Understanding the diversity and biogeography of Colombian edible plants

Despite being the second most biodiverse country in the world, hosting more than 7000 useful species, Colombia is characterized by widespread poverty and food insecurity. Following the growing attention in Neglected and Underutilized Species, the present study will combine spatial and taxonomic analysis to unveil their diversity and distribution, as well as to advocate their potential as key resources for tackling food security in the country. The cataloguing of Colombian edible plants resulted in 3805 species. Among these, the most species-rich genera included Inga, Passiflora, Miconia, Solanum, Pouteria, Protium, Annona and Bactris. Biogeographic analysis revealed major diversity hotspots in the Andean humid forests by number of records, species, families, and genera. The departments of Antioquia, Boyacá, Meta, and Cundinamarca ranked first both in terms of number of unique georeferenced records and species of edible plants. Significant information gaps about species distribution were detected in the departments of Cesar, Sucre, Atlántico, Vichada, and Guainía, corresponding to the Caribe and Llanos bioregions, indicating the urgent need for focusing investigation in these areas. Furthermore, a significant level of geographic specificity was found in edible plant species’ distributions between 13 different bioregions and 33 departments, hinting the adoption of tailorized prioritisation protocols for the conservation and revitalization of such resources at the local level.

Food represents the strongest form of interaction between humans and the environment. It lays at the foundation of human experience, shaping our relationship to other non-human living beings and embedding forms of intangible cultural legacy 1 . It is known that more than 7000 plants are edible 2,3 , meaning that "as a whole or their any part (roots, leaves or fruits) are acceptable for eating purpose by humans" 4 , p. 41). Many of them form part of the traditional gastronomic heritage of human populations and have the potential to support food security and develop sustainable agriculture around the world 3 . However, today almost the entire human caloric intake is made up of only ten species 5 . This incongruous trend was triggered by the green revolution 6 , which started to replace traditional landraces and wild species by a restricted assortment of modern commercial hybrids, favouring yield production 7 . The downsides of such direction were manifold. People started to diminish their interest in local edible plants while progressively decreasing their attention to the wellbeing of the ecosystems hosting them, which led to their degradation 8 . On the other hand, increasing pressure on a narrow portion of natural resources, together with unsustainable cultivation (e.g., substantial use of external agricultural inputs with impoverish the soil and damage pollinators' populations) and harvesting practices (e.g., overharvesting of a given species in the wild), resulted in a rapid depletion of the natural populations of edible plants [9][10][11] . What is more, demographic growth and increased urbanisation resulted in severe land and forest cover changes: ongoing shifts towards urban centres caused progressive land abandonment 12 , and increasing demand for cereals, oils and meat led to the conversion of natural ecosystems to pastures and croplands 13 .
In response to this trend, international policy frameworks aimed at combining biodiversity protection and sustainable development gained attention over the last decades 14 . As reported by Borelli et al. 15 , efforts have increased to revitalize and promote the use of "orphan crops" and wild edible plants. Over the last few decades, more than fifty years after the green revolution, Neglected and Underutilised Species (NUS)-defined as "useful plant species which are marginalized, if not entirely ignored, by researchers, breeders and policy makers" 16 , p. 9)-have been proved to hold crucial importance for building sustainable livelihoods and mitigating environmental deterioration 15,17 , Ulian et al. 2020 3 ). They also hold critical biocultural values, as they are linked to local agricultural systems and culinary traditions and practices, symbolizing the organic relationship between nature and culture. Growing evidence has demonstrated that the value peoples give to local plant resources can play a crucial role for their engagement in conservation and sustainable management 18,19 . "Conservation-through-use" approaches, aimed at encouraging nature conservation through the sustainable use of its resources, are increasingly being applied in conservation programs worldwide (Dulloo et al. 2017 20 , Oliveira Beltrame et al. 2018 21 ).
Colombia is one of the world's "megadiverse" countries 22,23 , hosting 10% of the global biodiversity 24 and bringing together an unequalled number of distinct natural ecosystems and human cultures. However, despite its great biocultural richness, Colombia is nowadays characterized by widespread poverty, with more than 54% of its population suffering from food insecurity 25 . What is more, new land uses are now causing habitat destruction, driven by export oriented industrial agricultural policies and unsustainable market conditions 26,27 . While local communities had historically benefited extensively from local plant diversity (Rivas et al. 2010 28 ), native plants consumption in the country has substantially decreased over time (López Diago & García 2021 29 ). In fact, following the global trend, over the course of the past decades local edible plants have been rapidly replaced by high yielding commercial varieties 15 and have become full-fledged NUS. Nevertheless, many underutilised species, although being rather unknown outside of the country, hold the potential to address environmental degradation, while creating sustainable livelihoods and boosting Colombian green growth 3,7 . Considering these circumstances, investigating and understanding Colombian Gastronomic Ethnobiology-the study of the complex interactions between people, food, and their environment 30 -acquires crucial importance for the formulation of targeted and effective conservation and sustainable development activities.
