numbers of taxa suggest Mayotte had slightly higher numbers of coral taxa (25.5\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;2.0 (SD)) taxa per 40 m\u003csup\u003e2\u003c/sup\u003e than the Comoros Islands (23.4\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;1.7).\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003ePredictions for fish\u003c/h2\u003e\u003cp\u003ePredictions of numbers of fishes in Madagascar indicate some differences compared to coral diversity predictions (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb). The number of fish was highest in Antsiranana Province but with some clear onshore-offshore patterns or higher diversity in offshore reefs. Again, the offshore Ankarea MNP was the predicted center of this fish diversity. Fish diversity was high and less variable than corals in Ambanja District but declined further to the south and only high again in the Barren Islands of Mahajanga Province. Fish diversity in the northeast was predicted to be higher than in corals. Locations of high diversity were predicted for the entire Masoala Peninsula in the Antalaha District. Diversity predictions declined in the Antoginil Bay of the Maroantsetra District, due to poor water quality conditions. However, south of Nosy-Boraha and Soanierana Ivongo there was also a high predicted fish diversity that extended south to the Toamasina District.\u003c/p\u003e\u003cp\u003eThe southwest Toliara Province had low predicted fish diversity overall but had high predicted diversity in the north of the Province around the Velondriake and near the Soarkiake MNPs. The province had the same onshore-offshore diversity patterns observed elsewhere. The southern part of Toliara and all of the Fianarantsoa Provinces were predicted to have few reefs and a low diversity of fish except for an offshore island in Fianarantsoa Province.\u003c/p\u003e\u003cp\u003ePredictions for numbers of fish species in Comoros and Mayotte indicated more spatial variability than corals, particularly for the onshore-offshore gradients. For example, high numbers of species were predicted on the outer reef of Mayotte, but numbers declined rapidly shoreward. Predictions for the west and south sides of Mayotte were to have more frequent species than the island's east and north sides. In Comoros, Mwali (Moheli) was predicted to have the most fish species, followed by Ngazida (Grande Comoros) and high spatial variability in species for Ndzuwani (Anjouan) island. The proposed Mitsamiouli Ndroudeb MNP in the north of Ngazida island was predicted to have high numbers of fish species. The proposed Coelacanth MNP on the southern tip of Ngazida has reefs only on the far southern end, which was also a location predicted to have high numbers of fish. Predictions for a mixture of moderate and high numbers of fish was made for locations within the proposed Shisiwani MNP located on the northwest of the Ndzuwani island. Destructive fishing is reported to be common in Ngazida and Ndzuwani and actual are expected to be lower than predicted numbers of species.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003ePredictions for total numbers of species\u003c/h2\u003e\u003cp\u003ePredictions using the biodiversity proxy for the total number of marine taxa indicate similar patterns with both coral and fish (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec). Notably, there was high predicted total diversity from Ankarea west to Mayotte and Comoros. Other patterns align with high diversity predictions offshore in the northwest. The nearshore locations around Masoala and southern reefs extending into the northern Toamasina Districts were predicted to have high total diversity. Similarly, high diversity was predicted for the Velondriake and Soarkiake MNPs in northern Toliara and Morombe Districts.\u003c/p\u003e\u003cp\u003eComoros and Mayotte reefs were predicted to have high numbers of marine taxa. Mwali had the highest number of taxa followed by Ngazida. Predicted taxa in Mayotte were high on the outer reef and declined shoreward. Predictions for Ndzuwani were similar but patterns were more variable along the coast.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\u003ch2\u003eComparisons with past reports and protected area locations\u003c/h2\u003e\u003cp\u003eSummarizing the findings of the past reports and the number of criteria used in the biodiversity evaluations (3 of taxa and 4 of spatial scales) indicated that there were no locations in Madagascar that fit all 12 WIO provincial criteria (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e; Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Therefore, by the model criteria, all selected areas were national and not WIO provincial priorities. Two past reports identified Grand Recif and Nosy Ve in Toliara Province and Helodrano in Antoginal Bay in Tomasina County. None of these were among the model\u0026rsquo;s high biodiversity selections. A large area that included Masoala was identified by the World Heritage report, and the windward or eastern ocean-exposed peninsula was identified by the model\u0026rsquo;s criteria. Masoala peninsula ranked high for the fish and total diversity metrics, but not by the coral criteria. The model and criteria selected new sites in the coastal reefs south of Antongil Bay or Ivontaka-Antanambe villages and those south to St Anne Island in Tomasina.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eSelection of priority diversity locations selected by the biodiversity machine learning model predictions for Madagascar, Mayotte, and Comoros. Marine protected areas listed were determined from the joint project of IUCN and UNEPs \u0026ndash; World Conservation Monitoring Center (WCMC) compilation or World Database of Protected Area (WDPA). Criteria are based on a maximum of 12 possible criteria described in McClanahan et al. (\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2024\u003c/span\u003ea) and as the number of past reports where they are listed as priorities. The 12 criteria are top selections for 3 measures of biodiversity (coral, fish, and their combination) and four spatial scales (nation, ecoregion, province, and reef clustering).\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"8\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cem\u003eEcoregion name\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u003cem\u003eCountry\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cem\u003eName\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u003cem\u003eWorld Heritage Marine Sites of Outstanding value\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u003cem\u003eSWIOFP Biodiversity hotspots\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003e\u003cem\u003eSum Criteria (McClanahan et ala. 2024a)\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u003cem\u003eSum Criteria (Regional Reports)\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u003cem\u003eMarine Protected area\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAntisaranana - Masoala\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eAntongil bay, Northeast Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eAntongil bay\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e2\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e2\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eMasoala\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eNosy Mitsio\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eNorth and Northeast Madagascar, Ambodivahibe - Sahamalaza\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e8\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eAnkarea\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eBaie Lotsaina - Ankazomalemy\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eNorth and Northeast Madagascar, Ambodivahibe - Sahamalaza\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e6\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eNosy Hara\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eNosy Sakatia/Nosy Be\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eNorth and Northeast Madagascar, Ambodivahibe - Sahamalaza\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e6\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eLokobe\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAnkazoberavina - Baie Androfiabe\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eNorth and Northeast Madagascar, Ambodivahibe - Sahamalaza\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e6\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eNosy Antsoha,\u003c/p\u003e\u003cp\u003eAmpasindava,\u003c/p\u003e\u003cp\u003eAnkivonjy\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eNosy Faly\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eNorth and Northeast Madagascar, Ambodivahibe - Sahamalaza\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e2\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eNosy Ambariovato\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eNorth and Northeast Madagascar, Ambodivahibe - Sahamalaza\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSoutheast Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eSoutheast Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eSouthern Madagascar (the deep south)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAnjiambe - Ampisikanana\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e6\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eAmbodivahibe,\u003c/p\u003e\u003cp\u003eAnalamerana,\u003c/p\u003e\u003cp\u003eLoky manambato, Oronjia,\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAmbohotrabo - Tsiadamaba\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e4\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eMangoky Ihotry wetland complex,\u003c/p\u003e\u003cp\u003eVelondriake,\u003c/p\u003e\u003cp\u003eManjaboaka, Soriake\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eVohemar bay\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e4\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eNosy Boraha - Toamasina\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e3\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSoutheast Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eNosy Faho\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e3\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eMaintirano\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e2\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAmbatonjanahary - Anjiabe\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e2\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eIvontaka - Anatanambe\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e2\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eMananara nord, Seranambe,\u003c/p\u003e\u003cp\u003eVohitralanana, Ambodimangamaro\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAnalalava - Mahabo\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSoutheast Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eNosy Dombala - Nosy Fonga\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSoutheast Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAmpanotoamaizina\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSoutheast Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eVohimasina\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSoutheast Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eManakara\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSoutheast Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eGrande recif\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eGrande Recif (11)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSoutheast Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eNosy Ve\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eNosy Ve\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e0\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMayotte\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eLongoni bay - Tsingoni\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eComoros - Glorieuses crescent\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e8\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eIlots M'Tzamboro,\u003c/p\u003e\u003cp\u003eMayotte,\u003c/p\u003e\u003cp\u003ePointes Et Ilots Du Nord,\u003c/p\u003e\u003cp\u003eBaie De Dzoumogne-Longoni\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMayotte\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eTsingoni - Baie de Kani\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eComoros - Glorieuses crescent\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e7\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eMayotte,\u003c/p\u003e\u003cp\u003eLittoral De Sada-Chiconi,\u003c/p\u003e\u003cp\u003eLittoral De Kani-Keli,\u003c/p\u003e\u003cp\u003eN'Gouja\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMayotte\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003ePamanzi bay - Baie de Kani\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eComoros - Glorieuses crescent\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e4\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eIlots De La Passe,\u003c/p\u003e\u003cp\u003eIlots De Dembeni,\u003c/p\u003e\u003cp\u003eIlots De Bandrele,\u003c/p\u003e\u003cp\u003eMayotte,\u003c/p\u003e\u003cp\u003eLa Vasiere des Badamiers,\u003c/p\u003e\u003cp\u003eLittoral De Mamoudzou,\u003c/p\u003e\u003cp\u003eLittoral De Bandrele,\u003c/p\u003e\u003cp\u003eIlots Mbouzi,\u003c/p\u003e\u003cp\u003ePointes Et Plages De Saziley Et Charifou,\u003c/p\u003e\u003cp\u003eCratere De Petite Terre,\u003c/p\u003e\u003cp\u003eLittoral De Dembeni,\u003c/p\u003e\u003cp\u003eVasiere Des Badamiers\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eComoros, Mwali Island\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eFomboni\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eComoros - Glorieuses crescent\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eMoheli (1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e10\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e2\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eMoheli - Zone de transition,\u003c/p\u003e\u003cp\u003eParc National de Moheli\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eComoros, Mwali Island\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eNioumachoa\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eComoros - Glorieuses crescent\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003eMoheli (1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e10\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e2\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eMoheli - Reserve Mea,\u003c/p\u003e\u003cp\u003eMoheli - Reserve Magnougni et dzaha 1,\u003c/p\u003e\u003cp\u003eMoheli - Reserve Magnougni et dzaha 2,\u003c/p\u003e\u003cp\u003eMoheli \u0026ndash; Reserve Nioumachioua,\u003c/p\u003e\u003cp\u003eMoheli - Reserve Ouenefou,\u003c/p\u003e\u003cp\u003eParc National de Moheli\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eComoros, Ngazida Island\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eMoroni- Singani\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eComoros - Glorieuses crescent\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e12\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eParc National Coelacanthe\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eComoros, Ngazida Island\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eNiamaoui - Chomoni\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eComoros - Glorieuses crescent\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e12\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eParc National Mitsamiouli Ndroude\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eComoros, Ngazida Island\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eDimani - Mohoro\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eComoros - Glorieuses crescent\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e10\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eComoros, Ndzouani Island\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eDjomani - Domoni\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eComoros - Glorieuses crescent\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e6\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eComoros, Ndzouani Island\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAnjouan SW\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eComoros - Glorieuses crescent\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e4\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003eParc National Shisiwani\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWestern and Northern Madagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eComoros, Ndzouani Island\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAnjouan N, NE\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eComoros - Glorieuses crescent\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u003cb\u003e2\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eOf these 60 WDPA in Madagascar, 8 specific locations overlapped with the model\u0026rsquo;s 21 larger biodiversity hotspot locations. These WDPA specific names of the protected area are Masoala, Ankarea, Nosy Hara, Lokobe, Nosy Antshoa, Ampasindaya, Ankivonjy, Ambodivahibe, Analmerana, Lokymanambato, Oronjia, Mangoky, Ihbotry wetland complex, Velondriake, Manjaboaka, Soriake, Mananara nord, Seranambe, Vohitralanana, and Ambodimangamoro.\u003c/p\u003e\u003cp\u003eDesignations are complicated in that several smaller protected areas are contained in larger protected areas. For example, all WDPA protected areas in Mayotte are within the larger national protected area. Therefore, the 3 hotspot locations identified from the model, namely Longoni bay \u0026ndash; Tsingoni, Tsingoni - Baie de Kani, and Pamanzi bay - Baie de Kani, are all contained in either the larger park or smaller protected areas. In Comoros, most of the coastal areas of the island of Mwali are in the Moheli park but this park contains several reserves. Four of these reserves are contained in the modelled biodiversity hotspot location named Nioumachoa. On Ngazida Island, the Coelacanth and Mitsamiouli Ndroude national parks are contained in the selected Moroni-Singani and Niamaoui\u0026thinsp;=\u0026thinsp;Chimoni hotspots, respectively. On the island of Ndzouani, the Shisiwani MNP is contained in the Anjouan southwest selected hotspot. Therefore, 3 of the 8 selected hotspots are not included among WDPA park designations. These are the Dimaini-Mohoro on Ngazida Island and the Diomani-Domoni and Ajouan northeast hotspots on Ndzouani. Therefore, of the 102 WDPA protected areas in these 3 countries, 44 were overlapping with half (16) of the models selected 32 biodiversity hotspots.