Meta-analysis of Red List conservation assessments of Mexican endemic and near endemic tree species shows nearly two thirds of these are threatened

Abstract

Mexico is floristically the fourth most species-rich country in the world. Currently, 3620 native tree species are reported from this country. Mexico has been a forerunner at global level in recognising the need for plant conservation, yet this is in stark contrast with governmental programmes and actions, past and present that negatively affect(ed) immense areas of primary vegetation. In the framework of the Global Tree Assessment, to date, we have assessed the conservation status of nearly 1500 Mexican endemic and near endemic tree species for the International Union for the Conservation of Nature (IUCN) Red List, the distribution data of which we use here for a first meta-analysis of conservation statuses and spatial distribution patterns in a newly delimited area we call Natural Mexico.Our database consists of 112,416 taxonomically and geographically very carefully curated distribution points of 1474 tree species endemic to Mexico and adjacent areas, belonging to 98 flowering plant families. The extensive curation methodology we consider essential for both research and conservation purposes is emphasised.Nearly 60% of the assessed tree species are threatened, almost double the percentage of threatened arborescent species at global level. Tropical rainforests and cloud forests house the highest proportions of threatened trees.These assessments have an extensive impact, as they not only are the starting point of urgent species conservation actions but also allow for comprehensive studies including extinction risk estimation, gap analyses for conservation planning and species reintroductions, all this in a framework of land use change, climate change and landscape composition and configuration.

INTRODUCTION

One of the few countries on our planet that houses two biodiversity hotspots: the Madrean pine-oak woodlands and a considerable part of Mesoamerica (Myers et al., 2000) and with more than 23,000 vascular plant species (Villaseñor, 2016), of which approximately 11,000 are endemic (Ulloa Ulloa et al., 2017; Villaseñor, 2016); Mexico is the fourth richest country in terms of vascular plant diversity, despite being only the 14th largest country in the world (Ramamoorthy et al., 1993). Apart from the abovementioned biodiversity hotspots, Mexico's extraordinary plant diversity is also due to the following unique factors: (1) the country is geographically located in two biogeographic realms, the Nearctic and the Neotropics (Udvardy, 1975), and especially the Yucatán Peninsular flora shares a high proportion with the Caribbean flora (Rzedowski, 1978), as a consequence of which a unique flora has developed (Rzedowski, 1993); (2) long-time isolation of several large particular areas, such as the Balsas basin and Tehuacán-Cuicatlán valley, as well as the Baja California peninsula (Rzedowski, 1978), leading to a high percentage of endemic species; and (3) most globally known vegetation types occur in Mexico and several of these do not occur anywhere else in the world (Holdridge, 1947; Miranda & Hernández-Xolocotzi, 1963).

This enormous plant diversity includes all known kinds of tropical and temperate forests, as well as arid vegetation. The species diversity of the dry tropical forests of the Mexican Pacific slope equals that of the Amazonian tropical rainforest (Mario Sousa, pers. comm., 2000). Some of these communities are unique on our planet, such as three completely different types of forests with cactus species with secondary growth, called nopaleras, tetecheras and cardonales (Miranda & Hernández-Xolocotzi, 1963), occurring throughout Mexico, as well as the otherworldly looking boojum forest of the Baja California Peninsula. The Yucatán Peninsula is another example of this unusual diversity, for example, the secondary speciation centre of the family Sapotaceae, which, although the area does not have visible barriers, 12% of its plants are endemic due to a large gypsum outcrop that acts as an edaphic barrier creating physiological aridity and calcium toxicity (Tom Wendt, pers. comm., 1997). All these vegetation types occur in an extraordinary mosaic, shaped by the complex and varied geological history setting the stage for one of the most heterogeneous physical-geographical scenarios on earth (Ferrusquía-Villafranca, 1993). The large elevational gradients in turn result in a great diversity of climatic conditions, on top of the large latitudinal climate variation that characterises Mexico (García, 2004).

The inventory of this huge floristic diversity is partial and incomplete, despite many efforts by many botanists (Mario Sousa, pers. comm., 1984). About 100 plant species new to science are described annually from this country (Alvarado-Cárdenas & Chávez Hernández, 2022). A total of 5 million herbarium specimens from Mexican plants is estimated to exist, which largely come from regional floristic projects (e.g. Flora de Baja California, Flora del Desierto de Sonora, Flora del Desierto de Chihuahua, Flora Mesoamericana, Flora de Chiapas, Flora de Veracruz, Flora Novo-Galiciana, Floras del Valle de México, Flora del Bajío y de regiones adyacentes, Flora de Guerrero, Flora de Oaxaca, Flora del Estado de México and Flora de Jalisco). Additionally, a national biological collection project involved collections in other zones that are not included in the abovementioned areas (CONABIO, 2016). Several great 19th and 20th century collectors explored throughout Mexico, such as Cyrus Guernsey Pringle, George B. Hinton, Carl Albert Purpus, Edward Palmer, Reid Moran, Howard Scott Gentry and others of the US Department of Agriculture, Rogers McVaugh, Jerzy Rzedowski, and Mario Sousa. In the last 20 years of the 20th century, many state herbaria were founded, but the collections of most of these are not online available. Whichever plant group that has been studied in detail in the last 40 years, including intensive exploration, resulted in about 30%–40% species new to science, as a consequence of which the estimations prior to the study of Villaseñor (2016) may be closer to the actual plant diversity in Mexico (30,000 to 32,000). Especially the large mountain chains (Sierra Madre Occidental, Sierra Madre Oriental, Sierra Madre del Sur and Sierra Madre de Chiapas) lack thorough exploration, because they have been and are relatively dangerous for a variety of reasons, as well as the large agricultural and cattle areas where no exploration was possible before the vegetation disappeared (e.g. many species descriptions from the state of Sinaloa are only known from their type localities). Finally, it is well-known that herbaria are a treasure vault for the discovery of new species (e.g. Bebber et al., 2010), so many species that will be recognised as new to science in the next decades are already available in herbarium collections. An estimated 16% of the total number of known vascular plant species in Mexico are trees (own unpublished data). The first comprehensive overview of the woody flora of Mexico is that of Standley (1920–1926). More recent research has focused on tree species of certain geographic or climatic zones such as those of Pennington and Sarukhán (2005) about the most common tropical trees, the study of Rzedowski (2015) of the trees of the Sierra Madre Oriental, and the list of Ibarra-Manríquez et al. (1995) of the arborescent species of the Yucatán peninsula, or synthesise information about specific tree-rich plant families (Ricker et al., 2013, 2016; Ricker & Hernández, 2010; Sousa et al., 2003).

