Landrace added value and accessibility in Europe: what a collection of case studies tells us

In the actual climate change scenario, in situ conservation of plant genetic resources for food and agriculture can significantly contribute broadening the diversity of our food system as well as increasing its resilience; nevertheless, landrace cultivation has been dramatically reduced in the last decades all over Europe. One of the most effective approaches to counteract the loss of landrace diversity in situ is facilitating its use. Aims of this study were to (i) describe how in situ maintenance of landraces occurs in different agro-environmental conditions in Europe and (ii) identify the main factors influencing landrace added values and accessibility as means to increase their adoption by new farmers. To the purpose, a collection of 95 case studies of both garden and open field landraces maintained in situ was analysed. A first description was obtained by classifying the information into 18 purposely defined categories. Data on landraces added values and accessibility were further transformed into weighted variables; the derived quantitative scores were then used as dependent variables in univariate and multivariate analyses. Results showed that farmers alone are still the main actors maintaining landraces in situ across different European biogeographical regions, mainly carrying out their activity under organic or low-input conditions, often in marginal areas. Results of the multivariate analysis showed that (i) type of actor involved in the multiplication, (ii) the main use of the product and (iii) presence of promotion actions significantly affect garden landraces added value and accessibility; presence of promotion actions was the only factor affecting added value of open field entries. Evidence arising from this work can contribute to the establishment of a fruitful ground of discussion for future European policies and strategies to protect and increase landrace use.