\u003c/p\u003e\u003cp\u003eIn northern Antsiranana Province there was a large area selected by the criteria and the World Heritage report. Many of these reefs were included in the marine parks of Ambodivahibe, Nosy Hara, Ankarea, and Ankivonjy. Specifically, these included the biodiversity selected locations of Vohemar, Anjabe Ampisikanana, Nosy Mitsio, Nosy Sakatia, Nosy Ambariovato, Baie Androfiabe, and Analalava Mahabo. The smaller area of Analalave Mahabo in northern Mahajanga Province was selected for its predicted high numbers of coral taxa. However, numbers of fish species were predicted to be low, and so was the total diversity. Vohemar was selected for high numbers of fish taxa but predicted to have modest numbers of coral taxa.\u003c/p\u003e\u003cp\u003eThere were also a few offshore islands in the Toliara and southeastern Fianarantsoa Provinces that were selected for their predicted high numbers of fish species. These include the offshore coral reef islands of Nosy Dombala, Nosy Fonga, and Ampanotoamaisina south of Toamasina. Some coastal rocky reefs of Vohimasina and near Manakara in the south were also identified by the same criteria. The western coast included a small coral reef island\u0026thinsp;~\u0026thinsp;15 km offshore from Maintirano contained in a RAMSAR designated area. Further south are the islands and the coastal fringing reefs from Andavakoaka south to Tsifota on the coast from the Ifaty forest. Many of these locations are included in Velondriake and Soarkiake MNPs.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec17\" class=\"Section2\"\u003e\u003ch2\u003eClimate change predictions\u003c/h2\u003e\u003cp\u003ePlots of the model\u0026rsquo;s predicted changes in coral community cover and number of taxa from 2020 to 2050 indicate widespread losses in cover and taxa (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e; Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). In the Business-as-Usual scenario (RCP 8.5) average coral cover was predicted to decline in Madagascar from 34.6\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;11.1% (SD) in 2020 to 24.6\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;9.9% in 2050. The number of taxa was also predicted to decline from 16.5\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;1.9 to 15.6\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;1.5 taxa per 40 m\u003csup\u003e2\u003c/sup\u003e. Therefore, most predictions lay in the bottom left quadrat where there were losses in coral cover and taxa. In fact, 72% of the reefs were predicted to have losses in both cover and taxa, while 7.3% were predicted to have gains in both coral and taxa. Mapped cells with gains were distributed among Mahajanga, Toliary, Toamasina and Antsiranana provinces.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eResults of impacts on coral cover and number of taxa for two climate change scenarios using CMIP5.0 variables or (a) business as usual and (b) carbon emission reductions. Predicted coral cover (%) and number of taxa (per ~\u0026thinsp;40m\u003csup\u003e2\u003c/sup\u003e) in 2020 and 2050. The number of coral reef cells predicted to have the 4 combinations of gains and losses (2050\u0026thinsp;\u0026minus;\u0026thinsp;2020) of coral cover and number of taxa.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"9\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCountry\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003e2020 Coral cover, %\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2050 Coral cover, %\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003e2020 Number of taxa\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003e2050 Number of taxa\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eCover loss/ taxa loss, n cells (%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003eCover loss/ taxa gain, n cells (%\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c8\"\u003e\u003cp\u003eCover gain/ taxa loss, n cells (%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c9\"\u003e\u003cp\u003eCover gain/ taxa gain, n cells (%)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"9\" nameend=\"c9\" namest=\"c1\"\u003e\u003cp\u003eBusiness-as-usual scenario RCP8.5\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e34.6 (11.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e24.6 (9.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e16.5 (1.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e15.6 (1.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e1643 (72)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e183 (8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e289 (12.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u003cp\u003e167 (7.3)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMayotte\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e47.1 (3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e44.8 (4.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e25.5 (2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e23.1 (1.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e214 (79.6)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e55 (20.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eComoros\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e31.5 (5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e29.5 (5.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e23.4 (1.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e22.6 (1.6)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e236 (99.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e2 (0.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colspan=\"9\" nameend=\"c9\" namest=\"c1\"\u003e\u003cp\u003eCarbon emission reduction scenario RCP2.6\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMadagascar\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e34.6 (11.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e31.4 (16)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e16.5 (1.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e15.7 (1.6)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e1259 (55.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e48 (2.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e656 (28.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u003cp\u003e319 (14)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMayotte\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e47.1 (3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e56.7 (9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e25.5 (2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e24.7 (1.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e17 (6.3)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e212 (78.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u003cp\u003e40 (14.9)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eComoros\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e31.5 (5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e20.7 (4.6)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e23.4 (1.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e22.3 (1.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e226 (95)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e12 (5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eIn the carbon emission reduction scenario (RCP2.6) average coral cover in Madagascar was predicted to decline less but more variably from 34.6\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;11.1% (SD) in 2020 to 31.4\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;16.0% in 2050. The number of taxa was predicted to be smaller and more variable from 16.5\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;1.9 to 15.7\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;1.6 taxa per 40 m\u003csup\u003e2\u003c/sup\u003e. Still, predictions for most reefs fell in the bottom left quadrat but with fewer losses in coral and taxa than the Business-as-Usual scenario. For example, 55.2% of the reefs were predicted to lose both cover and taxa, while 14.0% predicted to experience gains in both cover and taxa. Reef cells are predicted to have gains in cover, but losses in species increased from 12.7% to 28.7% when comparing the two climate change scenarios. The carbon emission scenario RCP2.6 was predicted to have better outcomes for Mayotte, Toliary, and Mahajanga reefs.\u003c/p\u003e\u003cp\u003eMayotte reefs were predicted to have among the highest coral cover in 2020 at 47.1\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;3.0% and to be reduced to 44.8\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;4.2% in 2050 in the Business-as-Usual RCP8.5 scenario. Mayotte was also predicted to have a higher number of taxa than Madagascar at 25.5\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;2.0 taxa per 40 m\u003csup\u003e2\u003c/sup\u003e and predicted to decline to 23.1\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;1.4 in 2050 if carbon emissions are not reduced. If carbon emissions are reduced, the model predicts an increase in coral cover above the 2020 baseline in 2050 to 56.7\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;9.0%. This will also produce a smaller loss of coral taxa to 24.7\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;1.8 per 40 m\u003csup\u003e2\u003c/sup\u003e relative to the 25.5\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;2.0 2020 baseline.