In the framework of the Global Tree Assessment, Beech et al. (2017) reported 3364 tree species for Mexico, positioning this country in the top 10 of the most tree species-rich countries, and due to our efforts in listing additional tree species since then, this has been increased to 3620 species (BGCI, 2022). The recently published report ‘State of the World's Trees’ (BGCI, 2021) summarises the importance of and main threats to trees, based on tree conservation assessments worldwide. According to this report, one-third of the trees at global level is threatened. However, some groups such as the Magnoliaceae show a much higher degree of threat. In contrast, our recent study on Mexican arborescent Asteraceae species endemic and near endemic to Mexico (i.e., those shared with the south of the United States of America north of Mexico, and with Central America south of the country) showed that relatively few daisy trees are currently seriously threatened, as many of them are relatively widespread (Redonda-Martínez et al., 2021).

A yet unknown number of Mexican endemic and near endemic arborescent species are threatened by land use change and/or climate change, which both are seriously and visibly affecting its biodiversity, and potentially urban and rural livelihoods. The impact of the decline and loss of tree species, the principal components of forest ecosystems which play an important role in the Earth's biogeochemical processes (Millennium Ecosystem Assessment, 2005), are largely understudied because their conservation status has not been comprehensively assessed (Newton et al., 2015). Only a few efforts have been realised in Mexico to comprehensively assess tree conservation statuses, one focusing on a specific habitat, the Red List of Mexican Cloud Forest Trees (González-Espinosa et al., 2011) and three focusing on a taxonomic group, the Magnolia Red List (Rivers et al., 2016) in which 15 Mexican Magnolia species were assessed, the Fraxinus Red List (Barstow et al., 2018), containing 16 endemic Mexican species, and the Red List of Oaks, which includes 164 Mexican Quercus species (Carrero et al., 2020). The first official Mexican standard including threatened organisms was established in 2002 (NOM-059-SEMARNAT-2002, SEMARNAT, 2002), but at least the section about plants lacks rigour and in some cases is subordinate to political and economic interests. The most recent version of this list, NOM-059-SEMARNAT-2010 (SEMARNAT, 2010), currently under revision for improvement, includes plant species that are widespread and not threatened, and lacks thousands of seriously threatened species, as the process does not involve a systematic review of all organisms, and instead depends on proposals that are presented upon invitation to the communities of specialists, which are then reviewed by committees of specialists. Additionally, research of species included in this list requires a special permit, likely diminishing the number of species proposed to be included. Finally, the inclusion of species in the list also restricts their use by local people. The recent study of Téllez et al. (2020) is a valuable first approach to patterns of tree diversity, distribution, uses and conservation in Mexico. However, an unfortunate caveat of this study is the lack of rigorous curation of distribution record data and taxonomic identification (Villaseñor, 2021), mainly due to errors in the original databases of the herbaria that house the specimens.

Mexico has been a forerunner in recognising the need for plant conservation and was one of the first countries in the world to develop a national strategy, the Mexican Plant Conservation Strategy, in line with the Global Strategy for Plant Conservation of the Convention on Biological Diversity (CONABIO, 2012). The National Commission for the Knowledge and Use of Biodiversity (CONABIO in Spanish, from ‘Comisión Nacional para el Conocimiento y Uso de la Biodiversidad’), founded in 1992 as an interministerial applied research organisation promoting, coordinating, supporting and carrying out activities aimed at knowledge of biological diversity, as well as its conservation and sustainable use for the benefit of society (CONABIO, 2022), was also a pioneer organisation of its kind at global level. The National Commission of Protected Natural Areas (CONANP in Spanish, from ‘Comisión Nacional de Áreas Naturales Protegidas’) currently administers 183 federal natural protected areas, representing 90,942,124 ha and supports 371 Areas Voluntarily Designated for Conservation with a surface of 596,867.34 ha (CONANP, 2021). Additionally, throughout the Mexican territory, there are also multiple protected areas at the state, municipal and local level, as well as initiatives of communities, ‘ejidos’ (communal property ruled by general rules where members own their own part) and private individuals.

However, these conservation arrangements are in stark contrast with other governmental programmes and actions, past and present, that destroyed, are destroying or plan to demolish immense areas of primary vegetation throughout the country, mainly for agricultural activities (Bravo Peña et al., 2010; Moreno Unda et al., 2019). Other important factors that have led or are leading to large-scale primary forest destruction are mining activities, fires, water management and tourism projects, each of these also leading to important socio-ecological conflicts (see Rodríguez Robayo et al., 2021, for a recent overview).

Since 2018, the authors of this paper, in the framework of the Global Tree Assessment under the auspices of Botanic Gardens Conservation International (BGCI) and the IUCN/SSC Global Tree Specialist Group, have been involved in the preparation of conservation assessments of Mexican endemic and near endemic trees for the IUCN Red List. The present study aims at a first comprehensive meta-analysis of the information we have generated since then and that largely has been published on the IUCN Red List in the meantime. Therefore, our objectives are, based on taxonomically and geographically very carefully curated distribution data of nearly 1500 Mexican endemic and near endemic arborescent species, to (1) emphasise the importance of data curation for conservation purposes; (2) define general patterns of distribution and conservation statuses of near endemic and endemic tree species; and (3) present a preliminary discussion of our data in the light of the UN Decade for Ecosystem Restoration and prospects for tree conservation in Natural Mexico.