The first comprehensive documentation of Colombian useful plant diversity was conducted by Pérez-Arbeláez (1978), who catalogued 1771 species, including hundreds of edible species. Subsequently, Romero- Castañeda (1991) 31 29 , agricultural studies have been carried out on less than 20% of Colombian wild fruits. This constitutes a significant limit to the formulation of conservation-through-use strategies. Moreover, most ethnobotanical studies, being focused on a narrow portion of geographic locations, are not geographically representative of the entire Colombian territory. Indeed, the academic coverage of this topic leaves significant geographic gaps, an example of which is the almost total deficiency of research carried out in the Caribbean region 29 .
The Useful Plants and Fungi of Colombia (UPFC) project-started in 2019 by the Royal Botanic Gardens, Kew-aims at turning the potential of Colombian biodiversity into an economic resource for improving local livelihoods and food systems of impoverished communities. Locating itself within this fascinating and yet complex research framework, the present work aims to provide detailed information on the composition and distribution of the Colombian edible flora. We catalogued Colombian edible plant diversity and displayed a first prototype of its biogeographical distribution. Furthermore, we present useful information and knowledge for driving future efforts towards edible plants revitalization and conservation-through-use in Colombia.

Materials and methods
Methods used for this study followed the approach employed by Diazgranados et al. 2 for compiling the World Checklist of Useful Plant Species (WCUP) containing key taxonomic and ethnobotanical information on 40 292 species. Among all the plants classified as useful, 7039 were classified under the category of "Human Food", following the Economic Botany Data Collection Standard 39 . The same taxonomic backbone of the World checklist of useful plant species was adopted here to reconcile taxonomically species coming from a combination of 10 new datasets and publications, both Colombian and international, in addition to the data coming from the Annotated Checklist of Useful Plants of Colombia 40,41 , New datasets were checked and cleaned in R 4.1.0 42 . Taxon names were reconciled to POWO 43 and, when no such data were available, to Tropicos 44 using "Plyr" and "Dplyr" packages 45,46 . Higher taxonomy information was obtained from the Global Biodiversity Information Facility 47 . Data on species edibility was retrieved from the cleaned datasets. Lastly, the number of Colombian edible NUS was obtained by checking the final dataset of Colombian edible species against FAO's global census of agricultural crops 48 .
To contemplate both ecological and political factors affecting biodiversity distribution, we employed Colombian departments (32 departments and 1 capital district), bioregions (13) and cells (10 × 10 km 2 ) as units for the present biogeographic analysis. We utilised a map of bioregions produced by Bystriakova et al. 49 , which combines the map of Terrestrial Ecoregions of the World (Olson et al. 2001 50 ) with the five continental ecoregions in Colombia, resulting in 13 distinct units (Fig. 1). Unique georeferenced records of herbarium specimens for edible species were downloaded from GBIF (2021) through "RGbif " package 51 and cleaned using "ShinyC-Cleaner" package 52 in R 4.1.0 42 . Filters for removing occurrences recorded within urban centres, in the sea, within institutions (e.g., botanic gardens, ex-situ repositories), and in the centre of Colombia (i.e., centroid points) were applied. Moreover, latitude and longitude points with less than three decimals were removed, as well as occurrences with equal latitude and longitude and occurrences with either latitude or longitude equal to zero. Grid analysis to measure species richness was carried out using "rgdal" package 53 , "raster" (Hijmans and van Etten 2012 54 ), and "sp" 55 . The complete resulting datasets are accessible in Figshare (https:// figsh are. com/s/ cf5c1 9832a d4fd1 695d7). ArcGIS pro 2.8.1 was employed to carry out biodiversity quantifications. SDMtoolbox 56 -a python-based toolbox for spatial analysis-was used to obtain biodiversity metrics such as species richness (i.e., sum of species per cell), weighted endemism (i.e., sum of the total number of cells each species in a grid cell is found, emphasizing areas rich in species with restricted distributional ranges), and corrected weighted endemism (i.e., weighted endemism divided by the total number of species in a cell, emphasizing areas rich in species with restricted ranges, but that are not necessarily species-rich), employing a geographic resolution of 0.1 degrees (~ 10 × 10 km 2 ).