\u003c/p\u003e\u003cp\u003eComoros reefs were predicted to have the lowest 2020 baseline coral cover of 31.5\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;5.0% and to be reduced to 29.5\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;5.5% in 2050 in the Business-as-Usual and 20.7\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;4.6 in the carbon reduction scenario. Thus, the prediction is for a greater decline in cover with reduced emissions, which is the opposite of Mayotte. Comoros has a predicted intermediate number of coral taxa in 2020 at 23.4\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;1.7 and small predicted to decline to 22.6\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;1.6 per 40 m\u003csup\u003e2\u003c/sup\u003e in 2050 in the Business-as-Usual scenario. If carbon emissions are reduced, the predicted numbers of taxa in 2050 were unchanged in 2020 at 22.3\u0026thinsp;\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e\u0026plusmn;\u003c/span\u003e\u0026thinsp;1.9 taxa per 40 m\u003csup\u003e2\u003c/sup\u003e. Predictions are based on ocean temperature conditions and did not include other important human influences such as fishing and watershed erosion.\u003c/p\u003e\u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThe methods used here were able to map biodiversity on a large scale and identify several new conservation or protected area priorities for Madagascar. Moreover, these were identified by an objective measure based on multiple environmental variables, and at a broader and finer spatial scale than past efforts. Past efforts used presence-absence data (Allnut et al. 2012) or prioritized at the larger WIO faunal province delineation (McClanahan et al. \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The higher diversity on the African continent makes it important to independently establish national level priorities in Madagascar. Additionally, the 6.25 km\u003csup\u003e2\u003c/sup\u003e scale of analysis provided a much finer scale for decision making that may be more realistic in scaling the costs and efficacy of management. Notably, many recent conservation initiatives are small-scale community managed areas (Gardner et al. \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). The outcomes for the reef coral and fish faunae environmental relationships showed a mixture of shared and divergent environmental influences on the taxa. For example, water temperature variability affected both groups, but fishes were more influenced by connectivity and larval retention, productivity, and fish biomass than corals. In contrast, physicochemical factors, such as waves, currents, dissolved oxygen, salinity, and chronic and acute temperature stresses most affected corals.\u003c/p\u003e\u003cp\u003ePast prioritization decisions represent the historical focus on planning large-scale protected areas (Wells et al. \u003cspan citationid=\"CR60\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). This approach has, however, been augmented in Madagascar with the development of smaller- or community-scale proposals that now cover large areas of Madagascar (Rocliffe et al. \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). These areas include many smaller LMMAs associated with many tropical habitats (UNEP-Nairobi Convention and WIOMSA 2021). The outcomes of both the 20 nationally gazetted protected areas and the \u0026gt;\u0026thinsp;200 LMMA for protecting biodiversity were a primary concern for assessing their relationship to underlying biodiversity patterns and potential for conservation impact. In many cases, high resource dependency and low funding for conservation has often undermined the effectiveness of larger nationally protected areas, where outcomes seldom differ from effective national fisheries restrictions management (McClanahan et al. \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Randrianarivo et al. \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Smaller community managed areas may have similar problems associated with their small size relative to the resource and areas needed to support many of the captured animals (McClanahan and Graham \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2015\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eGovernment partnerships with NGOs and the private sector may help to overcome some of these historical problems. Community management has been in place for sufficient time to provoke several lessons (Gardner et al. \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Common problems include the difficulties of promoting participation and good governance, applying rules, resolving conflicts with outsiders, influencing fish prices such that benefits exceed fishing costs, long-term commitments to monitoring human use and key resources, and sustainable funding (Parker et al. \u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Zafimahatradraibe et al. \u003cspan citationid=\"CR64\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Some important solutions may include the importance of co-management rather than community management, outside support to negotiate conflicts among neighboring communities, a focus on managing natural resources rather than the more common focus on distributing and selling captured resources, increasing the production of fish to reduce poverty and social disparity, collaborative decision making, diversification and entrepreneurial sourcing of income, and monitoring feedbacks to resource users to improve adaptive management (McClanahan and Abunge \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Gardner et al. \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Lessons such as these provide some guidelines to consider among many policy and management options. Nevertheless, much remains to be learned about how to support biodiversity while reducing poverty over the long term.\u003c/p\u003e\u003cp\u003eThe finer spatial scales of the presented maps compared to past reports may be more useful for emerging smaller-scale governance and LMMA conservation priorities. Moreover, ecological services are among key current priorities and include fisheries production, shoreline protection, and local biodiversity conservation. Therefore, ecological functions provided by fishes and corals align well with the human services prioritization approach that drives LMMA management. Indeed, sustaining high and stable fish catches is a major social concern throughout Madagascar and Comoros (Le Manach et al. \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Cinner et al. \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; McClanahan and Jadot \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2017\u003c/span\u003e; Gough et al. \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Zeller et al. \u003cspan citationid=\"CR65\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Ranaivomanana et al. \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Fish resources have been reported to be in better condition in Mayotte, associated with greater wealth and lower local natural resource dependency (McClanahan and Jadot \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2017\u003c/span\u003e).\u003c/p\u003e\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e\u003ch2\u003eConservation priorities\u003c/h2\u003e\u003cp\u003eVariability in environmental associations with the two taxa resulted in variability in the selection of locations. For example, some locations were selected because of high predicted numbers of fish, such as those in eastern Madagascar including Anjabe, Ampisikanana, Vohemar, Ivantake, Masoala, Antanambe, Nosy Dombolo, Nosy Fonga, Ampanotoamizina, Vohimasina, and Manakara. Locations in the northwest were picked for high numbers of corals, such as Antalava-Mahabo. Most of the other western sites were selected based on both corals and fish or the proxy for total numbers of species. Locations included many of the areas recently included as national marine parks, specifically Ambodivahibe, Nosy Hara, Ankarea, Ankivonjy, Velondriake, and Soariake. The combined taxa or proxy approach used here provided thorough coverage that should make moderate to good predictions for where high numbers of species are located. Many of these locations are heavily influenced by other human watersheds and fishing impacts (Bruggemann et al. \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Maina et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2013\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eMany existing MNPs appeared to coincide with the high diversity predictions. Therefore, the machine learning model supports national decisions with a few exceptions. Among the exceptions are the Kirindy Mite MNP north or Velondriake MNP, which was predicted by the model to have low but variable diversity among its islands. Some notable omissions of the national park selections would be the reefs south of Antongil Bay or the Ivontaka-Antanambe and St Anne Island villages. These reefs are, however, an area of dense LMMAs that may provide some conservation benefits (UNEP-Nairobi Convention and WIOMSA 2021). Additionally, there were some offshore coral islands in the east and west predicted to have high diversity, namely Nosy Dombala, Nosy Fonga, and Ampanotoamaisina in the islands offshore in the east. In the west coast, islands from Maintirano to Barren Islands and Iles Eparses were also identified by the model and as a RAMSAR designation. Most notable is the high overlap of the model\u0026rsquo;s biodiversity predictions and MNPs on the west coast of Madagascar.