The Analysis of Italian Plant Agrobiodiversity Databases Reveals That Hilly and Sub-Mountain Areas Are Hotspots of Herbaceous Landraces

Landraces are an agri-food and historical-cultural heritage but are undergoing losses worldwide. Italy is taking action to counteract this problem by following European guidelines. One of the most important measures is the Agrobiodiversity National Register (ANR), but 12 Italian regions currently appear without any landraces and around 80% of the landraces listed are trees, with less detailed data on herbaceous species. The aim of this study is to investigate the situation for Italian herbaceous landraces preserved on farms (in situ) by merging and analyzing data contained in the main databases on plant agrobiodiversity in Italy. Data were georeferenced, organized by botanic families and Italian regions, and analyzed by GIS and R. A total of 1615 herbaceous landraces were found (versus the 416 recorded in the ANR). Poaceae, Fabaceae, and Solanaceae together comprise 70% of all herbaceous landraces and are mostly preserved/grown in areas between 150 and 800 m a.s.l. Some hilly and sub-mountain areas of the Apennines and the Alps are hotspots of herbaceous landraces due to anthropic and environmental factors. The results of this research will be useful to enrich the ANR and trigger actions of characterization, conservation, and promotion of these


Introduction
The safeguarding of agrobiodiversity is an extension of the concept of biodiversity conservation that refers specifically to the varieties/races of plant, animal, and microbe species of agricultural interest, as well as crop wild relatives. The latter have contributed to crop domestication for millennia, providing important genetic resources adapted to the environment and climate changes and can represent a valuable resource for innovative lower-input agricultural systems [1]. In the past, there were many more landraces, understood as dynamic populations of cultivated plants that have a historical origin and distinct identity and lack formal crop improvement, as well as often being genetically diverse, locally adapted, and associated with traditional farming systems [2]. Landraces constitute unique genetic resources for genetic crop improvement programs [3][4][5] and a source of food diversity available to humans and other living beings [6].
These genetic resources have undergone (and are undergoing) heavy losses over the last decades [7,8]. The Food and Agriculture Organization of the United Nations (FAO) estimates that about 75% of global agrobiodiversity has been lost over the last century and that three quarters of food worldwide is produced by only 12 plant species and five animal species [7][8][9][10][11][12]. This loss represents a serious problem that has prompted governments at global and local levels to take immediate action: in recent decades, there have been several Italy, the first genebank was set up in Bari [1] in the 1970s, followed by many more under the National Research Council (CNR) and Council for Research in Agriculture (CREA).
Although genebank accessions could fulfil the conditions set by the EU directive for conservation varieties, their lists could be completed by a germplasm that is not part of the government genebank system and is actively managed by farmers and gardeners organized in seed savers' associations or other individuals engaged in the development of informal seed supply systems [37].
In Bulgaria, a study was conducted on local horticultural cultivars and populations to summarize the information available for the germplasm of local accessions collected from their typical growing regions, resulting in a collection of 2057 accessions and bases comprising important phenological, morphological, and economic traits, allowing their rational use in breeding and practice [38]. In Germany, a project started in 2007 for the development of a federal state reporting and documentation system for in situ (CWR) and on-farm (landraces) to implement the National Inventory on PGRFA (PGRDEU), containing passport data for more than 150,000 accessions of German ex situ holding institutions [37]. Further, in other continents, alliances of local and national governments, research centers, and farmers' associations that preserve landraces in situ proved fundamental to create an inventory of traditional varieties and prevent the genetic erosion of landraces [39].

Materials and Methods
The herbaceous landraces inventory was elaborated following the considerations of Maxted et al. in [40]. In detail, for the purposes of creating a complete and updated inventory of herbaceous landraces, the following inventories were considered: • The Agrobiodiversity National Register provided to the research team by MiPAAF   As regards Italy, even if, as explained above, the Agrobiodiversity National Register  currently contains poor data on herbaceous landraces, other open sources of data, such as  inventories produced by MiPAAF, universities and research centers, or foundations for the  conservation of agrobiodiversity such as Slow Food, can provide a considerable amount of  information on Italian herbaceous landraces. The aim of this study was to collect as much detailed and complete information as possible regarding herbaceous landraces in Italy. This target was pursued by merging and analyzing the data contained in the main databases of cultivars (preserved on-farm, in situ). Specifically, data were analyzed considering the appurtenance of each landrace to plant families and their distribution throughout the Italian territory, focusing on altitudinal and regional distribution, to allow the identification of herbaceous agrobiodiversity hotspots. The families with the highest number of landraces were further investigated by analyzing the altitudinal distribution and allocation in the main genus/species. This research was supported by the Department for Regional Affairs and Autonomies (DARA) of the Italian Presidency of the Council of Ministers to implement the data of the Agrobiodiversity National Register and launch actions to promote these resources.