MATERIALS AND METHODS Study area

The study area was defined taking as a starting point Megamexico 3 as delimited by Rzedowski (1991), with additions in the north and south according to natural characteristics (Figure 1). We present here these novel limits, based on diverse bibliographical sources and extensive field knowledge of the last author of this paper. Whenever possible, subdivisions are defined on the basis of the three main biologically relevant variables: topography, climate and vegetation (Jepson Flora Project, 2022). These three variables do not always change together, and in such cases, vegetation differences usually have priority (Jepson Flora Project, 2022). In some cases, transitions in vegetation occur gradually and there is no apparent botanical basis for drawing a sharp line (Tunnell & Judd, 2002). Where this occurs, the limits primarily follow river valleys and the continuation of the Laguna Madre (The START group, 2007). Together, the Laguna Madre of Texas and the Laguna Madre de Tamaulipas form the largest hypersaline system in the world (Delphi, 2020; Tunnell & Judd, 2002). Vegetation types in Spanish according to Miranda and Hernández-Xolocotzi (1963) are mentioned between quotation marks when it is relevant to refer to a specific vegetation type, given that the classification system of the aforementioned authors uses local names that find their origin in native languages that historically delimited those vegetation types, and are therefore very specific.

FIGURE 1. The study area named Natural Mexico (delimited in grey), in which the tree species we assessed for the International Union for the Conservation of Nature (IUCN) Red List occur. We defined this area by taking Megamexico 3 as delimited by Rzedowski (1991) as a starting point, with additions in the north and south according to natural characteristics related to vegetation, topology and climate.

The northern limit from west to east was established according to the following criteria:

1. South Coast Subregion, in the Southwestern California Region, of the California Floristic Province; matorral and chaparral vegetation predominate (Jepson Flora Project, 2022);
2. Peninsular Ranges Subregion excluding the San Jacinto Mountains, in the Southwestern California Region, of the California Floristic Province; the vegetation includes coniferous forest, deciduous forest, oak forest and pine forest (Jepson Flora Project, 2022);
3. Sonoran Desert Region in the Desert Province (Jepson Flora Project, 2022); characteristic plant species are Olneya tesota, Carnegiea gigantea, Parkinsonia florida, Fouquieria splendens and Justicia californica (Sánchez Escalante, 2007). This area is also the northern limit of the occurrence of jaguar (Center for Biological Diversity, 2022; Sanderson et al., 2002);
4. Chihuahuan Desert Ecoregion (U.S. National Park Service, 2022). This area is characterised by the following vegetation: grasslands (‘pastizales’), Yucca forest (‘izotales’), Opuntia scrub (‘nopaleras’), Agave scrub (‘magueyeras’), evergreen thorn scrub forest (‘mezquitales’), halophilic and gypsophyllous scrub, and riparious vegetation (Granados-Sánchez et al., 2011). Moreover, this is an area of high Cactaceae diversity (WWF, 2022); and
5. The lowland areas of the South Texas Plains, south of the Edwards Plateau and the lowest elevations south of the Gulf Coastal Prairies, until the region of Corpus Christi. These areas tend to merge imperceptibly. The characteristic vegetations are grasslands and many types of scrublands (The START group, 2007). This area is, together with another small zone in the Sonoran Desert, also the northern limit of the occurrence of ocelot (Lombardi et al., 2020; Paviolo et al., 2015).

The southern limit from north to south includes the following areas:

1. In Honduras: from La Ceiba to Juticalpa, following the mountains, with pine forests (excluding Pinus caribaea) and cloud forest, including the arid zone of Olanchito. Moreover, it included tropical dry forest (low and medium-sized), both deciduous and semideciduous, until Trojes at the border with Nicaragua (E. M. Martínez Salas, pers. comm., 2022);
2. In Nicaragua: the central mountains (‘Zona Norcentral’) with oak forest, pine forest (excluding P. caribaea), as well as cloud forest); the Pacific (‘Zona Pacífica’), and east and west of the lakes, with tropical dry forest (low and medium-sized), both deciduous and semideciduous (Stevens, 2022);
3. In Costa Rica until south of San José in areas with cloud forest, oak forest and semideciduous low and medium-sized tropical dry forest (E. M. Martínez Salas, pers. comm., 2022).

The northernmost distribution of deciduous tropical dry forest ends in the latter region, as well as the occurrence of pine forest excluding Pinus caribaea, and also many cloud forest species. It is also a barrier for plant species and families, as well as animals (such as sloths) of Amazonian origin.

We call this newly delimited zone from here onwards ‘Natural Mexico’, as it is clear that it can be considered a delimitation based on natural factors. The polygon was delimited using the geographic information system QGIS v. 3.16.3 (QGIS Development Team, 2019).

Tree definition, species selection and Red List assessments

The applied tree definition is the one agreed on by the IUCN/SSC Global Tree Specialist Group (GSTG), which has also been applied by Beech et al. (2017): a woody plant, usually with a single stem growing to a height of at least 2 m, or if multi-stemmed, then at least one vertical stem 5 cm in diameter at breast height.

The Mexican endemic and near endemic tree species list was extracted from GlobalTreeSearch (BGCI, 2022) and served as the starting point for our project. Next, this list was carefully checked with respect to the taxonomy of all species. The tree habit according to the abovementioned definition was confirmed or rejected based on the comprehensive field knowledge of the last author of this study. The resulting list was shared with the GTSG, followed by the generation of this list in the IUCN Species Information Service (SIS), the central IUCN database to store and manage species accounts and assessments for publication on The IUCN Red List.

Certain taxonomic groups, for example, oaks, conifers, cacti, palms, Annonaceae, Ebenaceae, Lauraceae and Sapotaceae had already been assessed or were assigned to specialists in these groups for assessment and are therefore not included in the present study. Different data sets may have been utilised for these other assessments, not following the same curation protocol for all taxonomic and distribution data that we developed before starting each species assessment (see further).

At the start of each assessment, the taxonomy of each species was checked again to certify there were no taxonomic or nomenclatural actualisations at that point. The preparation of the Red List assessments consisted of the evaluation of the number of locations and the state of habitat conservation or destruction throughout the distribution areas of each species. Moreover, occurrence data at state level, population data, habitat and ecology data, use and trade information at local, national and/or international level, known threats, and known and/or suggested conservation actions were completed based on literature, collection data labels from herbaria, and field knowledge. The most difficult topics to gather current information on, based on literature and specimen data, were population sizes and trends, as such data are usually not included on herbarium specimen labels or in taxonomic studies which do not include conservation aspects, but here again, the extensive field knowledge of the last author of this paper was invaluable. Red List categories were obtained according to the IUCN Red List criteria (IUCN, 2012).