\u003c/p\u003e\u003cp\u003eSome of the regions priority areas may have been ignored due to past efforts that focused on the WIO provincial rather than the national scale priorities. Scale of delineation can be quite important in terms of the numbers, areas, and distribution of selected sites (Grace et al. \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Madagascar does not fare well in provincial delineation selections for several reasons, including poor coverage of species data, the isolated island biogeographic affect, and the dispersed nature of the coral reefs. For example, the coarse scales used by the World Heritage Report included all of Mayotte and Comoros as a priority whereas the selection method identified 3 specific sites in Mayotte and 8 in Comoros. World Heritage selected two large areas at the northwestern and southern tips or Madagascar whereas criteria used here identified 21 more specific locations in Madagascar (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). SWIOFP identified 3 specific sites, but few of these selections were supported by this biodiversity analyses. SWIOFP used different criteria and therefore chose Antongil Bay for the hump-back whale breeding sanctuary, Nosy Ve for tropical breeding birds, and Grand Recif for coral diversity based on historical information that was based on extensive but spatially restricted sampling.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec20\" class=\"Section2\"\u003e\u003ch2\u003eManagement priorities\u003c/h2\u003e\u003cdiv id=\"Sec21\" class=\"Section3\"\u003e\u003ch2\u003eFish biomass\u003c/h2\u003e\u003cp\u003eDesignations such as MNR or LMMA do not necessarily provide evidence for protection or the status of the resources (Randrianarivo et al. \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Past work has shown that fish biomass is one of the single best indicators of resources status and biodiversity. Biomass is also expected to be manageable by fisheries regulations. An empirical field study of Madagascar and Mayotte reef fishes concluded that managing the biomass of fish is more important than site selection (McClanahan and Jadot \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). However, the model used here held biomass constant so that it was possible to visualize and compare sites for the same biomass (600 kg/ha) and depth (10 m). This produces spatial variability driven more by non-biomass factors, which exposes the underlying biodiversity patterns expected without the current high human fishing pressure. This approach produces more environmentally driven spatial variability and helped with the identification of several prioritization sites mentioned above. Nevertheless, managing biomass on large scales is likely to be one of the simplest and most effective targeted approaches to protect biodiversity while maintaining optimal fisheries yields. By focusing on this goal, it is possible to indirectly protect fish biodiversity, maximize yields, and manage other ecological processes (McClanahan \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2018\u003c/span\u003e, 2022).\u003c/p\u003e\u003cp\u003eFish biomass is managed through the usual restrictions on access, fishing gear, times, locations, and capture choices (McClanahan and Abunge \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). These preferences vary with fishers and communities, but people are more willing to agree to restrictions when local governance is effective. Some poorly known influential national attributes include fisheries governance, laws, compliance, and human dependence on fish resources. In general, managing fisheries for maximum sustained yields on large scales combined with closure areas approaching 30% of the nearshore is likely to achieve optimal management (Kerwath et al. \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; McClanahan \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2021\u003c/span\u003e, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2022a\u003c/span\u003e). An evaluation of fish biomass in the Northwest Madagascar Ecoregion found high variability but that 65% of the studied reefs had biomass greater than the recommended maximum sustained yield level of 600 kg/ha (McClanahan and Jadot \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). Most of the lower biomass levels were close to urban centers or in the southwest where dry conditions make fishing one of the few animal food options (Bruggemann et al. \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). Most reefs in Mayotte had biomass above this level but Comoros values are lower. Fish stock biomass is an important indicator of the success of tradeoffs between human needs and biodiversity conservation.\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec22\" class=\"Section2\"\u003e\u003ch2\u003eCoral diversity\u003c/h2\u003e\u003cp\u003eCoral biodiversity was primarily driven by environmental forces of temperature stresses, currents, and waves. These environmental forces will change as climate warms. Climate predictions suggest further degradation of the reef environment by 2050 for both climate scenarios. After accounting for fish biomass and depth, the strongest predictors of the total biodiversity proxy were median SST and metrics of variability (excess heat, kurtosis or chronic SST variability, rate of SST rise, bimodality, and skewness). Many of these variables are frequently among the key factors that influence coral bleaching and cover. For example, the standard climate-coral impact models frequently include cumulative excess heat, and the rate of SST rise as the main stresses (McClanahan \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2022b\u003c/span\u003e). However, studies in the WIO province have shown that SST variability metrics of kurtosis, bimodality, and skewness had equal or greater influences on both biodiversity and coral cover (McClanahan and Azali \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). This environmental complexity produces considerable spatial variability in the responses predicted for the two climate change scenarios.\u003c/p\u003e\u003cp\u003eBoth coral and fishes are affected by the historical patterns of acute and chronic stress and not just excess heat. In fact, modest amounts of excess heat promote corals acclimation to climate change (McClanahan and Azali \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Therefore, high environmental variability around Madagascar created the heterogeneous responses among the 2050 predictions. Heterogeneity is evident in the predicted losses of taxa, with both gains and losses of cover and biodiversity. Most reef cells were predicted by the Business-as-Usual scenario to have both losses of coral cover and taxa (55 to 72%). However, a smaller number of reefs may escape the consequences of climate change in the next 30 years if the carbon emission reductions are achieved and other human local disturbances reduced. The smaller group predicted to benefit from climate change was between 7 and 14% of reef cells. A modelling study of Madagascar provided evidence that deforestation of watersheds may be more detrimental to corals than climate change, even where rainfall is predicted to decline in some provinces (Maina et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). While southwest Madagascar may be a refuge when seen from climate impact perspective (Sully et al. \u003cspan citationid=\"CR57\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), the local fishing impacts are expected to undermine the climate refuge prediction (Bruggemann et al. \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2012\u003c/span\u003e; Ranaivomanana et al. \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e\u003cdiv id=\"Sec23\" class=\"Section3\"\u003e\u003ch2\u003eModel caveats\u003c/h2\u003e\u003cp\u003eSeveral technical advancements in conservation science were represented in the presented approach and outcomes (Pilowsky et al. \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Nevertheless, the availability of field data was core to the empirical predictions that allowed broader scale predictions. Fortunately, the collaborators shared methods that were comparable (McClanahan et al. \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2007\u003c/span\u003e). Moreover, the machine learning model accounted for observer and biomass effects when making predictions. In the case of fishes, the machine learning model held biomass constant and above the saturation level required to make comparable predictions of numbers of species between reef cells. The flexibility and ability to control for many factors is a key strength of the machine learning and partial effects approach.