Materials and Methods
The herbaceous landraces inventory was elaborated following the considerations of Maxted et al. in [40]. In detail, for the purposes of creating a complete and updated inventory of herbaceous landraces, the following inventories were considered: The above-mentioned inventories and data sources were merged, eliminating overlapping data, and considering only herbaceous plants. Where possible, landrace data were checked so that they corresponded to the characteristics stated by [2]. To understand if the landraces were used for crop production or just preserved/grown by custodian farmers, a web survey regarding cultivation in agricultural holdings was performed for landraces for which no data were present in the databases analyzed.
Data were organized in Excel sheets with different fields: common name/names of the landraces, species/genus, family, elevation, latitude, longitude, municipality, region, cultivator/conservator (farmer/hobbyist and/or farm). Data were organized by botanic families and regions and georeferenced. Geographic coordinates (latitude and longitude) and elevation were acquired through Google Earth and/or Google Maps from the address of the private grower/association/municipality or farm of the specific landrace obtained from the data sources mentioned. Landraces were sorted according to multidimensional scaling (MDS) analysis in order to highlight the main geographical variables that differentiated the samples. MDS was performed using Euclidean distance and considering the latitude, longitude and elevation of each landrace. MDS was performed using the "vegan" package of R 3.5.2 software [50].
Distribution maps of the landraces were created by ArcGIS 10 software (®Esri, Redlands, CA, USA). The same software was used to identify hotspots by the kernel density method [51,52] using the spatial analysis tool. The elevation feature was displayed through boxplots using R software [50]. Boxplots were created by splitting the landraces into families and, for the most numerous (more than 200 landraces), a focus on the main genus and/or species was realized.
The scientific name of each taxon is in accordance with the Portal of Flora of Italy 2020.1 (http://dryades.units.it/floritaly/index.php).

Results and Discussion
The survey showed that the total number of herbaceous landraces (obtained merging the data of the inventories mentioned in Material and Methods) is 1615, while the National Register currently mentions only 416 varieties, that is, about 26% of the total varieties observed. This means that the total number of herbaceous Italian landraces identified could increase by about 290%. This information highlights the need to merge all existing databases into the Agrobiodiversity National Register. Although they belong to the same institution (MiPAAF), the Italian Traditional Agri-food Products ("Prodotti Agroalimentari Tradizionali"-PAT) list and the National Register have not yet been merged.
The catalogues created by the Perugia and Milan Universities provided a considerable amount of data, testifying to the success of projects led by research centers, since the survey activity of both was financed by regional, national, or supranational (EU) funds. However, apparently there was a gap in communication towards regional and national institutions since, as mentioned in the introduction, Lombardy (the region in which the University of Milan is located) has not yet communicated any varieties to the bodies in charge of the creation of the National Register. The most complete Italian inventory of landraces can be considered the one derived from the "PGR Secure project" (www.pgrsecure.org) for which the Department of Applied Biology of the University of Perugia compiled "The First Italian Inventory of In Situ Maintained Landraces" [41]. The inventory reported data until January 2013 and accounted for over 20% of the data collected in this study. Furthermore, the inventory of the Centre of Applied Studies for the Sustainable Management and Protection of Mountain Areas (CRC Ge.S.Di.Mont.) can be considered the most complete for the Alpine center-eastern area [30,[46][47][48][49].
In Figure 2, the data collected throughout the Italian peninsula are split into families, of which Poaceae, Fabaceae, and Solanaceae are those containing the greatest number of landraces. These three families together account for 70% of all herbaceous landraces, while the remaining families comprise 30%. cure.org) for which the Department of Applied Biology of the University of Perugia compiled "The First Italian Inventory of In Situ Maintained Landraces" [41]. The inventory reported data until January 2013 and accounted for over 20% of the data collected in this study. Furthermore, the inventory of the Centre of Applied Studies for the Sustainable Management and Protection of Mountain Areas (CRC Ge.S.Di.Mont.) can be considered the most complete for the Alpine center-eastern area [30,46,47,48,49].