All our assessments were reviewed by a specialist or a member of the GSTG, followed by submission to the IUCN Red List by BGCI staff, and a final thorough revision by the IUCN Red List staff.

Curation of taxonomic and distribution data

Prior to the preparation of each conservation assessment, distribution data were downloaded from GBIF (2022) via GeoCAT (Bachmann et al., 2011) for each species, including its synonym(s). Each of these occurrences was carefully curated, including checking herbarium specimens, generally in the herbaria MEXU and MO (acronyms according to Thiers, continuously updated), as well as others (F, G, GH, LL, MICH, NY, TEX and US) when the image of the specimens was online available, including photos of the type specimens on JSTOR Global Plants (JSTOR, 2022). When there were no occurrences in GBIF, or when data were incomplete or incorrect, these were completed with collection information from the herbaria MEXU and MO (for the family Asteraceae, data from US, MICH, GH and RSA, listed in order of importance, were also included), and georeferenced in Google Earth (2022) or directly in GeoCAT when they had no geographic coordinates, for example, in the case of historical collections. All geographical coordinates are in WGS 84 datum.

In the following paragraphs, we detail some of the most recurrent situations we encountered during the curation of the retrieved data. iNaturalist records were deleted unless a specialist of that particular plant group had certified the identification of each of the records. Monographic studies were consulted when available, which allowed the certification of distribution patterns endorsed by the taxonomic specialist of each species. When inconsistencies were detected in GBIF data, collection labels were revised, leading to confirmation or correction of the geographic location. We verified that all records were within the geographic range and vegetation type(s) of each species.

Records of the herbarium F are always automatically located in the centroid of the country. When collection data were available, the record was placed in the correct location. When these were not available, the record was deleted. This effort was especially done in the case of type specimens, using other sources of information, such as JSTOR Global Plants (JSTOR Global Plants, 2022), herbaria, monographs, and protologues. We detected that the geographic records of some herbaria were consistently erroneous; in the case of these institutions, all records were carefully revised. An estimated 30,000 records were not available in GBIF, mainly of the National Herbarium of Mexico MEXU (UNAM, 2022) and the herbarium MO (Tropicos, 2022), in the case of Asteraceae also those of the herbaria MICH and US, and these were consequently georeferenced and added to the maps of each species.

We identified conflicts of interest between research groups and institutions working on the same plant groups. In such cases, we evaluated the certainty of each taxonomic determination giving priority to the criterion of the specialists in these plant groups or to the professional trajectory of particular taxonomists. Moreover, each source institution was also evaluated (e.g. the specimens of the collections of Dennis Breedlove in the herbarium CAS are consistently changed to a non-existing person Álvarez Álvarez in GBIF, albeit with the same collection number); if we know that their collection is not curated by specialists, or they do not share information it was not taken into account for the species assessment. When specimens are not available to botanists, but all localities are in GBIF (e.g. herbaria FCME and HEM), and they included new records in areas previously explored by specialists, none of the GBIF data were taken into account. Another situation is when a specialist only certified part of a herbarium collection, and these are from a specific geographic area, and those from other areas were not certified. As a consequence, each specific case was evaluated to decide whether the uncertified information was taken into account or not.

When the distribution map was generated and certain records were far away from the conglomerate of records, or when coordinates were assigned by automated systems such as those of CONABIO, these were not taken into account unless the original collection label could be checked. This was especially problematic in the case of historical collections with common locality names (e.g. San José) that were located in erroneous points. When duplicates of the same collection were deposited in different herbaria, only one specimen and its distribution record was retained. This is indeed a very time-consuming process, but without any doubt, it has resulted in high-quality distribution maps reflecting the true occurrence of the assessed species. In general, only between one third and one quarter of the GBIF data could be maintained based on the abovementioned criteria, focusing on data supported by herbarium specimens.

Finally, plant groups that do not possess taxonomic revisions or only deficient monographic studies were not assessed.

Spatial distribution analyses

Using the combined species data, analyses of spatial distribution were undertaken. The databases were reviewed using the statistical software R v. 4.0.5 (R Core Team, 2018) and the packages rgdal v. 1.5–23 (Bivand et al., 2021), dplyr v. 1.0.4 (Wickham et al., 2021) and raster (Hijmans, 2021). Species whose distribution extended more than two geographic degrees from the limits of the study area were excluded. For each of the points, the corresponding information was extracted from the layers of protected natural areas (CONABIO, 2020; SEMARNAT and CONANP, 2020), biogeographic provinces (CONABIO, 1997; Löwenberg-Neto, 2014), land use and vegetation (CONABIO, 1999). Distribution maps of each of the species and families were prepared using R and the raster and ggplot2 packages (Wickham, 2016). Species diversity of each biogeographic province and natural reserve was obtained using R and QGIS.

The complete dataset with all distribution points is available upon request from the corresponding author, for well-argumented scientific and conservation purposes only. Unfortunately, this dataset cannot be made publicly available, given that an important number of the species is threatened and that their occurrence data are particularly sensitive from a conservation point of view. It should be mentioned that none of the occurrence points of the threatened species we have assessed is available on the IUCN Red List either and that a very high percentage of these differs from the points of the same specimens on GBIF.

RESULTS Tree species diversity, distribution and IUCN red list categories

Our database consisted of 112,416 very carefully curated distribution points of 1474 tree species endemic and near endemic to Natural Mexico, belonging to 98 flowering plant families (Table 1, Dataset S1). Therefore, the spatial distribution analyses resulted in 1474 species maps and 98 family maps (not shown), as well as eight maps according to each Red List conservation category (LC, NT, VU, EN, CR, EX, EW and DD; not shown) and one combined map of the three ‘threatened’ categories (VU, EN and CR; Figure 2). Also resulting from the spatial distribution analyses are six maps according to layers: federal protected natural areas (Figure 3), state-protected natural areas, species diversity at country and state levels, and biogeographic provinces/regions according to Morrone (Löwenberg-Neto, 2014, not shown) and CONABIO (1997; Figure 4), respectively. Finally, there were also five maps focusing on specific regions with different histories of soil use, vegetation types and species richness patterns: the Baja California peninsula (Figure 5), the Bajío region (including parts of the states of Guanajuato, Jalisco, Michoacán and Querétaro; Figure 6), Guerrero (Figure 7), Veracruz (Figure 8), and Chiapas plus Guatemala (Figure 9).