\u003c/p\u003e\u003cp\u003eTests of model performance using cross-validation procedures indicated good predictive ability. Yet, when making predictions for many cells on large scales, there is the possibility of overfitting and missing important local conditions, especially below the 6.25-km\u003csup\u003e2\u003c/sup\u003e scale. Moreover, the model cannot account for local unmodelled variables, such as damaging fishing methods, rare cyclones, or point-source pollution. Nevertheless, the model represents a considerable advance in marine spatial modelling in a country lacking broad-scale and comparable historical data. Challenges remain to test predictions and account for human and other local factors not currently available. The outcomes will, however, depend on the metrics and values of the assessments, such as ecological functions of taxa, their various values, uniqueness, evolutionary relatedness, and threats (Brooks et al. \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2006\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe methods used here did not focus on some important conservation concerns, such as remoteness, large body sizes, species with connections to land, and rarity. Nevertheless, hard coral is a well sampled invertebrate and a taxon with many other dependent species. When corals are combined with fish, the diversity proxy includes multiple taxon known for high species richness. The positive correlation found for these two groups indicated that these taxa should often correlate with a more diverse group of unsampled taxa, such as species associated with seagrass and mangroves (McClanahan et al. \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The modest fit for the coral-fish correlation suggests that the two faunal groups combined should provide a general but not necessarily accurate proxy for total richness. Past prioritization efforts often share selection choices or biases in terms of selected taxa and human impacts. Many of these selections were justifiable based on immediate threats and needs of rare, space-requiring, and sensitive species. However, the model used here made a first estimate of broad-scale total diversity at finer scales than previous efforts.\u003c/p\u003e\u003cp\u003ePast selections have disproportionally focused on large-bodied species with broad distributions and threats. Some other species, such as the southern humpback whales (\u003cem\u003eMegaptera novaeangliae\u003c/em\u003e) and nesting birds have broad ranges, and the identified sites focused on their nesting locations, habitats, invasive species, and reproductive needs (Feare et al. \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2007\u003c/span\u003e; Le Corre et al. \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2015\u003c/span\u003e; Amaral et al. \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). However, the large migrations of these species suggest population viability may require more than sites for reproduction. Management across these species\u0026rsquo; migratory ranges needs to be combined with the protection of specific breeding locations. Nevertheless, mating and nesting sites can be critical for many species and are therefore a good attribute for making conservation decisions. While visible and charismatic species may promote ecotourism, tourism often faces economic limits and instability in countries with volatile politics, such as Madagascar and Comoros (Spash \u003cspan citationid=\"CR56\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Therefore, providing smaller-scale information that focuses on species with important ecological services should be useful for decision makers that need to balance food production with healthy and diverse ecosystems.\u003c/p\u003e\u003cp\u003eThe biodiversity proxy method will fail to identify locations with specific species attractions, such as breeding birds and marine mammals. It should, however, be more objective for other subtidal taxa and suggests that Grand Recif was not more diverse than other locations. Grand Recif received much attention from taxonomists, and this may have created the perception of high diversity but also represents high sampling effort. Nevertheless, without comparing Grand Recif with other locations using the same sampling effort, the conclusions can be challenged. Field studies comparing corals for similar levels of sampling between regions concluded there were smaller differences among provinces (Randrianarivo et al. \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Other studies suggest that the northwest Nosy Be-Ankarea locations had the highest number of coral taxa due to rare taxa often found at depth (McClanahan et al. \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2009\u003c/span\u003e, \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). Randrianarivo et al. (\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) found that fishing and the abundance of herbivorous fishes was a more important influence on coral diversity than region. The model makes fair comparisons by holding factors constant to avoid these confounding problems. Thus, the model predictions better represent a baseline than an empirical post-human impact evaluation.\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e"},{"header":"Conclusions","content":"\u003cp\u003eThe above approach provided a more spatially refined view of biodiversity than previous national, global, and regional efforts (Allnut et al. 2012; Jenkins and Van Houtan \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2016\u003c/span\u003e). The focus on Madagascar ensures the ranking of sites is relevant to this island region and not diminished by potentially higher diversity locations, such as the African coastline or Coral Triangle (Kusumoto et al. \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; McClanahan et al. \u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Previous efforts have used the presence/absence of key taxa, limited sampling, and charismatic and visible taxa rather than the smaller-scale environmental conditions that create niches that support species and affect their distributions (van der Elst and Everett 2016). Environmental modelling provides an alternative broad-scale approach by accounting for environmental variability and ecological processes at smaller scales. These methods were better able to identify potential patterns of causation and small to modest-sized area-based conservation planning and management needs. Given more testing, the methods should eventually replace or augment past and current practices of mapping diversity from sparse or selective information. The method requires empirical data, so models are expected to improve as these data increase in scale and resolution.\u003c/p\u003e\u003cp\u003eMapped cell-level predictions showed that smaller scale environmental factors were among the strongest influences on predictions. Therefore, failure to account for this variability is likely to produce weaknesses in past recommendations based on single species and coarser scale biogeographic factors. Clearly, ground-truthing and evaluating environmental and important demographic influences will be an important next step in the prediction and prioritization process. However, the current model provided a selection of 32 locations of modest size in Madagascar, Mayotte, and Comoros that should be among those considered for focused management (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Fortunately, a high percentage of these sites have already either been included in national or LMMA management systems. The challenge now is to ensure that selected delineations receive appropriate and effective management to secure their biodiversity. The pressures of climate change, increasing cyclones, and the continual demand for fish and forest resources will continue to challenge the need to resolve nature and human need tradeoffs.\u003c/p\u003e\u003cp\u003eAddressing climate change will require participation towards carbon emission reductions goals. This requires political action to promote emission reductions nationally and globally. Reforestation that addresses watershed management and carbon emissions is also another high priority solution (Maina et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Research in fisheries is showing there are win-win outcomes for biodiversity and fisheries yields when managing catch for maximum sustained yields (McClanahan \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2018\u003c/span\u003e, \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2022a\u003c/span\u003e). Therefore, keeping fish stocks near or above optimal levels (600 kg/ha) is a practical goal to get yield and biodiversity benefits. Decisions to act on the recommended priorities will require strengthening governance institutions and commitments at multiple scales to achieve adaptation goals.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eT.R.M. concepualized the study, raised the funds, collected the field data, supervised the analysis of data and production of tables and figures, and wrote and edited the manuscript\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eLinks to environmental data are available in cited data source publications (McClanahan et al. 2024). Access to field data collected in the study sites requires a formal request to the author.