In Figure 2, the data collected throughout the Italian peninsula are split into families, of which Poaceae, Fabaceae, and Solanaceae are those containing the greatest number of landraces. These three families together account for 70% of all herbaceous landraces, while the remaining families comprise 30%. This result is coherent with inventories and data collection in other areas of the world. For example, 7 landraces of beans (Phaseolus spp.) and 26 landraces of corn (Zea mays L.) were found in the mountain area of the Arkansas Ozarks (USA), with these two species representing the most numerous plant species among the 39 considered [53]. Likewise, 101 landraces of bean (Phaseolus spp.), 44 of tomatoes (Solanum tuberosum L.), and 28 of corn (Zea mays L.) were found in the Appalachians (North America). Beans, tomatoes, and This result is coherent with inventories and data collection in other areas of the world. For example, 7 landraces of beans (Phaseolus spp.) and 26 landraces of corn (Zea mays L.) were found in the mountain area of the Arkansas Ozarks (USA), with these two species representing the most numerous plant species among the 39 considered [53]. Likewise, 101 landraces of bean (Phaseolus spp.), 44 of tomatoes (Solanum tuberosum L.), and 28 of corn (Zea mays L.) were found in the Appalachians (North America). Beans, tomatoes, and corn are therefore the most important landraces being maintained across the American highlands [53]. In Portugal, according to the Second Report on the State of the World's Plant Genetic Resources for Food and Agriculture [8], a total of 45,375 accessions were conserved, of which 59.7% were cereal species and 23.3% were grain legumes. This is probably linked to the dietary role of these plant resources in providing starch, as regards Poaceae (in particular, corn and wheat) and Solanaceae (potatoes), and proteins, as regards Fabaceae (legumes in general). This is also confirmed by the homogeneous distribution of Fabaceae, Poaceae, and Solanaceae throughout the Italian territory ( Figure 3).
Considering, instead, the landraces split into families for each region of Italy (Figure 4), interesting study cases can be observed. For example, only the varieties that can develop at higher altitude such as Fabaceae and Poaceae can be found in Aosta Valley, an almost totally mountainous region, while Sicily is rich in Poaceae landraces, also due to the presence of short-cycle grains (such as the variety "Tumminia") that can grow in the specific conditions of the island, where a second wheat yield can be achieved during the year [54]. In Molise, 50 varieties of beans were identified, contributing greatly to the richness of this family at a national level. highlands [53]. In Portugal, according to the Second Report on the State of the World's Plant Genetic Resources for Food and Agriculture [8], a total of 45,375 accessions were conserved, of which 59.7% were cereal species and 23.3% were grain legumes. This is probably linked to the dietary role of these plant resources in providing starch, as regards Poaceae (in particular, corn and wheat) and Solanaceae (potatoes), and proteins, as regards Fabaceae (legumes in general). This is also confirmed by the homogeneous distribution of Fabaceae, Poaceae, and Solanaceae throughout the Italian territory ( Figure 3). 4), interesting study cases can be observed. For example, only the varieties that velop at higher altitude such as Fabaceae and Poaceae can be found in Aosta Va almost totally mountainous region, while Sicily is rich in Poaceae landraces, also the presence of short-cycle grains (such as the variety "Tumminia") that can grow specific conditions of the island, where a second wheat yield can be achieved dur year [54]. In Molise, 50 varieties of beans were identified, contributing greatly to t ness of this family at a national level. Apulia, historically considered one of the most important Italian areas for whea duction [55], turned out to be very poor in Poaceae landraces, differently from most o other Italian regions, probably due to the use of modern commercial varieties wit consequent loss of traditional cultivars. Lombardy, despite being a region characte by industrialized agriculture in the plains (Po Plain in particular), has preserved m Poaceae, Solanaceae, and Fabaceae thanks to the presence of traditional mountain v ties of Zea mays, Solanum tuberosum, and Phaseolus spp. [30,48,49]. For example, the m landraces "Spinato di Gandino", "Rostrato Rosso di Rovetta", "Scagliolo di Caren "Nero Spinoso", and "Mais delle Fiorine" are good examples of agro-ecotypes of maize for "polenta" production developed over the centuries [56]. Landraces of Cal were almost all included in the Fabaceae family. Emilia Romagna and Sardinia h high percentage of Cucurbitaceae.