TABLE 1 List of the 98 flowering plant families which contain the tree species assessed by the authors for the International Union for the Conservation of Nature (IUCN) Red List

Family	Genera	Species	Assessment
			LC	NT	VU	EN	CR	EW	EX	DD
Achariaceae	1	1				1
Achatocarpaceae	1	2				2
Actinidiaceae	1	17	3	4	5	5
Amaranthaceae	1	2	2
Anacardiaceae	10	19	10		5	3	1
Annonaceae	2	7				4	3
Apocynaceae	3	8	2	1	2	3
Aquifoliaceae	2	6	2			2	2
Araliaceae	3	12	6		2	3	1
Arecaceae	13	21	12		3	5				1
Asparagaceae	4	22	3		11	5	2	1
Asteraceae	40	115	51	23	17	21	2			1
Berberidaceae	1	6	4			2
Bignoniaceae	2	12	3		2	5	2
Burseraceae	2	84	25	9	25	21	4
Buxaceae	2	2			1	1
Capparaceae	2	8	2	1	2	2	1
Caricaceae	1	1			1
Celastraceae	8	38	9	1	8	14	3			3
Chloranthaceae	1	1	1
Chrysobalanaceae	3	3				2				1
Cleomaceae	1	1	1
Clethraceae	1	15	5		4	6
Clusiaceae	1	4	3	1
Convolvulaceae	1	3	2			1
Cordiaceae	1	12	5		3	3	1
Cunoniaceae	1	1				1
Dichapetalaceae	1	1					1
Dipentodontaceae	1	1	1
Ehretiaceae	2	10	2	1	1	5			1
Elaeocarpaceae	1	5				4	1
Ericaceae	5	11	10	1
Erythropalaceae	1	1					1
Erythroxylaceae	1	1	1
Euphorbiaceae	17	95	22	3	23	37	6			4
Fabaceae	48	214	82	8	40	65	18			1
Fouquieriaceae	1	10	2		3	5
Garryaceae	1	5	2	1	1	1
Hamamelidaceae	3	3	1			2
Hernandiaceae	1	1			1
Hypericaceae	1	1				1
Icacinaceae	1	1					1
Juglandaceae	4	7	2	2	1	1				1
Lamiaceae	3	5	2		2	1
Lauraceae	3	20	2	1	2	11	3			1
Lecythidaceae	1	1					1
Loasaceae	1	1	1
Lythraceae	1	1	1
Magnoliaceae	1	2				2
Malpighiaceae	4	8	4		1	1	2
Malvaceae	17	68	26	1	10	22	7			2
Melastomataceae	5	13	3		2	7	1
Meliaceae	3	10	1		2	3	4
Menispermaceae	1	2	1			1
Metteniusaceae	3	4	2			1	1
Monimiaceae	1	1				1
Moraceae	2	4	4
Myrtaceae	3	28	5	2	7	10	3			1
Nyctaginaceae	2	2	1		1
Ochnaceae	1	5			2	2	1
Oleaceae	1	5	3			2
Onagraceae	1	1			1
Opiliaceae	1	3	1			1	1
Papaveraceae	1	4			2	2
Pentaphylacaceae	3	10	1	1	4	2	2
Peraceae	1	1				1
Petenaeaceae	1	1				1
Phyllanthaceae	2	2	1			1
Picramniaceae	1	6	1		2	2	1
Picrodendraceae	1	1			1
Platanaceae	1	2	1		1
Polygonaceae	2	4	1	1		1	1
Primulaceae	5	10	3		1	4	2
Putranjivaceae	1	1	1
Resedaceae	2	7	4			2	1
Rhamnaceae	7	33	17	1	3	11				1
Rosaceae	5	30	7		6	15				2
Rubiaceae	31	141	37	3	17	48	22	1	1	12
Rutaceae	11	49	6	2	12	17	6			6
Sabiaceae	1	7	2			3	2
Salicaceae	8	21	6	1	5	8	1
Sapindaceae	7	13	5	2		5				1
Sapotaceae	2	21	5	1	7	5	2			1
Schisandraceae	1	1			1
Schoepfiaceae	1	3	1			2
Scrophulariaceae	2	6	4	1		1
Simaroubaceae	1	1				1
Solanaceae	6	37	14	1	11	9	2
Staphyleaceae	2	3	2			1
Styracaceae	1	10	1	2	1	4	1			1
Surianaceae	1	4	1		1	1	1
Symplocaceae	1	14	2	2	1	7	2
Thymelaeaceae	2	14	1		4	8				1
Urticaceae	3	7	4		2	1
Verbenaceae	3	16	7	1	2	4	2
Viburnaceae	1	12	6		4	2
Violaceae	1	3			1	2
Ximeniaceae	1	2	2
Zygophyllaceae	1	1					1
Total	374	1472	476	79	280	468	123	2	2	41

Note: The number of genera and species assessed is indicated, as well as the number of species in each of the Red List conservation categories.

Abbreviations: LC = Least Concern, NT = Near Threatened, VU = Vulnerable, EN = Endangered, CR = Critically Endangered, EW = Extinct in the Wild, EX = Extinct, DD = Data Deficient.

FIGURE 2. Spatial distribution patterns of the threatened tree species in Natural Mexico (delimited in grey) assessed by the authors for the International Union for the Conservation of Nature (IUCN) Red List. Yellow dots = vulnerable (VU), orange dots = endangered (EN), red dots = critically endangered (CR)

FIGURE 3. Species richness of tree species in federal protected areas (CONANP, 2021) in Natural Mexico assessed by the authors of this study for the International Union for the Conservation of Nature (IUCN) Red List. The colour scale at the right indicates the approximate number of species.

FIGURE 4. Spatial distribution patterns of the threatened tree species according to biogeographic regions following CONABIO (1997) in Natural Mexico assessed by the authors for the International Union for the Conservation of Nature (IUCN) Red List. The colour scale at the right indicates the approximate number of species.