Introduction
Future food supplies depend on few major crops that are exposed to increasingly extreme and uncertain impacts of climate change (Nelson et al. 2009); in this scenario, we are facing an exceptionally high rate of biodiversity loss (FAO 2019). Biodiversity is central for resilience of ecosystems as it provides buffering effects to cope with drastic environmental fluctuations that affect communities within an ecosystem (Altieri 1999;Newton et al. 2009).
Agrobiodiversity-a subset of biodiversity that includes plant genetic resources for food and agriculture (PGRFA)-generally refers to the level of richness with regard to the number of species together with among-and within-species diversity, in a defined area. At species level, the presence of populations, different races and/or varieties, including their wild relatives, can counteract the current agrobiodiversity decline (CBD 1992;Maxted et al. 1997Maxted et al. , 2002Negri et al. 2009). Nowadays, a portion of agrobiodiversity is still managed and maintained in complex agroecosystems worldwide (van de Wouw et al. 2010), but progressively lost. Such loss is recognised as one of the major threats that agriculture faces as consequence of climate change (Lobell et al. 2008;FAO 2019).
Preservation of agrobiodiversity encompasses a range of diverse strategies and consequent actions to prevent loss of genetic, species and ecosystem diversity. Currently, a relevant portion of PGRFA is safely conserved ex-situ (i.e. in genebanks) since this conservation means has been regarded as the most practicable security measure 'to save as much as possible' from a short-term perspective (Gomez-Campo 1985). However, removing species, and their associated variability, from their environments of origin freezes their evolutionary and adaptive potential (Brush 2004;de Haan et al. 2013). A widely recognised solution to this issue is to integrate ex situ with in situ conservation by maintaining PGRFA in the areas where they have evolved over time and produced their distinctive traits (CBD 1992). This approach brings important advantages, in fact, its application can support functioning, resilience and, to some extent, productivity of agroecosystems (Tilman 1999(Tilman , 2000. In Europe, in situ conservation is generally associated with conservation of wild plants in protected areas; however, some actions focusing on landraces have been also carried out (Negri et al. 2013;Vetelainen et al. 2009). In the past, most of such actions arose locally or nationally and after a number of international documents and agreements (CBD 1992(CBD , 2010aFAO 2001FAO , 2012. Currently, through the formalisation of the EU Biodiversity Strategy for 2030 (European Commission 2020), the EU has created a favourable environment for the development of coordinated strategies to facilitate conservation of landraces and their sustainable use.
Although most of European agricultural production has been relying on registered and uniform cultivars, landraces, obsolete cultivars, and also other non-uniform varieties are still grown on-farm and in home gardens, i.e. in situ Galluzzi et al. 2010). However, to date, no systematic studies were carried out to depict in situ conservation and use of landraces in Europe; in fact, diverse conservation situations and actors maintaining these materials make their systematic tracking rather complex and challenging at the same time. On the other hand, understanding and properly maintaining such diversified agricultural systems can favour equitable mechanisms of income generation and help society to better benefit of these resources, hence improving livelihoods and well-being (Bellon et al. 2015), whilst facing the challenge posed by the effects of climate change (Bellon et al. 2017). It is also noteworthy that landraces still hold great variability that can support future crop improvement for sustainable farming systems .
Most importantly, understanding which factors increase (i) added value of landrace products and (ii) access to their propagation materials (e.g. seeds or seedlings), can favour the establishment of appropriate policies promoting use and adoption of landraces by new farmers across Europe, in line with the principles of its Biodiversity Strategy (European Commission 2020). Understanding the state of in situ conservation of landraces across Europe is not only a matter of collecting information about where and how these resources are cultivated, but also of recognising mechanisms and actions that have been put in place to favour their cultivation; for example, in other regions, collective market actions and ad hoc labelling strategies already created virtuous landrace in situ conservation conditions (Devaux et al. 2009;Krishna et al. 2010).
In order to get insights on the current situation about landrace in situ conservation, there are two possible general approaches. The first would privilege quantitative analysis of few variables across large samples (nomothetic approach) while the other would focus on qualitative, multi-aspect, in depth study of one or few case studies (idiographic approach). This difference brings to the key matter of studying many issues in few cases or few issues in many observations. On the other hand, the use of the so-called 'case study methodology' (CSM) can bridge the gap between nomothetic and ideographic methods and transcend, to a certain extent, their specific limitations (Larsson 1993). CSM takes advantage of the ability of case studies to deal with processual and multiple stakeholder considerations through multisource data; the nature of case studies and their complexity can make statistical examination of patterns, through different situations, rather difficult. However, this method tries to overcome this limitation by considering large sets of case studies.
In this work CSM was used to perform quantitative analysis on a large collection of case studies, gathering information about different in situ maintained landraces crossing a range of different socio-economic, pedo-climatic and cultural contexts in Europe as such offering insights that might not be achieved using other methodological approaches. In particular, the application of this methodology, allowed to summarise some of the major elements affecting in situ maintenance of landraces in Europe and the conditions that influence their added value and accessibility, both key factors that can contribute increasing adoption of landraces by next generations of farmers.
A recent approach-formalised into the Concept for on-farm conservation and management of PGRFA by the European Cooperative Programme for Plant Genetic Resources (ECPGR 2017)-rather than giving additional and/or broader definition to the term landrace itself, focused on different materials that are maintained in situ: (i) true landraces (i.e. sensu stricto, Polegri and Negri 2010;Negri et al. 2009); (ii) introduced landraces (i.e. 1 3 allochthonous, Zeven, 1998); (iii) selection from landraces ); (iv) crosscomposite populations and varietal mixtures (European Commission 2014; Goldringer et al. 2006;Raggi et al. 2016a, b); (v) historical varieties (Klaedtke et al. 2017). Being characterised by a certain level of genetic heterogeneity, all the above-mentioned materials-hereafter landraces intended in a broad sense-were the object of this study.

Collection of case studies
Case studies analysed in this work were retrieved from the public repository In situ landraces: best practice evidence-based database, hosted by the ECPGR and available at https ://www.ecpgr .cgiar .org/in-situ-landr aces-best-pract ice-evide nce-based -datab ase. The online database encompasses a number of case studies of in situ maintained landraces, representing successful (or potentially successful) examples of use, management and valorisation of these resources.