Tuscany is the richest region in herbaceous landraces (197) (Figure 4 and Figu probably due to the unique logistics of the agri-food sector in this region [57], where tronomic heritage has always been a key factor in touristic success. This region is follo Apulia, historically considered one of the most important Italian areas for wheat production [55], turned out to be very poor in Poaceae landraces, differently from most of the other Italian regions, probably due to the use of modern commercial varieties with the consequent loss of traditional cultivars. Lombardy, despite being a region characterized by industrialized agriculture in the plains (Po Plain in particular), has preserved many Poaceae, Solanaceae, and Fabaceae thanks to the presence of traditional mountain varieties of Zea mays, Solanum tuberosum, and Phaseolus spp. [30,48,49]. For example, the maize landraces "Spinato di Gandino", "Rostrato Rosso di Rovetta", "Scagliolo di Carenno", "Nero Spinoso", and "Mais delle Fiorine" are good examples of agro-ecotypes of flint maize for "polenta" production developed over the centuries [56]. Landraces of Calabria were almost all included in the Fabaceae family. Emilia Romagna and Sardinia have a high percentage of Cucurbitaceae.
Tuscany is the richest region in herbaceous landraces (197) (Figures 4 and 5), probably due to the unique logistics of the agri-food sector in this region [57], where gastronomic heritage has always been a key factor in touristic success. This region is followed by Campania (185) and Lombardy (138). However, considering the number weighted on the surface of the region, Molise was the region with the highest density of landraces and Campania was the second, while Lombardy was tenth ( Figure 5). This richness in plant agrobiodiversity of the Molise region and the neighboring Apennine areas (Figures 3 and 5) is probably determined by a variety of environmental and anthropic conditions (such as different bioclimatic levels that reach from sea level to mountain altitudes over a short distance) and agrosystems and related practices distributed among low-input and familybased agriculture [58]. by Campania (185) and Lombardy (138). However, considering the number weighted on the surface of the region, Molise was the region with the highest density of landraces and Campania was the second, while Lombardy was tenth ( Figure 5). This richness in plant agrobiodiversity of the Molise region and the neighboring Apennine areas (Figures 3 and  5) is probably determined by a variety of environmental and anthropic conditions (such as different bioclimatic levels that reach from sea level to mountain altitudes over a short distance) and agrosystems and related practices distributed among low-input and familybased agriculture [58]. Apart from a few exceptions, the areas richest in number of landraces and in different crops cultivated as landraces were located inland in hilly and mountainous areas of the Apennine chain that are characterized by high environmental diversity. The Apennines is a land system that preserves, and in some case enhances, the biological diversity threatened by changes in land use and by diffuse abandonment [58,59].
As can be observed in the map in Figure 5, some areas (Apennines of Molise, Abruzzo, Campania, and Calabria, the Central Apennines of Tuscany, Umbria, and Emilia-Romagna, and Central Eastern Alps on the border between Lombardy and Trentino-Alto-Adige) can be considered "hotspots" of herbaceous landraces and this may be linked to an anthropical explanation, such as the existence of research and inventory initiatives (as already explained for the case of the VNR catalogue [41] or CRC Ge.S.Di.Mont. [30,46,47,48,49,56] or due to the existence of particularly active farmers' associations). Agricultural and environmental conditions obviously also contributed to this phenomenon. In the research found in [60], in fact, areas with a high level of agrobiodiversity in which to set or enhance in situ conservation of plant genetic resources (most appropriate areas (MAPAs)) were identified through a methodological approach using three criteria: pres- Apart from a few exceptions, the areas richest in number of landraces and in different crops cultivated as landraces were located inland in hilly and mountainous areas of the Apennine chain that are characterized by high environmental diversity. The Apennines is a land system that preserves, and in some case enhances, the biological diversity threatened by changes in land use and by diffuse abandonment [58,59].
As can be observed in the map in Figure 5, some areas (Apennines of Molise, Abruzzo, Campania, and Calabria, the Central Apennines of Tuscany, Umbria, and Emilia-Romagna, and Central Eastern Alps on the border between Lombardy and Trentino-Alto-Adige) can be considered "hotspots" of herbaceous landraces and this may be linked to an anthropical explanation, such as the existence of research and inventory initiatives (as already explained for the case of the VNR catalogue [41] or CRC Ge.S.Di.Mont. [30,[46][47][48][49]56] or due to the existence of particularly active farmers' associations). Agricultural and environmental conditions obviously also contributed to this phenomenon. In the research found in [60], in fact, areas with a high level of agrobiodiversity in which to set or enhance in situ conservation of plant genetic resources (most appropriate areas (MAPAs)) were identified through a methodological approach using three criteria: presence of landrace diversity, presence of wild species, and agroecosystem ecological diversity. Considering their additive strategy (AS), it is not surprising that they match the hotspots identified.