FIGURE 5. Satellite map of the southern part of the Baja California peninsula showing the threatened tree species assessed by the authors for the International Union for the Conservation of Nature (IUCN) Red List. Yellow dots = vulnerable (VU), orange dots = endangered (EN), red dots = critically endangered (CR). The cities are included for orientation purposes.

FIGURE 6. Satellite map of the Bajío region (including (parts of the states of Guanajuato, Jalisco, Michoacán and Querétaro) showing the threatened tree species assessed by the authors for the International Union for the Conservation of Nature (IUCN) Red List. Yellow dots = vulnerable (VU), orange dots = endangered (EN), red dots = critically endangered (CR). The cities are included for orientation purposes.

FIGURE 7. Satellite map of the state of Guerrero showing the threatened tree species assessed by the authors for the International Union for the Conservation of Nature (IUCN) Red List. Yellow dots = vulnerable (VU), orange dots = endangered (EN), red dots = critically endangered (CR). The cities are included for orientation purposes.

FIGURE 8. Satellite map of the state of Veracruz showing the threatened tree species assessed by the authors for the International Union for the Conservation of Nature (IUCN) Red List. Yellow dots = vulnerable (VU), orange dots = endangered (EN), red dots = critically endangered (CR). The cities are included for orientation purposes.

FIGURE 9. Satellite map of the state of Chiapas, Mexico, and Guatemala, showing the threatened tree species assessed by the authors for the International Union for the Conservation of Nature (IUCN) Red List. Yellow dots = vulnerable (VU), orange dots = endangered (EN), red dots = critically endangered (CR). The cities are included for orientation purposes.

Of the species we assessed, the majority occurred in Mexico (1465), followed by Guatemala (337), Honduras (140), Belize (88), El Salvador (87), Nicaragua (59), Costa Rica (46), and the United States of America (17). The number of species within this region assigned to IUCN Red List categories was the following: Extinct (EX): 2; Extinct in the Wild (EW): 2; Critically Endangered (CR): 124; Endangered (EN): 468; Vulnerable (VU): 280; Near Threatened (NT): 79 and Least Concern (LC): 476, and the remaining (41) were Data Deficient (DD) (Table 1, Dataset S1). Therefore, 59.2% of the assessed species fell in one of the three threatened categories.

Following the biogeographic provinces sensu Morrone (2014), the five most species-rich provinces were, in order of importance (from here onwards, the first number between brackets is the species number, the second one is the number of threatened species [CR, EN, VU], while the third one is the percentage of threatened species with respect to the total number of species): Pacific lowlands (819/410/50.1%), Veracruz (793/418/52.7%), Sierra Madre del Sur (481/192/39.9%), Transmexican Volcanic Belt (405/132/32.6%), and the Chiapas highlands (372/181, 48.7%), followed by the Balsas Basin (364/134/36.8%), the Nearctic region (320/78/24.4%), the Sierra Madre Occidental (212/44/20.6%), and the Sierra Madre Oriental (186/56/30.1%), with the remaining provinces containing less than 100 species each.

Following the Mexican biogeographic regions sensu CONABIO (1997), the five most species-rich provinces were, in order of importance: Pacific coast (699/325/46.5%), Mexican Gulf coast (541/288/53.2%), Sierra Madre del Sur (495/216/43.6%), Transmexican Volcanic Belt (441/158/35.8%), followed by the Sierra Madre Oriental (327/122/37.3%).

The richest vegetation types with the highest number of threatened species were tropical dry forest (817 species/383 species threatened/46.8%), oak forest (735/318/43.3%), pine forest (734/345/47%), cloud forest (599/308/51.4%) and tropical rainforest (451/246/54.5%).

Protected areas versus areas impacted by anthropogenic factors

The most species-rich federal protected areas with the highest number of threatened species were the Tehuacán-Cuicatlán Biosphere Reserve (154 species/49 threatened species/31.8%), the Sierra de Manantlán Biosphere Reserve (152/46/30.3%), C.A.D.N.R. 043 Estado de Nayarit (151/32/21.2%), Sierra Gorda (148/41/27.7%), and Los Tuxtlas, (123/53/43.1%). The most diverse state protected areas were Sierra de Nanchititla (84/21/25%), Cordón Pico El Loro-Paxtal (66/26/39.4%), El Cielo (49/11/22.4%), Archipiélago de Bosques y Selvas de la Región Capital del Estado de Veracruz (43/10/23.2%) and Cerro Colorado (42/12/28.6%). The accumulative total number of tree species in federally protected areas was 3127, whereas this is 1763 in state protected areas. The accumulative total number of threatened species in both area types was 762 and 340, respectively. As a proxy for unprotected areas and the impact of anthropogenic factors, we mention here the species numbers for areas covered by important cities (277/47/17%) and agricultural, livestock and forest (plantations) management (1173/624/53.2%), according to the delimitations by CONABIO (1999).

DISCUSSION Importance of data curation for conservation aims

Distribution data for most of the species we have assessed available in global and national databases are notoriously erroneous, due to limited taxonomic curation of the accessions and/or inaccurate georeferencing, most often originating in herbarium collections. Indeed, as we are noting in our ongoing taxonomic studies (e.g. Samain et al., 2021), GBIF occurrence data have to be interpreted with great caution, in the light of a potential lack of curation of both taxonomic and distribution data. Given the immense potential of occurrence data, not only for all kinds of analyses, but especially for conservation purposes, starting with the IUCN Red List assessments, major efforts should be made for thorough curation by knowledgeable people, both taxonomists who are specialists of specific groups, and individuals familiar with particular geographic areas. The recent report from the Global Taxonomy Initiative expands on the fundamental role of taxonomy and systematics for the implementation of the Convention on Biological Diversity (Abrahamse et al., 2021), a role which is often underestimated, not in the least by the community itself, but also by myriad other instances, from institutions to governments. This project that was entirely carried out by taxonomists shows that the impact of this knowledge is invaluable and far-reaching.