Edible species distribution & diversity hotspots.
Across the country, 172 species (4.5%) are found in more than twenty departments, only fifteen species (0.3%) are found in more than thirty departments, and only one species-Eleusine indica, an introduced grass from the Tropical and Subtropical Old World-is found in thirty-three out of thirty-three departments. Only four of the species present in more than thirty departments (i.e., Gynerium sagittatum (Aubl.) P.Beauv., Lasiacis procerrima www.nature.com/scientificreports/   (Fig. 3). Grid analysis revealed several hotspots for edible species diversity across the country (Fig. 4). Both the Species Richness (SR) and Weighted Endemism (WE) analysis emphasised the northern and north-western Andean region as crucial repositories of edible species (Fig. 4A, B). In particular, the highest SR was recorded between the areas of Antioquia, Boyacá, Cundinamarca and Caldas, which extend across three distinct bioregions: the Andean dry forest, humid forest, and páramo. In contrast, the Corrected Weighted Endemism (CWE), emphasising areas that are characterised by a high proportion of species with restricted distributional ranges, displayed a more scattered distribution of numerous restricted hotspots (represented by dark brown cells in Fig. 4C). A considerable density of narrow-distribution edible plants was recorded in the Amazonian bioregions, in the departments of Amazonas, Caquetá and Vaupés, as well as in the departments of Guainía, Nariño and La Guajira, in the proximity of Sierra Nevada de Santa Marta.

Discussion
Richness and taxonomic diversity of Colombian edible plants. This study provides a comprehensive overview of the edible plant diversity present in Colombia. Results show that the diversity of edible plants in Colombia is remarkable. Half of the species with reported human uses present in the country are edible, for a total of 3805 food plants. This figure far exceeds the previously known numbers for Colombia. Furthermore, this number acquires crucial significance if compared to the data showcased by the WCUP 2 : in fact, the proportion between useful species and edible species at the global level (respectively, 40 292 and 7039 in number) is remarkably lower than the one specific to Colombia. While, according to the Economic Botany Data Collection Standard 39 , at the global level only the 17.4% of useful plants have been recorded as "Human food" 2 , in Colombia this percentage raises up to 53.6%, making this country a global reservoir for edible plant diversity. Nevertheless, only 2457 species are known to be edible in the country, meaning that more than one third of the total diversity of edible plants present in Colombia are unknown or neglected from a gastronomic perspective. This data stresses the untapped potential of Colombian edible plant diversity, and creates an interesting space for future research on the possible reasons behind such gap.
The factors responsible for this notable diversity are many. Among these, it is worth mentioning the unrivalled ecosystem diversity of Colombia: according to the Institute of hydrology, meteorology, and environmental studies (IDEAM, 2017) the country hosts 93 general ecosystems, including 15 coastal ecosystems, as well as 42 terrestrial ones. Some of them are regarded as of global conservation importance, such as the páramos, the Andean forests and the tropical rainforest of the Chocó department 49 , due to the great diversity of life forms they host, as well as their complexity and sensitivity to current environmental changes. What is more, such ecosystem   58 . Over the past few decades, there has been increasing recognition of the inextricable link between cultural and biological diversity, and of the countless modalities in which they mutually affect each other 59 . Considering this, Colombia provides a concrete example of how these two dimensions go hand in hand.
Results of the present taxonomic analysis are partially in line with what Diago and García 29 found: according to these authors, the richest families of edible wild fruit species in Colombia were Fabaceae, Arecaceae, Passifloraceae, Sapotaceae, Moraceae and Melastomataceae. If we consider the totality of Colombian edible plants, results concerning the most important families (i.e., Fabaceae, Asteraceae, Poaceae, Arecaceae, Rubiaceae) differ quite significantly. However, if we only take native edible species into account, results align, highlighting Fabaceae, Moraceae and Melastomataceae as some of the most significantly rich families in edible species. This inconsistency highlights the fact that in Colombia most of the introduced species come from a small number of families, including Fabaceae, Poaceae, Arecaceae. Not surprisingly, these families are among the most important ones for the number of cultivated species, both at the national and at the global level 3 . As the predominance of some introduced cultivated species suggests, the Colombian population has progressively moved away from native foods over the last decades 29 . Numerous native species became NUS due to the progressive transformation of Colombian peoples and their traditional cultural heritage following the advent of European conquerors 60 and the country's gradual introduction into the globalized trade market of natural products 27 . Today, Colombian gastronomy reflects this country's complex history, with European and African influence from the times of the colonisation, mixed with the rich indigenous cultural heritage and local biodiversity 61 .