Data classification
As from the original structure of the database, the units of analysis of this research are 'the case studies', each referring to a distinct in situ maintained landrace. Information of each case study was initially classified into 18 purposely defined categories (Table 1). Altogether, the defined categories, describe basic characteristics and cross-sectional features that show how and where landraces are cultivated, managed, promoted on the market and accessed. The case studies were categorised according to their description by ascertaining the presence of single or multiple sentences showing clear correlation with categorical values in each defined category (Table 1).
In particular, the first six categories describe basic features such as local name, taxonomic classification, type of in situ conserved resource, location where cultivation occurs and mating system of each species (Table 1). The following three (7-9, Table 1) describe actors managing the landraces, their main uses and actions to promote their products. Categories from 10 to 12 describe the agronomic contexts where landraces are cultivated and some of their management features; categories from 13 to 15 are about market extent and use of geographical indications or other labels to add value to landrace products. The last three (from 16 to 18) categorise factors that can increase landrace accessibility and their potential adoption by new farmers (Table 1).
Due to the positive effects that product added value and improved accessibility can have in increasing current and future use of landraces-by increasing the value of their product and the number of actors able to access their propagation materials-the categories describing these aspects, added value (13-15) and accessibility (16-18), were converted into weighted variables. This transformation allowed to weight differences among categorical values in creating added value or improving access to propagation materials (Table 2).
In particular, as for categories defining the added value, an increasing value from 1 to 7 was assigned from products having no market, used for self-supply only (value = 1), up to products characterised by an international market (value = 7) passing through local (value = 3) and national market (value = 5) ( Table 2). The selected weighted values reflect the idea that an increasing market extent can directly contribute to add value to a local product allowing to reach a larger number of potential consumers. Regarding labels, it has been demonstrated that concerning high quality food products, consumers value more those holding a Geographical Indication (GI), in comparison to non-GI labelled products (Menapace et al. 2009). In addition, the promotion of linkages between local producers, their local areas and their food products through GIs has been recognised as a crucial pathway to nutritious food systems and sustainable development for rural communities by both Food and Agriculture Organisation (FAO) and the European Bank for Reconstruction and Development (Vandecandelaere et al. 2018). For these reasons, a weighted value of 5 was assigned to landrace products holding a GI label while of 0 to those without any. Similarly, a weighted value of 3 was assigned to products holding other types of quality labels while of 0 to those without ( Table 2). As for categories defining accessibility of landraces propagation material, the highest weighted value (7) was only assigned to entries for which seeds or seedlings are available on the market (Table 2); indeed, in these cases, a relevant number of seeds or seedlings is expected to be accessible with no restrictions. On the contrary, a lower weighted value (5) was assigned to landraces whose propagation material can only be accessed from ex situ storage facilities due to some limitations imposed to their access, such as the signature of a Standard Material Transfer Agreement, and low number of seeds that can be retrieved from gene banks. Finally, increasing weighed values were assigned to landraces listed in local, national or international inventories (1, 3 and 5, respectively) since the inclusion in such lists can help landraces to reach potential users at local, national or international level. Weighed values were assigned to each entry according to the corresponding categorical values (Table 2). For each entry, an 'Added Value Quantitative Score' (AddValQS) was calculated as the sum of the assigned weighted values for the categories from 13 to 15 (Table 2); the same procedure, applied to categories from 16 to 18 (Table 2), was used to estimate the 'Accessibility Quantitative Score' (AccQS).

Data analysis
A general picture of the dataset was initially obtained by calculating the distributions of the case studies into the 18 categories, expressed as percentages. For each case study, average latitude and longitude of the cultivation area were retrieved and imported into GIS. The distribution of the landrace cultivation areas into the different European biogeographical regions was obtained by mapping the entries against 'the cartographic elaboration of the European biogeographical regions', obtained from the European Environment Agency web site (https ://www.eea.europ a.eu/data-and-maps/data/bioge ograp hical -regio ns-europ e-3).
In order to reduce complexity of the dataset, and make comparisons among more homogeneous groups, case studies were divided into garden and open field entries according to their usual classification and not on the average size of the area in which a specific landrace is cultivated.
In the univariate analysis, differences in medians of AddValQS and AccQS among the categories Multiplication actor, Use and Promotion were initially tested by means of nonparametric Kruskal-Wallis test (Rohlf and Sokal 1980). In addition, to test the effect of a specific categorical value of the same categories on AddValQS and AccQS, a set of binary variables were created (Calvet-Mir et al. 2011) as follows: (1) Multiplication actor, took the value of 1 if the landrace is multiplied by 'Farmers' and 0 otherwise; (2) Use, took the value of 1 if the landrace is mainly used as 'Home-supply' and 0 otherwise and (3) Promotion, took the value of 1 if the level of promotion of the landrace is limited to maintainers or local actors only and 0 otherwise. In the multivariate analyses, the above-described binary variables were used as independent variables against AddValQS and AccQS, used as dependent variables. In the analysis, distinct models were tested for garden and open field entries and case studies characterised by ≥ 1 missing data were not included.