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAllen GR, Werner TB (2002) Coral reef fish assessment in the \u0026lsquo;coral triangle\u0026rsquo; of southeastern Asia. Environ Biol Fish 65:209\u0026ndash;214\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAllnutt TF, McClanahan TR, Andr\u0026eacute;fou\u0026euml;t S, Baker M, Lagabrielle E, McClennen C, Rakotomanjaka AJ, Tianarisoa TF, Watson R, Kremen C (2012) Comparison of marine spatial planning methods in Madagascar demonstrates value of alternative approaches. PLoS ONE 7:e28969\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAmaral AR, Loo J, Jaris H, Olavarria C, Thiele D, Ensor P, Aguayo A, Rosenbaum HC (2016) Population genetic structure among feeding aggregations of humpback whales in the Southern Ocean. Mar Biol 163:1\u0026ndash;13\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAndrello M, Darling ES, Wenger A, Su\u0026aacute;rez-Castro AF, Gelfand S, Ahmadia GN (2022) A global map of human pressures on tropical coral reefs. Conserv Lett 15:e12858\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBrooks TM, Mittermeier RA, Da Fonseca GAB, Gerlach J, Hoffmann M, Lamoreux JF, Mittermeier CG, Pilgrim JD, Rodrigues ASL (2006) Global biodiversity conservation priorities. Science 313:58\u0026ndash;61\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBruggemann JH, Rodier M, Guillaume MM, Andr\u0026eacute;fou\u0026euml;t S, Arfi R, Cinner JE, Pichon M, Ramahatratra F, Rasoamanendrika F, Zinke J, McClanahan TR (2012) Wicked social\u0026ndash;ecological problems forcing unprecedented change on the latitudinal margins of coral reefs: The case of southwest Madagascar. Ecol Soc 17\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBurke L, Reytar K, Spalding M, Perry A (2012) Reefs at risk revisited in the coral triangle 978-1-56973-791-0. World-Fish Center, International Coral Reef Action Network, Washington, D. C\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCinner JE, Huchery C, MacNeil MA, Graham NA, McClanahan TR, Maina J, Maire E, Kittinger JN, Hicks CC, Mora C, Allison EH, D\u0026rsquo;Agata S, Hoey A, Feary DA, Crowder L, Williams ID, Kulbicki M, Vigliola L, Wantiez L, Edgar G, Stuart-Smith RD, Sandin SA, Green AL, Hardt MJ, Beger M, Friedlander A, Campbell SJ, Holmes KE, Wilson SK, Brokovich E, Brooks AJ, Cruz-Motta JJ, Booth DJ, Chabanet P, Gough C, Tupper M, Ferse SCA, R. U., Sumaila, Mouillot D (2016) Bright spots among the world\u0026rsquo;s coral reefs. Nature 535:416\u0026ndash;419\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCooke A, Lutjeharms JRE, Vasseur P (2003) Marine and coastal ecosystems. In: Goodman SM, Benstead JP (eds) The Natural History of Madagascar. University of Chicago Press, pp 329\u0026ndash;474\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCowburn B, Samoilys M, Obura D (2018) The current status of coral reefs and their vulnerability to climate change and multiple human stresses in the Comoros Archipelago, Western Indian Ocean. Mar Pollut Bull 133:956\u0026ndash;969\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCrochelet E, Roberts J, Lagabrielle E, Obura D, Petit M, Chabanet P (2016) A model-based assessment of reef larvae dispersal in the Western Indian Ocean reveals regional connectivity patterns\u0026mdash;Potential implications for conservation policies. Reg Stud Mar Sci 7:159\u0026ndash;167\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eElith J, Leathwick JR, Hastie T (2008) A working guide to boosted regression trees. J Anim Ecol 77:802\u0026ndash;813\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFeare CJ, Jaquemet S, Le Corre M (2007) An inventory of Sooty Terns (Sterna fuscata) in the western Indian Ocean with special reference to threats and trends. Ostrich-Journal Afr Ornithol 78:423\u0026ndash;434\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFontoura L, D\u0026rsquo;agata S, Gamoyo M, Barneche DR, Luiz OJ, Madin EMP, Eggertsen L, Maina JM (2022) Protecting connectivity promotes successful biodiversity and fisheries conservation. Science 375:336\u0026ndash;340\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGardner CJ, Cripps G, Day LP, Dewar K, Gough C, Peabody S, Tahindraza G, Harris A (2020) A decade and a half of learning from Madagascar's first locally managed marine area. Conserv Sci Pract 2:e298\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGardner CJ, Nicoll ME, Birkinshaw C, Harris A, Lewis RE, Rakotomalala D, Ratsifandrihamanana AN (2018) The rapid expansion of Madagascar\u0026rsquo;s protected area system. Biol Conserv 220:29\u0026ndash;36\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGough CL, A. KM, Dewar BJ, Godley E, Zafindranosy, Broderick AC (2020) Evidence of overfishing in small-scale fisheries in Madagascar. Front Mar Sci 7:317\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGrace MK, Ak\u0026ccedil;akaya HR, Bennett EL, Boyle MJW, Hilton-Taylor C, Hoffmann M, Money D, Prohaska A, Young R, Young R, and B. L (2022) The Impact of Spatial Delineation on the Assessment of Species Recovery Outcomes. Diversity 14:742\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHansen JE, Sato M, Simons L, Nazarenko LS, Sangha I, Kharecha P, Zachos JC, von Schuckmann K, Loeb NG, Osman MB (2023) Global warming in the pipeline. Oxf Open Clim Change 3:kgad008\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJenkins CN, Van Houtan KS (2016) Global and regional priorities for marine biodiversity protection. Biol Conserv 204:333\u0026ndash;339\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKerwath SE, Winker H, G\u0026ouml;tz A, Attwood CG (2013) Marine protected area improves yield without disadvantaging fishers. Nat Commun 4:2347\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKuhn M, Johnson K (2013) Applied predictive modeling. Springer, New York\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKusumoto B, Costello MJ, Kubota Y, Shiono T, Wei CL, Yasuhara M, Chao A (2020) Global distribution of coral diversity: Biodiversity knowledge gradients related to spatial resolution. Ecol Res 35:315\u0026ndash;326\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLe Corre M, Danckwerts DK, Ringler D, Bastien M, Orlowski S, Rubio CM, Pinaud D, Micol T (2015) Seabird recovery and vegetation dynamics after Norway rat eradication at Tromelin Island, western Indian Ocean. Biol Conserv 185:85\u0026ndash;94\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLe Manach F, Gough C, Harris A, Humber F, Harper S, Zeller D (2012) Unreported fishing, hungry people and political turmoil: The recipe for a food security crisis in Madagascar? Mar policy 36:218\u0026ndash;225\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMaina J, De Moel H, Zinke J, Madin J, McClanahan T, Vermaat JE (2013) Human deforestation outweighs future climate change impacts of sedimentation on coral reefs. Nat Commun 4:1\u0026ndash;7\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMaina J, McClanahan TR, Venus V, Ateweberhan M, Madin J (2011) Global gradients of coral exposure to environmental stresses and implications for local management. PLoS ONE 6:e23064\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMaina JM, Gamoyo M, Adams VM, D'agata S, Bosire J, Francis J, Waruinge D (2020) Aligning marine spatial conservation priorities with functional connectivity across maritime jurisdictions. Conserv Sci Pract 2:e156\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMaire E, Cinner J, Velez L, Huchery C, Mora C, Dagata S, Vigliola L, Wantiez L, Kulbicki M, Mouillot D (2016) How accessible are coral reefs to people? A global assessment based on travel time. Ecol Lett 19:351\u0026ndash;360\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcClanahan TR (2018) Multicriteria estimate of coral reef fishery sustainability. Fish Fish 19:807\u0026ndash;820\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcClanahan TR (2021) Marine reserve more sustainable than gear restriction in maintaining long-term coral reef fisheries yields. Mar policy 128:104478\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcClanahan TR (2022a) Fisheries yields and species declines in coral reefs. Environ Res Lett 17:044023\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcClanahan TR (2022b) Coral responses to climate change exposure. Environ Res Lett 17:073001\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcClanahan TR (2023) Local heterogeneity of coral reef diversity and environmental stress provides opportunities for small-scale conservation. Divers Distrib 29:1324\u0026ndash;1340\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcClanahan TR, Abunge CA (2019) Conservation needs exposed by variability in common-pool governance principles. Conserv Biol 33:917\u0026ndash;929\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcClanahan TR, Abunge CA (2020) Perceptions of governance effectiveness and fisheries restriction options in a climate refugia. Biol Conserv 246:108585\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcClanahan TR, Ateweberhan M, Graham NAJ, Wilson SK, Sebasti\u0026aacute;n CR, Guillaume MM, Bruggemann JH (2007) Western Indian Ocean coral communities: Bleaching responses and susceptibility to extinction. Mar Ecol Prog Ser 337:1\u0026ndash;13\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcClanahan TR, Ateweberhan M, Omukoto J, Pearson L (2009) Recent seawater temperature histories, status, and predictions for Madagascar's coral reefs. Mar Ecol Prog Ser 380:117\u0026ndash;128\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcClanahan TR, Azali MK (2021) Environmental variability and threshold model\u0026rsquo;s predictions for coral reefs. Front Mar Sci 8:1774\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcClanahan TR, Friedlander AM, Wickel J, Graham NAJ, Bruggemann HJ, Guillaume MMM, Chabanet P, Porter S, Schleyer MH, Azali MK, Muthiga NA (2024) Testing for concordance between predicted species richness, past prioritization, and marine protected area designations in the western Indian Ocean. Conservation