According to 62 , despite the unrivalled richness of edible plant species characterising the Colombian territory, today 90% of the natural ingredients marketed in the country are imported. In response to these issues, the Colombian government has recently been investing resources in developing a bioeconomy strategy with the aim of facilitating the future green growth of the country. The Decree 63 , published by the Departamento Administrativo de la Función Pública (Governmental department of Civil Service), provides the first detailed regulatory and legal framework for the sustainable use of non-timber wild edible species. This gives local people the opportunity to commercialise NUS for the first time within a regulated system. Against this background, and based on the abovementioned figures, more in-depth ethnobotanical investigation is needed to identify priority species for revitalization and conservation-through-use initiatives. Numerous examples exist in the literature of how investigations aimed at understanding the relationships between human populations and the natural resources they have traditionally coexisted with and depended on for their subsistence and cultural expression can play an important role in the process of defining conservation priorities 64,65 . Conservation cannot ignore the relationship between the human and non-human dimensions of nature: the daily lives of people following traditional lifestyles are closely connected to the local natural environment (Pei et al. 2020 66 ), and the role and perception of a given plant resource within a given socio-cultural context can be of great importance for engaging people in conservation activities and enhancing conservation success 9 .  (Fig. 5), Lupinus mutabilis Sweet, Amaranthus caudatus L., Theobroma grandiflorum (Willd. Ex Spreng.) K. Schum. (Fig. 5)), based on their nutritional properties and great resilience and adaptability to a wide range of environmental conditions [67][68][69][70][71][72] . Here, we want to explore and highlight the four genera that in the present analysis were characterised by having the highest edible species richness and diversity, and conceivably holding great culinary potential and versatility: An example is the genus Passiflora Kunth-known as Granadilla de monte, Granadilla silvestre, Chulupa de mico or Gulupa 73 -are wild harvested for own consumption or sale in local markets. There is substantial morphological variation within the genus (e.g., fruits and flowers' colour and shape). The fruits can be eaten raw or cooked, or even used to make drinks, as in P. antioquiensis (Fig. 5) 73 . The pulp is very aromatic, and flowers and leaves are also edible sometimes. In the case of P. foetida L. (Cincollaga, Cocorilla) (Fig. 5), leaves are cooked and used as an ingredient in soups. The pulp of fruits is very variable in terms of taste: it can be sweet (e.g., P. ligularis), juicy and acid-flavoured (e.g., P. coccinea Aubl.) or aromatic and mildly biting (e.g., P. cumbalensis (H.Karst.) Harms).
The genus Solanum can also be considered particularly important from an economic perspective. In addition to Solanum lycopersicum L. (Tomato), and Solanum tuberosum L. (Potato), this genus has 61 edible species, many of which today have less economic importance, and are known, cultivated and consumed exclusively locally. In Colombia, these are especially present in the Andean region and are still largely unexplored from a taxonomic, agronomic and bromatological perspective. Examples include S. cajanumense Kunth, a fast-growing evergreen shrub whose golden-yellow fruits are eaten fresh when fully ripe; S. capsicoides All. (Fig. 5), whose poisonous fruits can be eaten when roasted or cooked; and S. pectinatum Dunal-commonly known as Huevo de gato, Naranjuelo or Toronja 73 -whose pale orange flash, characterized by a sweet-acidic flavour, is delicious when cooked with sugar (Food Plants International, 2021 74 ).  (Fig. 5)-also known as Churimo or Guabo-is the best known and most consumed species. Both the seeds and the white, jelly pulp surrounding them can be eaten. The pulp is characterised by a sweet and highly aromatic taste 76 and it is usually eaten raw. The seeds are eaten cooked, usually boiled or roasted, as in the case of I. ilta T.D.Penn (Guamo de semilla). When immature, they can also be eaten raw, blanched and salted, and added to salads (Food Plants International, 2021 74 ). Inga trees also hold great environmental value: they are commonly placed in coffee or cacao plantations to provide shade to the surrounding environment 77,78 .