The MDS biplot ( Figure 6) shows that along the first axis (MDS1), which accounts for 98.34% of the total variance in the dataset, the samples are sorted by elevation (which increases along the MDS1).
Diversity 2021, 13, x FOR PEER REVIEW 11 of 17 ence of landrace diversity, presence of wild species, and agroecosystem ecological diversity. Considering their additive strategy (AS), it is not surprising that they match the hotspots identified. The MDS biplot ( Figure 6) shows that along the first axis (MDS1), which accounts for 98.34% of the total variance in the dataset, the samples are sorted by elevation (which increases along the MDS1). Landraces are concentrated in the sub-mountain, hilly, and foothill areas, while there are only a few cases in the great industrial plains (Figures 3 and 7). Comparing Figure 2, Figure 6, and Figure 7, it is clear that the families that contribute most to the number of landraces (Fabaceae, Poaceae, and Solanaceae) are those containing landraces cultivated in a wide altitudinal range (from sea level to over 1000 m a.s.l.) but concentrated in hilly and sub-mountain areas (150-800 m a.s.l.). This is due to the fact that these families contain plants adaptable to, and grown in, these environments, such as beans (Phaseolus spp.), rye (Secale cereale), potatoes (Solanum tuberosum), and barley (Hordeum spp.) [48,61,62,63]. As shown by this study, most of the landraces that are part of the Fabaceae family are cultivated in hilly and sub-mountain areas and 70% are beans (Figure 8). Moreover, most of the Poaceae landraces (excluding rice landraces) are cultivated in hilly and sub-mountain areas, especially rye landraces, while as regards the Solanaceae family, only potato landraces (25% of all Solanaceae) are mostly grown from 300 to 1200 m a.s.l. (Figure 8). Very few landraces are grown over 1500 m a.s.l. because at the highest elevations, anthropic activities (such as agriculture) are very limited due to the harsh climatic conditions. Landraces are concentrated in the sub-mountain, hilly, and foothill areas, while there are only a few cases in the great industrial plains (Figures 3 and 7). Comparing Figures 2, 6 and 7, it is clear that the families that contribute most to the number of landraces (Fabaceae, Poaceae, and Solanaceae) are those containing landraces cultivated in a wide altitudinal range (from sea level to over 1000 m a.s.l.) but concentrated in hilly and sub-mountain areas (150-800 m a.s.l.). This is due to the fact that these families contain plants adaptable to, and grown in, these environments, such as beans (Phaseolus spp.), rye (Secale cereale), potatoes (Solanum tuberosum), and barley (Hordeum spp.) [48,[61][62][63]. As shown by this study, most of the landraces that are part of the Fabaceae family are cultivated in hilly and sub-mountain areas and 70% are beans (Figure 8). Moreover, most of the Poaceae landraces (excluding rice landraces) are cultivated in hilly and sub-mountain areas, especially rye landraces, while as regards the Solanaceae family, only potato landraces (25% of all Solanaceae) are mostly grown from 300 to 1200 m a.s.l. (Figure 8). Very few landraces are grown over 1500 m a.s.l. because at the highest elevations, anthropic activities (such as agriculture) are very limited due to the harsh climatic conditions. Diversity 2021, 13, x FOR PEER REVIEW 12 of 17 Therefore, as for many other areas of the planet [53,61,64,65,66], mountainous and hilly territories represent hotspots of plant agrobiodiversity in Italy. This is firstly linked to the remoteness of such areas compared to the great plains (such as the Po Plain), where highly productive commercial varieties or hybrids are grown. The particular environmental conditions of hilly and mountain areas and difficulties in communication have guaranteed that cultivars specifically adapted to those territories developed, becoming landraces. At the same time, in the great industrialized areas (which are generally located in the plains), commercial cultivars rapidly substituted traditional cultivars, resulting in the disappearance of traditional agricultural systems and the knowledge associated with landrace development, maintenance, management, and use, an occurrence reported in many parts of the world [67]. However, it should be noted that small islands often show a remarkable wealth of landraces together with traditional methods of agriculture and horticulture, and, even if this is not clear from the kernel density map, they are worth mentioning as agrobiodiversity spots, particularly as regards the crop wild relatives, as can be seen from previous studies in Italy [68,69]. This is probably due to the same reasons as for mountain and hilly areas (neglection, isolation, and traditional rural methods) and can be observed in small islands belonging to very different climatic fascia, for example, the case of the Outer Hebridean islands of Western Scotland highlighted in research for the UK national landrace inventory [70]. Sometimes, we can find convergent situations, where a small island is also a mountain territory with a high number of not easily accessible small villages, as in the cases of Sicily, Sardinia, and Corsica. An exploratory mission in the latter collected more than 112 valuable landraces, belonging to several species of pulses, vegetables, cereals, forages, spices, and condiments [71]. Therefore, as for many