The consequences of incompletely curated data on the conservation of individual species are evident, as they should truly reflect conservation units. However, these can also be observed at a general scale, as becomes clear when comparing our data with those of Téllez et al. (2020). These authors show a much higher number of occurrence points of threatened species in the Yucatán Peninsula and a lower number of distribution points of threatened species in the Mexican Pacific lowlands. In the first case, Villaseñor (2021) reported the presence of sometimes thousands of distribution points of the same species in the same locality, showing how insufficient curation may artificially increase occurrence point densities. In the second area, we added many distribution points from herbarium specimens present in the herbaria MEXU and MO, as well as from species protologues, that were not in GBIF. Amongst others, this concerned occurrence points of the family Rubiaceae, of species described in the last 20 years, and of historical collections older than 100 years. Interestingly, as we used a larger study area than that of Téllez et al. (2020), our analyses show more endemic threatened species in the Yucatán Peninsula, as these are shared with the northern and Pacific regions of Central America (see, e.g. Estrada Loera, 1991).

Evaluating endemic species assessments, general distribution patterns of threatened species and main factors impacting tree conservation statuses

Given that the number of species we have assessed is about 40% of the currently estimated tree species number of Mexico and about 50% of the number of endemic species, and all data additionally have been very carefully curated, we consider our information to be representative for trees of Mexico in specific and Natural Mexico in general. However, as we did not assess conifer and oak species, and therefore did not include these in our analyses, the patterns in the Sierra Madre Mountain chains may not be entirely reflecting the general tree conservation status in those temperate and semi-humid areas.

Conifers and oaks are very significant components of the tree flora of Mexico. Within the country oak and pine forests occur mostly in mountainous regions with temperate and semi-humid climates. These temperate forests cover about 21% of the country and include approximately 24% of the recorded flora. Unfortunately, 25% of the original temperate forests have been converted to agriculture or livestock use (Rzedowski, 1978). We can see from our analyses that approximately 45% of the trees we assessed from these regions are under threat. Nevertheless, it will be important to attempt to combine different tree conservation data sets in further studies.

Information about distribution, population status, habitat, ecology, uses, trade, threats and conservation of many Mexican endemic and near endemic trees is often limited, sometimes even non-existent, and scattered over a large number of publications and databases. As a consequence, our IUCN Red List assessments, based on very careful curation of the distribution data and on the extensive field knowledge of the last author of this paper, are generally the first publications where all this information is gathered in one place, freely accessible to all interested users.

The spatial distribution analyses following biogeographic provinces and vegetation types showed that the most species-rich provinces/types generally also have the highest numbers of threatened species. A second reason for these high numbers of both trees in general and threatened species specifically is that these flat and relatively easily accessible areas have been intensively collected by botanists, and there are also several field stations that facilitate explorations even more, such as the Los Tuxtlas and Chamela research stations of the National Autonomous University of Mexico. On the other hand, this easy accessibility has also allowed quicker deforestation for agricultural activities, as well as the installation of water dams, at their turn negatively affecting the conservation status of many arborescent species.

The main issue impacting Mexico's vegetation is the exponentially increasing demand for resources because of the tenfold population growth in the last century (INEGI, 2022). Especially water is a priority for both humans and ecosystems (CONAGUA, 2018), so the watershed management for cities and agricultural activities has negatively affected the conservation status of a high portion of the tree species we assessed.

The Mexican Pacific lowlands show the highest diversity and the second-highest number of threatened endemic tree species. This area is mainly characterised by tropical dry forest, the fourth most threatened vegetation type based on our assessments, which has been largely affected by industrial agriculture, followed by large touristic development projects along the coast (e.g. Acapulco, Mazatlán, Puerto Vallarta, Manzanillo, and Zihuatanejo, Huatulco being the only example of a more sustainable tourism project; Presidencia de la República EPN, 2014; Gobierno de Nayarit, 2017). The Gulf of Mexico area, the second most diverse zone with the highest number of threatened arborescent species, originally largely covered with tropical rainforest with the highest species diversity of Natural Mexico, and which currently is highly fragmented and consists of few relatively small remnants, has been mainly affected by oil development, generating a much higher income for the population in the whole area, which was then invested in cattle industry (Tudela, 1989), at its turn influencing the water resources in the whole area. The second most threatened vegetation type is the cloud forest, of which it was already known that about 60% of its tree species in Mexico is threatened (González-Espinoza et al., 2011). The main intrinsic factors affecting the different forest types mentioned here are the generation time and growth speed of adult trees which is very slow in tropical rainforest (200–400 years), 40–60 years in tropical dry forest, and intermediate in the cloud forest, albeit closer to that of the tropical rainforest (E. M. Martínez Salas, pers. comm., 2022).

As shown by Rodríguez Robayo et al. (2021), socio-ecological conflicts are more common in the most biodiverse areas. Some of these conflicts have resulted in serious deterioration of ecosystems, such as the Zapatist movement which has practically eliminated the Selva Negra forest in Chiapas (Expansión, 2011; Francisco & Mendoza, 2018). However, as already mentioned above, several governmental projects have also led to irreversible destruction of primary vegetation throughout Mexico, such as the National Program for Deforestation (‘Programa Nacional de Desmonte’, Moreno Unda et al., 2019) and the ‘Ley de Tierras Ociosas’ (Velázquez Fernández, 2016) in the past, and ‘Sembrando Vida’ (El Universal, 2021) at present.

Although it is clear that protected areas, at federal (Figure 4) or state level, play an essential role in maintaining species within their natural ecosystems, the number of threatened species they house is smaller than expected, and effective management is required to prevent the decline of species' populations within them. However, the highest tree diversity is not to be found in protected areas, although this does not necessarily mean that this concerns unprotected areas. A good example of this is the north of the state of Oaxaca, where inhabitants of the region have carried out forest management that has been adequate and successful since, in addition to generating jobs and resources for the inhabitants of the region, the forest area has increased in the last four decades (Redonda-Martínez et al., 2021). Indeed, few formally protected areas, but instead several zones protected by communities, house species with few records that we assessed as Least Concern, as there were no threatening anthropogenic events resulting in a threat category. Another reason for the fact that the highest tree species numbers are not necessarily present in protected areas is that many of these areas have been designated in remote areas of difficult accessibility and have therefore been relatively little investigated, as a consequence of which their floristic inventories may be incomplete.