Finally, Bactris counts 28 edible species in Colombia, and all of them are native. However, only B. gasipaes Kunth (Chontaduro, Cachipay, Pipire) (Fig. 5) is currently cultivated. Fruit of most species are inedible raw 79 . They are usually boiled in salted water for thirty to sixty minutes and eaten as a vegetable (Food Plants International, 2021 74 ). The pulp is characterised by a nutty flavour and a floury texture, as well as remarkably high nutritional properties due to their great protein and carbohydrate content. Fruits can also be made into a flour and baked into bread, cakes, and other processed foods 79 , Food Plants International, 2021 74 ). Seeds can be consumed raw 80 , as nuts, as in the case of B. major Jacq. (Lata arroyera, Albarica, Uvita de tigre) and B. gasipes. The palm heart of some species (e.g., B. riparia Mart., B. corossilla H.Karst.) is also eaten raw, in salads, or cooked. Finally, the fruits of B. guineensis (L.) H.E.Moore (Lata de corozo) (Fig. 5) can be fermented and used to produce a drink, which in Colombia is known as "Chicha de corozo" 81,82 . Distribution of occurrence records. Results clearly show how occurrences records are unequally distributed among Colombian edible species. This constitutes a substantial limitation to the study of the distribution patterns of the Colombian edible flora and can be mainly attributed to the lack of scientific coverage of some areas of the country 49 . In fact, inequalities in the distribution of georeferenced records does not only apply to species diversity but also to entire Colombian departments and bioregions. Linear regressions revealed a significant correlation (p < 0.001) between species richness and number of georeferenced records within Colombian departments, making it possible for us to argue that while some of the striking differences in species richness between Colombian departments (Fig. 2D) can be attributed to environmental and anthropogenic factors, other may be the result of the lack of adequate on-site investigation.
These results are consistent with what 49 reported on the distribution of useful plants across Colombian bioregions: while exhaustive sampling in areas such as the Andean region allows the comprehensive understanding of the edible flora of the Andean humid forest, the Andean dry forest and the Páramo bioregions, the regions of Caribe and Llanos (Cf. Fig. 1), to date, remain largely unexplored. This represents a substantial gap to the reliability of the present figures.
As 40 point out, both historical and security factors may have contributed to an unbalanced sampling effort across these regions. The internal conflict that Colombia has witnessed over the past six decades has prevented scientists from carrying out field work in several parts of the country, including the departments of Cesar, Norte de Santander and Arauca 40 . Both the Llanos and Caribe areas are known for violent episodes and illegal activities, explaining the scarcity of scientific studies there, and the scarcity of recorded edible plant species. Therefore, entire bioregions, located in such under-surveyed areas of the country, are likely to be underrated by the present results. These include the Caribbean dry forests and páramo, the humid and dry forests of the Llanos, the Amazonian humid forest and the savannas of the Orinoquia region.

Biogeographic patterns of Colombian edible plants. A significant level of geographic specificity was
found in the distribution of Colombian edible species across 33 departments, with most of the species only recorded within five of them or fewer, stressing the remarkable biological difference between the numerous bioregions and ecosystems across the country. This figure is confirmed by the CWE analysis (Fig. 4C), which instead of revealing few clear hotspots for narrow-distribution edible plants, such as in the case of SR and WE (Fig. 4A, B), displays numerous cells (10 × 10 km 2 ) of high endemic value in several areas of the country. In fact, areas that were not highlighted by the SR metric, such as the regions of Amazonia, Llanos and Caribe, are shown to be of great conservation importance due to the presence of unique species. This suggests that, although bioregions such as the Andean humid and dry forests have recorded the highest amount and diversity of edible species, especially within departments of Antioquia and Boyacá, such species were not characterised by restricted distribution ranges. In fact, these regions contain the main agricultural areas of Colombia and are therefore characterised by the presence of cultivated species with broad distributions. Considering this, the CWE metric is key to emphasise the hidden biological importance of unexplored Colombian bioregions, as well as to recognise the limitations of our current understanding of the biogeographic distribution of local edible plants.