Limitations
The dataset analysed in this work is vast and complex, encompassing entries of different crops species cultivated using different agricultural practices and, most notably, cultivated for different purposes. In addition, the unit of analysis is 'the landrace', not 'the farm', 'the farmer' nor 'the gardener'; this limit questions that can potentially be addressed. Finally, not all the European countries are covered.

The collected case studies
A number of 95 case studies (hereafter entries) were retrieved from a total of 105; the remaining 10 case studies were discarded being related to non-European materials (Bellon and Brush 1994) or to fruit trees, which investigation is beyond the scope this work.
Entries belong to 46 different crop species classified as garden and open field; the most represented crops are: Solanum lycopersicum (tomato, 12), Phaseolus vulgaris (common bean, 8) and Secale cereale (rye, 6); 8 cereal entries belonging to the Triticum genus are also present ( Table 3).
The collection encompasses entries from 15 different European countries; Italy, Spain and Hungary are represented with the highest number: 16, 13 and 13, respectively ( Table 3). As from the geographical distribution of entries across Europe, almost all the main European biogeographical regions are covered: Mediterranean (31 entries), Continental (16), Atlantic (14), Pannonian (14), Boreal (11), Alpine (7) and Macaronesia (2) (Fig. 1). Garden entries appear to be mostly located in southern Europe while, for open field, no clear geographical distribution can be observed (Fig. 1).
Farmers-single (84.1%) or grouped in consortia (6.8%)-are the main actors carrying out landrace multiplication (n = 88). The 27.4% of the entries resulted mostly cultivated for 'Home-supply' while the rest for commercial purposes (n = 95). According to the categorised data, the level of promotion (based on actions supporting the landrace) is 'Absent' in the 20.7%, 'Local' in 39.5%, 'National' 34.1% and 'International' in the 6.1% of the entries (n = 81).
When the relative size of the cultivation area is considered 46.8% of entries are categorised as 'Small', 39.4% 'Medium' and 13.8% 'Large' (n = 94). Most entries (83.3%) are cultivated under 'Low-input/organic' conditions and the others following conventional agronomic practices (16.7%) (n = 72). Interestingly, about half of the entries (43.3%) is cultivated in marginal areas (i.e. areas characterised by unfavourable conditions) (n = 90). The full list of 95 case studies with the corresponding categorical values is available in Table S1 (Supplementary Materials).
Raw or processed landrace products are mainly commercialised in local markets (58.9%) and, to a minor extent, at national level (23.3%), while a lower but significant portion is cultivated by gardeners or farmers only to satisfy their needs (14.4%); in few cases the products are commercialised in other countries (international markets, 3.3%) (n = 90). In most cases no geographical indication to promote raw or processed products on the market is reported (82.1%); indeed, only few entries hold a EU geographical indication (10.6%) or other types of geographical indication (7.4%) (n = 95). However, for 35.4% of the case studies other types of commercial indications or brands are present such as 'Slow Food Presidium', 'Green heritage' and 'Pro Specie Rara'. Finally, only six case studies hold a geographical indication and another brand or indication at the same time. Regarding accessibility of seed or seedlings, most of landraces can be obtained through ex situ Table 3 List of species with relative number of entries and country of origin of 95 in situ maintained landraces case studies

Added values and accessibility quantitative scores
When garden entries are considered, the added value quantitative score (AddValQS) reached a maximum value of 15 in the case of the Italian Aglione della Val di Chiana (Allium ampeloprasum)-the only landrace characterised by having an international market and both geographical indication and commercial labels-followed by the broccoli Broccolo Fiolaro di Creazzo (13) and the Finnish potato Puikula (12). Among open field entries, highest AddValQS characterise the bread wheat Solina (13), the maize Rheintaler Ribelmais (10) and the grass pea Fava Feneou (10). Regarding the accessibility quantitative score (AccQS) more than one entry scored the maximum value (17): six garden tomatoes from Hungary, the pepper Glikokafteri Mpachovou and, among open field entries, the maize Nostrano di Storo, the barley Mix48, the einkorn wheat Kaploutzas and the emmer

Factors affecting landrace added values and their accessibility
Regarding garden crops, results of the Kruskal-Wallis test revealed the existence of significant differences among medians of the AddValQS for Multiplication actor (P ≤ 0.05, n = 59), Use (P ≤ 0.001, n = 62) and Promotion (P ≤ 0.001, n = 55), while AccQS values were different among medians of the category Promotion (P ≤ 0.05, n = 55). Regarding open field crops, AddValQS was only different among categorical values of Promotion (P ≤ 0.05, n = 32), while AccQS among the categorical values of Use (P ≤ 0.05, n = 41).
Results of the multivariate regressions (Table 4), where the explanatory variables Multiplication actor, Use and Promotion-coded as binary-were considered at the same time, were quite consistent with those obtained in the univariate analysis.