other areas of the planet [53,61,[64][65][66], mountainous and hilly territories represent hotspots of plant agrobiodiversity in Italy. This is firstly linked to the remoteness of such areas compared to the great plains (such as the Po Plain), where highly productive commercial varieties or hybrids are grown. The particular environmental conditions of hilly and mountain areas and difficulties in communication have guaranteed that cultivars specifically adapted to those territories developed, becoming landraces. At the same time, in the great industrialized areas (which are generally located in the plains), commercial cultivars rapidly substituted traditional cultivars, resulting in the disappearance of traditional agricultural systems and the knowledge associated with landrace development, maintenance, management, and use, an occurrence reported in many parts of the world [67]. However, it should be noted that small islands often show a remarkable wealth of landraces together with traditional methods of agriculture and horticulture, and, even if this is not clear from the kernel density map, they are worth mentioning as agrobiodiversity spots, particularly as regards the crop wild relatives, as can be seen from previous studies in Italy [68,69]. This is probably due to the same reasons as for mountain and hilly areas (neglection, isolation, and traditional rural methods) and can be observed in small islands belonging to very different climatic fascia, for example, the case of the Outer Hebridean islands of Western Scotland highlighted in research for the UK national landrace inventory [70]. Sometimes, we can find convergent situations, where a small island is also a mountain territory with a high number of not easily accessible small villages, as in the cases of Sicily, Sardinia, and Corsica. An exploratory mission in the latter collected more than 112 valuable landraces, belonging to several species of pulses, vegetables, cereals, forages, spices, and condiments [71]. In general, the absence of detailed knowledge on agrobiodiversity prevents the application of methodological tools that could successfully assist in biodiversity conservation, and inventories are a first step to assessing biodiversity richness and distribution patterns and monitoring changes [36]. Concomitantly, the conservation of agroecosystems and traditional farming methods together with the protection of buffer areas of biodiversity such as dry walls, green infrastructures, and ecotone belts (and other typologies of habitat defined by directive 92/43 ECC) can guarantee the conservation of a high level of biodiversity and the conservation of species also including crop wild relatives.
The data contained in the National Register must be considered the "official" data on Italian agrobiodiversity, since this is the tool by which the Italian government implemented European Community guidelines on agrobiodiversity conservation. The data reflect the situation globally, as it is recognized that, worldwide, our knowledge of biodiversity includes only 20% of the total estimated number of species [36]. Coordination of the stakeholders in charge of agrobiodiversity conservation and study therefore appears fundamental.
As shown by this study, marginal territories such as mountain and hilly areas are hotspots of agrobiodiversity (confirming the data of the few studies conducted in Italian provinces/regions [30,60,72]), and these areas are, today, subject to climatic and anthropic imbalances. The reasons that led to the abandonment of mountain landraces are many but are mainly linked to the abandonment of mountain areas by human beings. This phenomenon has intensified in the Alps since the middle of the last century and is still in progress [36,73,74,75,76,77,78]. The abandonment of mountain areas (and related agricultural practices) has caused changes to the landscape (loss of agricultural land due to the expansion of forests), imbalances to ecosystems (increase in susceptibility to hydrogeological instability phenomena), and the loss of agrobiodiversity and all it involves (typical dishes, traditions, and popular culture linked to agriculture) [48]. It is therefore of paramount importance to reverse this trend to preserve this heritage through the characterization (agronomic, genetic, nutritional, and historical) of landraces and the promotion of sustainable, innovative, and quality agri-food chains [30] other than the conservation in seed banks [79,80]. These actions should be implemented as soon as possible to avoid the loss of such resources, as, from our survey, only 33.4% of landraces are grown by professional farmers, while the remaining (66.6%) are grown by hobbyist farmers. In general, the absence of detailed knowledge on agrobiodiversity prevents the application of methodological tools that could successfully assist in biodiversity conservation, and inventories are a first step to assessing biodiversity richness and distribution patterns and monitoring changes [36]. Concomitantly, the conservation of agroecosystems and traditional farming methods together with the protection of buffer areas of biodiversity such as dry walls, green infrastructures, and ecotone belts (and other typologies of habitat defined by directive 92/43 ECC) can guarantee the conservation of a high level of biodiversity and the conservation of species also including crop wild relatives.