Species richness. The Colombian Andes form part of the Tropical Andes, which extend across the north of Chile, Argentina, Bolivia, Peru, Ecuador, Colombia and Venezuela for over 1.5 million km 2 , between the latitudinal range of 11° N to 30° S. They are characterised by an elevational range of approximately 500 to 6000 m a.s.l. (Bax and Francesconi 2019 83 ). According to Meyers et al. (2000) 102 , the Tropical Andes support approximately 45 000 plant species, with nearly half of them being endemic to the Andean ecoregion. However, due to increasing human-driven alterations of natural ecosystems, as well as to the progressive impact of climatic variations, ecosystems such as the Tropical Andes are nowadays recognised as one of the most critically threatened ecoregions in the tropics 84 . This constitutes a particularly serious hazard to the native endemic edible species inhabiting the region, characterised by extremely specific habitat needs. In light of this, Colombian Andes are expected to lose www.nature.com/scientificreports/ a significant proportion of their native plant diversity by reason of environmental degradation 84,85 . This would represent a significant loss not only for the biological heritage characterising the region, but also for its sociocultural one, embedded in local traditional agricultural practices and gastronomy. Andean bioregions and the edible flora characterising them must therefore be protected and further investigated from an ethnobotanical perspective, in order to understand the mechanisms and socio-cultural practices underlying their remarkable gastronomic heritage.
Corrected weighted endemism. A significantly high density of narrow-distribution edible plants was recorded in the northern part of the Huila department. Huila may be considered one of the richest regions of Colombia in terms of plant biodiversity, due to its great variety of ecosystems, from páramos on top of the mountains to extensive areas of tropical dry forest and rainforest. About 120 000 hectares of the department are localised within the páramo biome, with elevations ranging between 2900 and 5000 m a.s.l. Notably, the area has been described as one of the richest in the country in terms of diversity of the genus Passiflora 86 . Another interesting hotspot for its unique edible flora can be spotted in the department of La Guajira (Caribe), in the vicinity of the Sierra Nevada de Santa Marta (SNSM), a UNESCO Biosphere Reserve. According to Durán-Izquierdo and Olivero-Verbel 87 , the SNSM can be considered as the most irreplaceable nature reserve in Colombia, due to its extraordinary ecological diversity. However, today SNSM's precious biomes, as well as the ecosystem services they provide, are increasingly being jeopardized by anthropogenic activities such as mining, agricultural expansion and tourism 87 . High concentrations of edible species with restricted distribution can also be found in the humid forests of Amazonia and Llanos, encompassing the departments of Amazonas, Caquetá and Guainía. These are some of the least accessible areas of the country due to the poor road coverage 49 . High species density cells occur especially around four national parks: Parque Nacional Natural Yaigoje Apaporis, Parque Nacional Natural Cahuinarí, Reserva Nacional Natural Nukak and Reserva Nacional Natural Puinawai. These places are focal points of biocultural diversity for the whole country, and over the years they all have witnessed, and to a large extent still witness, effective examples of indigenous resistance against mining expansion 88,89 . According to Bystriakova and colleagues 49 , only 8% of the Amazonian humid forest has been displaced by human activities. However, despite the institution of protected areas, deforestation and land use change still represent major threats to these sites. The insufficient sampling effort in these regions, together with the increasing pressure on natural resources, could result in the loss of a consistent portion of edible plant diversity which to date has still not been fully documented. One final hotspot is visible in the Nariño department, across the humid forests of the Pacific and the Andean regions. In particular, the highest concentration of edible species is scattered along the banks of the Mira River. In addition to such bioregions, the department also includes Andean humid and dry forests, páramos, and mangroves. Such biological diversity corresponds to an equally rich cultural diversity, with 17.8% of its population being Afro-Colombian and 15.7% belonging to various indigenous peoples 57 . Therefore, Nariño can be regarded as a remarkable hotspot for the Colombian biocultural diversity.

Food security & food sovereignty.