Garden
Open field • the multiplication is carried out by consortia, seed companies or public bodies. In other words, it is expected that when the seed or seedling production of a landrace is managed by organised actors, instead of non-organised farmers, the final product has wider market and holds single or multiple indications or brands; • the use of the landrace is mainly intended for the market, meaning that it is rather difficult that a landrace used for self-supply holds a geographical or a commercial label; • the promotion has been developed at National or International level, in comparison to local promotion or absence of promotion, meaning that extensive promotion actions positively affect market size and labels held by a certain landrace.
For the same type of entries (garden) accessibility quantitative score (AccQS) increases when: • entries are used for self-supply being frequently in ex situ collections, listed in registers or sold on the market; • promotion has been carried out at national or international level, in comparison to no or local promotion, meaning that when a landrace underwent international or national promotion actions it is likely to be more accessible for potential new users.
Regarding open field entries, promotion at national or international level is the only factor positively affecting the added value quantitative score (AddValQS); it means that also for open field entries promotion actions positively affect market size and presence of labels (Table 4). None of the three factors significantly affects accessibility quantitative scores (AccQS) ( Table 4). In this case, the low number of case studies strongly reduced our ability to predict open field entries AddValQS and AccQS starting from the values of the three considered binary variables.

Discussion
This collection of 95 case studies gives a cross-sectional picture of some of the key features that can contribute sustaining and possibly increase landrace cultivation across Europe. Indeed, a better understanding of factors positively affecting added value and accessibility of such genetic resources can be the ground of discussion for future measures aiming at improving the wealth of landrace cultivation. In fact, it has been already suggested that an effective, sustainable and long-term in situ conservation of landraces can only occur with adoption in cultivation by new farmers (Casañas et al. 2017;Maxted et al. 2009); these conditions are fostered when landraces are characterised by an appealing market potential (i.e. added values) and by clear rules and protocols that favour their accessibility (Spataro and Negri 2013). Today, the need of passing landraces in the hands of new generations is more important than ever, as a matter of facts it has been shown that European farmers using landraces are elderly (Calvet-Mir et al. 2011).
From the beginning, the idea behind the analysis of a large set of case studies was to include as much information as possible about the main features that characterise in situ conservation of crops' diversity. This brought the need of including, together with true landraces, examples of a range of different cultivated materials such as introduced landraces, historical varieties, selections from landraces and other types of populations that also encompass an intrinsic genetic diversity. Indeed, nowadays, cultivation and use of these 1 3 materials is seen as positive means to foster a diversified and more sustainable food system, able to support, to a certain extent, healthy agro-ecosystems, their functioning, resilience and productivity, especially in low-input conditions (ECPGR 2017). The CSM approach used in the present study helped in overcoming major drawbacks of single case study analysis, that is usually characterised by the inability of examining cross-sectional themes and to generalise conclusions and findings to large and diversified systems (Larsson 1993). For this reason, even considering the complexity of this topic, we are fairly confident that the evidence arising from this work well describes the context of in situ conservation of landraces as well as its diversified features. Indeed, collated information covers many diverse in situ conservation contexts that characterise Europe; in fact, as many as 46 crop species, of both garden and open field crops, are represented and the case studies are well distributed over the main European biogeographical regions.
Results of this research indicated that landrace cultivation frequently occurs in marginal areas contributing to the income of farmers working in conditions in which conventional agriculture cannot be easily carried out. Landraces, and especially true landraces-the majority of entries in this study-have adapted, and are continuously adapting, to the specific environmental conditions in which they have been cultivated. For this reason, such materials are of particular interest for marginal areas as well as for being used under organic and low-input management systems, where few external inputs are available for counteracting limiting factors. As a matter of facts, most of entries in our dataset are cultivated under organic or low input conditions.