Biology:e14256\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcClanahan TR, Graham NAJ (2015) Marine reserve recovery rates towards a baseline are slower for reef fish community life histories than biomass. Proceedings of the Royal Society B: Biological Sciences 282:20151938\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcClanahan TR, Graham NAJ, MacNeil MA, Cinner JE (2015) Biomass-based targets and the management of multispecies coral reef fisheries. Conserv Biol 29:409\u0026ndash;417\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcClanahan TR, Jadot C (2017) Managing coral reef fish community biomass is a priority for biodiversity conservation in Madagascar. Mar Ecol Prog Ser 580:169\u0026ndash;190\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMcClanahan TR, Maina JM, Muthiga NA (2011) Associations between climate stress and coral reef diversity in the Western Indian Ocean. Glob Change Biol 17:2023\u0026ndash;2032\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eObura DO, Church JE, Gabri\u0026eacute; C (2012) Assessing Marine World Heritage from an Ecosystem Perspective: The Western Indian Ocean. World Heritage Centre, United Nations Education, Science and Cultural Organization (UNESCO), Paris, France\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eParker QM, Longosoa HT, Long S, Jones PJS (2024) A longitudinal governance analysis of a locally managed marine area: Ankobohobo wetland small-scale mud crab fishery. Madagascar Mar policy 163:106138\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eParravicini V, Vill\u0026eacute;ger S, McClanahan TR, Arias-Gonz\u0026aacute;lez JE, Bellwood DR, Belmaker J, Chabanet P, Floeter SR, Friedlander AM, Guilhaumon F, Vigliola L, Kulbicki M, Mouillot D (2014) Global mismatch between species richness and vulnerability of reef fish assemblages. Ecol Lett 17:1101\u0026ndash;1110\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePilowsky JA, Colwell RK, Rahbek C, Fordham DA (2022) Process-explicit models reveal the structure and dynamics of biodiversity patterns. Sci Adv 8:eabj2271\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRanaivomanana HS, Jaquemet S, Ponton D, Behivoke F, Randriatsara RM, Mahafina J, L\u0026eacute;opold M (2023) Intense pressure on small and juvenile coral reef fishes threatens fishery production in Madagascar. Fish Manag Ecol 30:494\u0026ndash;506\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRandrianarivo M, Botosoamananto RL, Guilhaumon F, Penin L, Todinanahary G, Adjeroud M (2023) Effects of Madagascar marine reserves on juvenile and adult coral abundance, and the implication for population regulation. Mar Environ Res :106080\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRandrianarivo M, Gasimandova LM, Tsilavonarivo J, Razakandrainy A, Philippe J, Guilhaumon F, Botosoamananto RL, Penin L, Todinanahary G, Adjeroud M (2024) Assessing recovery potential of coral reefs in Madagascar and the effects of Marine Protected Areas. Regional Studies in Marine Science:103710.\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRandrianarivo M, Guilhaumon F, Tsilavonarivo J, Razakandrainy A, Philippe J, Botosoamananto RL, Penin L, Todinanahary G, Adjeroud M (2022) A contemporary baseline of Madagascar\u0026rsquo;s coral assemblages: Reefs with high coral diversity, abundance, and function associated with marine protected areas. PLoS ONE 17:e0275017\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRocliffe S, Peabody S, Samoilys M, Hawkins JP (2014) Towards a network of locally managed marine areas (LMMAs) in the Western Indian Ocean. PLoS ONE 9:e103000\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSamoilys M, Alvarez-Filip L, Myers R, Chabanet P (2022) Diversity of coral reef fishes in the western Indian Ocean: Implications for Conservation. Diversity 14:102\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSmallhorn-West P, Cohen PJ, Kukiti E, Rabbitt S, Sori F, Tapala S, Warren R, Boso D, Foale S (2022) Ten years of dynamic co-management of a multi-species reef fishery. Coral Reefs 41:1449\u0026ndash;1464\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSpash CL (2021) The economy\u0026rsquo;as if people mattered: revisiting critiques of economic growth in a time of crisis. Globalizations 18:1087\u0026ndash;1104\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSully S, Hodgson G, van Woesik R (2022) Present and future bright and dark spots for coral reefs through climate change. Glob Change Biol 28:4509\u0026ndash;4522\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTyberghein L, Verbruggen H, Pauly K, Troupin C, Mineur F (2012) and O. De Clerck. Bio-ORACLE: A global environmental dataset for marine species distribution modelling. Global Ecology and Biogeography 21:272\u0026ndash;281\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eVan der Elst R, Everett B (2015) Offshore fisheries of the Southwest Indian Ocean: Their status and the impact on vulnerable species. Oceanographic Research Institute (ORI) and the Western Indian Ocean Marine Sciences Association (WIOMSA), Durban\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWells S, Ray GC, Gjerde KM, White AT, Muthiga N, Bezaury JE, Creel BD, Causey J, McCormick-Ray R, Salm S, Gubbay, Kelleher G (2016) Building the future of MPAs\u0026ndash;lessons from history. Aquatic Conservation: Marine and Freshwater Ecosystems 26:101\u0026ndash;125\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWilson K, L. DP, Tittensor B, Worm, Lotze HK (2020) Incorporating climate change adaptation into marine protected area planning. Glob Change Biol :1\u0026ndash;17\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eConvention UNEP-N and WIOMSA (2021) Western Indian Ocean Marine Protected Areas Outlook: Towards achievement of the Global Biodiversity Framework Targets. UNEP and WIOMSA, Nairobi, Kenya\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eYeager LA, Marchand P, Gill DA, Baum JK, McPherson JM (2017) Marine socio-environmental covariates: Queryable global layers of environmental and anthropogenic variables for marine ecosystem studies. Ecology 98:1976\u0026ndash;1976\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZafimahatradraibe JA, Ranaivomanana LNJ, Rakotomahazo C, Randriamanantsoa B, Hartmann AC, Todinanahary GGB (2025) Authority, capacity, and power to govern: Three marine protected areas co-managed by resource users and non-governmental organizations. Mar policy 177:106647\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZeller D, Vianna G, Ansell M, Coulter A, Derrick B, Greer K, No\u0026euml;l SL, Palomares MLD, Zhu A, Pauly D (2021) Fishing effort and associated catch per unit effort for small-scale fisheries in the Mozambique Channel region: 1950\u0026ndash;2016. Front Mar Sci 8:1048\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-7932631/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7932631/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003ePurpose\u003c/h2\u003e\u003cp\u003eTo determine the distribution of biodiversity in coral reefs of Madagascar, Comoros, and Mayotte.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e\u003cp\u003eA Gradient Boosting Model (GBM) evaluating seventy environmental spatial databases predicted reef biodiversity field data to create spatial predictions in 2854 6-km\u003csup\u003e2\u003c/sup\u003e mapped reef cells in 2020 and 2050. Predicted biodiversity were compared to past provincial protected area prioritization activities and the current listing of marine national parks (MNP), Locally Managed Marine Areas (LMMAs), and the World Protected Area Database (WDPA).\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e\u003cp\u003eTwenty-one national high biodiversity priority cells were selected for Madagascar, 3 for Mayotte, and 8 for Comoros. Sixteen of the 32 selected high biodiversity locations were contained in 44 of the 102 possible listed WDPA protected areas. The east and coastal reefs south of Antongil Bay and offshore coral reefs islands were notably excluded from national but not LMMA designations. Madagascar\u0026rsquo;s west coast was better represented than the east coast in WDPA locations. Based on surface temperate predictions, coral cover declined in 55% and gained in 7%, while numbers of taxa declined in 72% but gained in 14% of the grid cells between 2020 and 2050.\u003c/p\u003e\u003ch2\u003eConclusions\u003c/h2\u003e\u003cp\u003eSpatial cells with minor climate-induced changes or gains in coral reef cover and diversity attributes were broadly scattered among governance authorities. However, most locations with little climate change effects were in southwest Madagascar where overfishing is likely to undermine their climate refugia potential.\u003c/p\u003e","manuscriptTitle":"Modelling coral reef biodiversity for prioritizing marine protected area in Madagascar and adjacent islands ","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-19 13:38:43","doi":"10.21203/rs.3.rs-7932631/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"c7cf2127-a096-42a8-9a52-99217bff0d74","owner":[],"postedDate":"November 19th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-11-19T13:38:43+00:00","versionOfRecord":[],"versionCreatedAt":"2025-11-19 13:38:43","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7932631","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7932631","identity":"rs-7932631","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}