The uneven distribution of occurrence records among edible species-with introduced cultivated species registering outstanding numbers of georeferenced records compared to native ones-closely reflects the process of establishment and expansion of large-scale plantations of commercial species, primarily destined for the international market. A concrete example is provided by Hurtado-Bermudez and colleagues 25 , whose work examined the increasing spread of sugarcane plantations (Saccharum officinarum) across the regions of Magdalena and Cauca. According to the authors, such expansion has led to increasing land dispossession, farmers displacement and food insecurity in the regions 25 . Indeed, Colombia reports one of the highest rates of inequalities regarding land ownership 90 , which often goes hand in hand with decreasing food security and sovereignty. In 2015, the prevalence of food insecurity in rural households in Colombia was 54.2% 91 . Moreover, targeted quantitative examination of indigenous and Afro-Colombian households revealed a much higher prevalence of food insecurity compared to national figures, ranging between 70 and 85% 25 . Studies demonstrate how large-scale industrial agricultural systems have negatively impacted Afro-descendant and indigenous peoples from a cultural and economic perspective, causing significant socio-environmental transformations, as well as progressive loss of autonomy over their territory and consequent impoverishment [92][93][94] . In response, we stress the potential of Colombian edible NUS for tackling these issues. We advocate the need to build locally controlled food systems, rooted in the local environment, culture and traditions, and the urgent need to create new sustainable livelihoods for local peoples, based on the cultivation and commercialisation of native edible plants. NUS are highly promising resources for agriculture, novel products and nutritional improvement 15 . It is therefore vital to promote new agricultural models that revalorise their use and provide an alternative approach to the spread of monocultures and intensive farming.
An example of best practice: the Guáimaro (Brosimum alicastrum Sw.). The Guáimaro (Brosimum alicastrum Sw.), also known as Caucho, Lechero and Sande 73 is a wild edible NUS belonging to the Moraceae family, native to tropical dry forests in Mesoamerica and the Caribbean 95,96 (Fig. 6). Due to its great environmental, social and economic potential, the Guáimaro was identified as a priority species by the Useful Plants and Fungi of Colombia (UPFC) project in one of its three pilot areas in the country: Becerríl (Cesar). Besides being an important ecological indicator of the health of the forest, this species is of great importance for the protection of soil and water bodies, as well as for the feeding of wildlife 95 . What is more, the Guáimaro is characterised by great nutritional qualities, such as a high carbohydrates' content and antioxidant activity 97 , as well as a great culinary versatility. The seed (Fig. 6E, F)  www.nature.com/scientificreports/ the original Guáimaro forest cover remains 99 . In the area of Becerril, main causes of disappearing of this species include coal mining, sowing of African palm for oil extraction and bovine and ovine extensive farming 99 . Promoting its revitalisation through sustainable consumption and commercialisation practices in the community of Becerril has proven to be an effective tool for ecological conservation and forest restoration: a community-based facility for local processing of the Guáimaro fruits was established (Fig. 6D), and direct sale of the processed nuts increased in local markets. Additionally, commercial connections were built between farmers in Becerril and restaurants across the country through UPFC and local partnering NGOs (Fig. 6G), giving rise to a new demand for this NUS and increasing its perceived value. Finally, in addition to economic and environmental benefits, increase in the consumption of Guáimaro-based flour in traditional preparations such as arepas, empanadas, soups and enyucados, has the potential to strengthen the food security of the community.

Conclusion
By disclosing the richness, diversity, and potential of Colombian edible plant diversity, and identifying current knowledge gaps at the geographic level, the present analysis constitutes a strong empirical basis for directing further research efforts targeting least explored areas of Colombia. The in-depth characterisation of Colombian edible plant resources is important to achieve their effective protection, to guarantee their survival and to encourage their recovery and valorisation. This process requires the joint forces of numerous disciplines, ranging from taxonomy, biogeography, ethnobotany and bromatology, together with the generation of more complete and detailed information on the population size, distribution range and threats monitoring of species 100 . This study has contributed to the preliminary characterization of edible plant resources in Colombia both from a taxonomic and biogeographic perspective. At the taxonomic level, Colombian edible plants cover an unrivalled variety of families, genera and species, many highly localised. Thanks to its unique and diverse natural ecosystems, as well as to their exhaustive sampling, the Andean region scored the highest number of edible species. On the other hand, regions like the Amazon, the Caribbean, and Llanos still remain poorly explored from a scientific standpoint and should therefore be prioritised for future, focussed research. www.nature.com/scientificreports/ Today, the preservation of Colombian NUS and the encouragement of their use are more crucial than ever. After more than six decades of internal conflict, the country is currently going through fast changes, which will determine the fate of its natural resources. Agricultural expansion and urban development are leading to deforestation and habitat loss, resulting in unprecedented levels of biodiversity erosion. Local NUS hold great potential for supporting local livelihoods and developing a bioeconomy based on the sustainable use of local natural resources. Therefore, further targeted ethnobotanical, bromatological and agricultural studies are urgently needed to achieve the full characterisation of these resources, and direct future prioritisation efforts toward their revitalisation and conservation-though-use. www.nature.com/scientificreports/ Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.