Going deep into the case study descriptions it emerges that several entries seem to be characterised by peculiar adaptive traits making their cultivation still attractive in specific pedo-climatic conditions. For example, tolerance to different types of limiting factors are reported for many entries. One of the most interesting examples is the Greek tomato landrace Tomataki Santorinis, traditionally cultivated in the volcanic island of Thira (Santorini), where soil is poor in organic matter and water, and plants undergo stressful conditions also due to high temperatures and radiation (Koutsika-Sotiriou et al. 2016); another example is the Austrian onion landrace Laaer Zwiebel that can be cultivated without any irrigation. Among the conservation varieties, the Finnish Hermanni and Iivo rye landraces are characterised by winter hardiness and Italian Solina wheat by adaptation to the specific conditions of the mountain area where it is cultivated (Porfiri et al. 2001).
There is evidence in the literature that local actions like local fairs (Castellini 2005), meals associated with particular recurrences (Papa 1999) and contests for the best landrace product (Mendes-Moreira et al. 2014) have a positive impact on the maintenance of landraces in situ (Negri 2012). Our data also showed that creation of added value is often linked to attribution of geographical indications (e.g. PDO, PGI), trademarks and quality labels. Such labels ascertain the uniqueness of the landrace product-in comparison with others-its particular traits, the link with the cultural and traditional values of the territory or, simply, the way it was obtained. However, the use of a geographic indication resulted not so extensive in the analysed dataset suggesting that there is still room for improvement. In some recent reports FAO (2009) showed that geographical indications can be used as a driver for a sustainable development of rural areas; indeed, geographical indications can link the European heritage of food diversity, genetic resources and local knowledge with the market dimension (Vandecandelaere 2011).
As from the description of the analysed case studies, a high number of landraces underwent different types of promotion initiatives so that we decided to test the effect of promotion on landraces added value (AddValQS) and access (AccQS) measured as quantitative scores; according to our results, promotion was the only factor increasing AddValQS and AccQS of both garden and open field entries (Table 4). Starting from this evidence-in order to make landraces more appealing and accessible to the next generation of European farmers-we would suggest that policies should leverage on actions promoting landraces at national and international level. The application of this strategy could rise landraces use in Europe as a consequence of the potential higher added value of their product and of an easier access to their propagation materials.
From our analysis, other key elements to sustain future cultivation of landraces across Europe emerged. Indeed, results showed that farmers are the main actors carrying out landrace multiplication being, the facto, the most relevant players in managing, using and conserving landrace diversity. From what we observed, seed companies have only a marginal role in producing and commercialising landrace seeds or seedlings. However, testing the effect of different multiplication actors against landraces added value and accessibility, it emerged that products derived from garden landraces have higher added values (AddValQS) when managed by farmer consortia, seed companies or public bodies while this effect was not observed for open field entries (Table 4). In other words, the creation of consortia around specific landraces can enhance added values of their commercial products resulting in a wider market and acquisition of geographical indications or labels, at least in the case of garden crops. In this regard, it is noteworthy that the European seed legislation frame offers several possibilities to register landraces as 'conservation varieties ', 'amateur varieties' or 'populations' (European Commission 2008 favouring the involvement of private seed companies and other stakeholders in the use of landraces. Then, while a small portion of landraces took advantage of these potentially beneficial tools (Spataro and Negri 2013), registration might represent an important means to scale up conservation and accessibility, use and market of garden landraces. Indeed, when seeds or seedlings are present on the market the access is facilitated considering both possible use and the number of seeds/seedlings that can be retrieved. It has already been reported that landraces in situ conservation is favoured by incentives to farmers and collaboration with institutions and social organisations (Bellon 2004;Brush 2004;Zimmerer 2010); in this context an external source of support mainly assumed the form of promotion activities to increase added value and visibility of landrace products on the market, calling consumer attention and interest.

Conclusions
Future policies aimed at promoting landrace conservation and their sustainable use should leverage on strategies favouring promotion to add value and increase accessibility. Indeed, factors that positively affect these aspects should properly be considered when planning national or European actions that want to increase and promote use of landraces in situ. We believe that a joint European long-term strategy can foster in situ conservation by helping the transition of these resources to the next generation of farmers, on behalf of a future diversified European agriculture.