Furthermore, climate change is shifting the distribution of species, as noted by Redonda-Martínez et al. (2021), regardless of protected area boundaries. Several recent studies have evidenced this species distribution shift for Mexican trees and forest types, especially in cloud forest and tropical dry forest (e.g. Brienen et al., 2010; Jiménez-García & Peterson, 2019; Prieto-Torres et al., 2016; Rojas-Soto et al., 2012), and there is even a more general study of 25 years ago about the impact of climate change on all forest ecosystems in Mexico, showing that the most sensitive vegetation communities are those which grow in temperate climate conditions (Villers-Ruiz & Trejo-Vázquez, 1997). The fundamental importance of conserving trees also needs to be recognised as part of broader landscape management. The importance of considering sustainable management options as part of land use policy, with alternative livelihoods for local people, is emphasised in the Red List assessments for many of the threatened tree species of Natural Mexico. The development of agricultural monocultures is one of the main threatening factors within the region. If such plantations are considered essential, mitigating factors should be emphasised such as local protection for threatened tree species and buffer zone areas of natural forest (see for example Rainforest Alliance, 2020).

Analysing our regional maps, the same recurrent issues are resulting in tree species to become threatened, although specific according to each region. The southern area of the Baja California Peninsula was a conserved region until touristic projects were developed; the highest number of threatened species can be observed in between the touristic beach cities of La Paz and Los Cabos (Figure 5). The Bajío region has been the basis for livestock (especially goats) and agriculture (starting with sorghum) for more than 400 years, and later major cities developed (Figure 6). The state of Guerrero has suffered anarchical development, and although abrupt geography has impeded industrial development, a major conglomeration of threatened species is observed near cities and mining projects (Figure 7). Agriculture has less affected the vegetation in this state, but the touristic development of the important beach cities of Acapulco and Zihuatanejo has done so. In Guerrero, we also observed that the mountainous areas have been studied more, and therefore, the recorded endemism in these areas is very high. The state of Veracruz is one of the most floristically diverse states and ranks first at the national level in vegetation loss, with less than 10% of its original vegetation more or less conserved (Ellis et al., 2011). The forests in this state since about 100 years have succumbed to agriculture and livestock activities, as well as petrol industry and population growth, probably one of the highest in our whole study area, with the most threatened species in the central area, which is also the most studied one from a floristic point of view (Figure 8). Finally, the state of Chiapas plus Guatemala has also shown a very high demographic growth in the last 50 years (Figure 9). Apart from agricultural development and coffee plantations that have destroyed both rainforest and cloud forest, the construction of several large water dams resulted in a major relocation of the population, who then went on to apply the harvesting models they were familiar with in a different kind of forest, leading to major destructions of the vegetation (Martínez, 2003).

UN Decade for Ecosystem Restoration and prospects for tree conservation in Natural Mexico

The tree flora of Mexico is of immense importance both ecologically and culturally. In preparing the conservation assessments of tree species for the IUCN Red List, the uses of individual species are recorded. This information can help support forest management and prioritise conservation actions linked to rural livelihoods (see, e.g. CONABIO, 2006).

Forest landscape restoration is recognised globally as an important means to respond to biodiversity conservation and climate change targets. The forest landscape restoration approach specifically recognises the need to both regain ecological integrity whilst enhancing human well-being. Forest landscape restoration can specifically address the decline in tree species if indigenous and especially threatened tree species are incorporated into restoration schemes. The Bonn Challenge and New York Declaration on Forests have set global restoration targets to support the UN Decade on Ecosystem Restoration (2021–2030). Mexico has made a Bonn Challenge commitment to restore 6,500,000 ha by 2030 (Stanturf & Mansourian, 2020). This provides a significant opportunity to restore populations of threatened tree species of Mexico as far as possible prioritising those with local livelihood values. As noted by Di Sacco et al. (2021), the Global Tree Assessment data provide useful information on threatened tree species for restoration schemes. They note that rare and threatened species are less likely to colonise through natural succession and should therefore be carefully reintroduced through planting at the appropriate stage of forest maturity. The recent reference work of Méndez-Toribio et al. (2018) about the restoration of terrestrial ecosystems in Mexico presents the actual state, need and opportunities for landscape-level restoration in the country.

CONCLUSIONS AND OUTLOOK

Our study showed that almost 60% of the species that we have assessed up to now are threatened, double the percentage of threatened trees at global level (BGCI, 2021), in addition to several extinct species, especially in the state of Veracruz, mainly due to agriculture, livestock, mining and tourism. Therefore, the potential impact of these assessments is wide-ranging, as they may be of interest to people seeking species-specific information, as well as the starting point for urgent conservation actions. In addition, they allow comprehensive studies to be carried out that include the estimation of extinction risk, gap analyses for conservation planning and even species reintroductions, all within a framework of land use change, climate change and landscape composition and configuration. Our ongoing research aims at defining factors that have influenced tree conservation status throughout the study area at general and fine scales, also including assessments we have not realised ourselves such as conifers and oaks, and use this information to propose areas of conservation and restoration.

ACKNOWLEDGEMENTS

We are grateful to the following colleagues who have supported our work on the IUCN Red List assessments since 2018: Emily Beech, Katharine Davies, Barbara Goettsch, Luz María González-Villarreal, Luis Hernández, Craig Hilton-Taylor, José Linares, Kate Marfleet, Rosalinda Medina, Anna Puttick, Clara Ramos, Malin Rivers and Janet Scott. Thanks are also due to the two anonymous reviewers and the PPP editorial team. We dedicate this paper to the memory of Dr. Mario Sousa Sánchez, visionary promoter and executor of large-scale floristic inventories in Mexico and Mesoamerica.

AUTHOR CONTRIBUTIONS

M.S.S., S.F.O. and E.M.M.S. designed the study, M.S.S. is the project responsible, E.M.M.S curated the distribution and taxonomic data for our IUCN Red List assessments with the help of K.M.M., A.C.D.F. and A.G.Z.C., who all prepared the conservation assessments for the IUCN Red List based on the knowledge of E.M.M.S. R.R.M., D.V.M. and F.A.A.N. prepared the conservation assessments of specific families. M.S.S. reviewed and corrected all conservation assessments before submission to BGCI, S.G.D. analysed the distribution data of our assessments and prepared the maps, M.S.S. wrote the manuscript with the help of E.M.M.S. and S.F.O. All authors contributed to the discussion, review and approval of the final manuscript.

CONFLICT OF INTERESTS

There is no conflict of interest.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.