The data contained in the National Register must be considered the "official" data on Italian agrobiodiversity, since this is the tool by which the Italian government implemented European Community guidelines on agrobiodiversity conservation. The data reflect the situation globally, as it is recognized that, worldwide, our knowledge of biodiversity includes only 20% of the total estimated number of species [36]. Coordination of the stakeholders in charge of agrobiodiversity conservation and study therefore appears fundamental.
As shown by this study, marginal territories such as mountain and hilly areas are hotspots of agrobiodiversity (confirming the data of the few studies conducted in Italian provinces/regions [30,60,72]), and these areas are, today, subject to climatic and anthropic imbalances. The reasons that led to the abandonment of mountain landraces are many but are mainly linked to the abandonment of mountain areas by human beings. This phenomenon has intensified in the Alps since the middle of the last century and is still in progress [36,[73][74][75][76][77][78]. The abandonment of mountain areas (and related agricultural practices) has caused changes to the landscape (loss of agricultural land due to the expansion of forests), imbalances to ecosystems (increase in susceptibility to hydrogeological instability phenomena), and the loss of agrobiodiversity and all it involves (typical dishes, traditions, and popular culture linked to agriculture) [48]. It is therefore of paramount importance to reverse this trend to preserve this heritage through the characterization (agronomic, genetic, nutritional, and historical) of landraces and the promotion of sustainable, innovative, and quality agri-food chains [30] other than the conservation in seed banks [79,80]. These actions should be implemented as soon as possible to avoid the loss of such resources, as, from our survey, only 33.4% of landraces are grown by professional farmers, while the remaining (66.6%) are grown by hobbyist farmers.

Conclusions
This research contributed to knowledge on herbaceous landraces cultivated (preserved in situ) in Italy by merging the most important databases. The study allowed the creation of an updated database of 1615 herbaceous landrace varieties, meaning that the total number of Italian herbaceous landraces identified could increase by about 290%. Their distribution throughout Italy was analyzed (considering single regions), highlighting interesting regional differences and locating herbaceous landrace hotspots, situated mainly in the hilly and sub-mountain marginal areas of Italy.
Parallels can be drawn between the Italian situation and the hilly, sub-mountain, and, in general, marginal areas (such as small islands and archipelagos) that can be considered hotspots of agrobiodiversity and require scientific attention for the protection of biodiversity and agrobiodiversity. Data on the latter, in particular horticultural and herbaceous agrobiodiversity, particularly at risk due to its characteristics, need merging and analyzing as inventories as the first step towards the study, conservation, and promotion of agrobiodiversity.
It is to be hoped that the dataset of this research, currently the most comprehensive and updated available in Italy, will be useful to all stakeholders (civil society, farmers, and governance as well as researchers) as it provides information on Italian herbaceous landraces and can be a starting point to enrich the Agrobiodiversity National Register and trigger actions of characterization, protection, and promotion of these resources. For these purposes, the data gathered during the research work will be forwarded to the Department for Regional Affairs and Autonomies (DARA) of the Italian Presidency of the Council of Ministers. Furthermore, an online interactive map reporting the main information regarding each landrace will be available open source at the following link: https://www.unimontagna.it/servizi/mappatura-agrobiodiversita-vegetale/. It is also to be hoped that, in the future, the database will be supplemented with maintainer details, cultivation, usage, and specific threats and that information will be made available via a web-enabled database and periodically updated to monitor changes in on-farm maintained landrace diversity following the recommendations of [40].
Funding: This research was supported by the Department for Regional Affairs and Autonomies (DARA) of the Italian Presidency of the Council of Ministers (DARA-CRC Ge.S.Di.Mont. agreement), FISR-MIUR "Italian Mountain Lab" project, and by the "Montagne: Living Labs di innovazione per la transizione ecologica e digitale" project.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.