The participatory study was conducted in the western and eastern parts of Odisha. Odisha has long been a pioneer in promoting agro-biodiversity. The state is among the leading conservators of rice germplasm in India. The state has registered over 900 farmers’ varieties with the Protection of Plant Varieties and Farmers’ Rights Authority. Landraces from more than 500 remote villages have been mapped and documented under flagship programmes such as Shree Anna Abhiyan and the Special Programme for the Promotion of Integrated Farming in Tribal Areas (Department of Agriculture and Farmers’ Empowerment, Odisha Millet Mission, 2020). Odisha is home to 64 tribal communities and 13 Particularly Vulnerable Tribal Groups (PVTGs), constituting 23% of the state’s population (Ota et al., 2020). These communities traditionally practice mixed farming systems, cultivating native varieties of millets, pulses, rice, oilseeds, tubers, and vegetables (see Figure 1).

For the purpose of this study, custodian farmers are defined as farmers who actively conserve, cultivate, and maintain indigenous crop landraces over multiple seasons, primarily through on-farm conservation. These farmers save their own seed, participate in seed exchange or community seed bank activities, and possess experiential knowledge related to the agronomic, cultural, and ecological significance of landraces. In contrast, non-custodian farmers are those who do not cultivate or conserve indigenous landraces and predominantly rely on modern or commercially available varieties.

Tribal community farmers were not explicitly stratified or separately identified during sampling; however, the study was conducted in regions with a high prevalence of tribal and traditional farming communities, particularly in eastern Odisha (Jeypore tract, Koraput district), a well-recognised agrobiodiversity hotspot. The selected villages were Umuri, Nauguada, Machhara, Sanuguadi, and Patraput, which are predominantly inhabited by tribal communities such as the Kondh, Paraja, and Gadaba, who traditionally practice landrace-based cultivation and on-farm seed conservation. In western Odisha, the study covered the districts of Jharsuguda, Sundargarh, and Bargarh. Although parts of these districts are experiencing increasing urbanisation and agricultural intensification, pockets of traditional farming communities, but not tribal farmers, continue to conserve indigenous crop landraces. In particular, Bargarh district has emerged as an important centre for landrace conservation, where several custodian farmers have adopted conservation-oriented practices inspired by local innovators. According to local seed bank Desi Bihan Surakhya Samiti, farmers in the district have collectively documented and conserved 925 local paddy germplasm accessions since 2014. Furthermore, of the 30 black rice landraces reported in India, six varieties are conserved within the indigenous seed germplasm of Bargarh district, underscoring the region’s continued relevance in agrobiodiversity conservation.

To gain deeper insights into the “Who” and “Why” behind the conservation and continued cultivation of landraces by certain sections of farmers, we employed the Four Square Analysis (FSA) approach. Focused group discussions (FGDs) were conducted with custodian farmers to collect both qualitative and quantitative data, which were further analysed using appropriate statistical techniques. This method enabled us to assess the extent and distribution of various landraces across the study regions (Grum et al., 2008). It effectively captured the localized spread of farmer-managed landraces, which are typically unequally distributed both among farms (households) and within farms (cultivated area). The analysis was carried out through the following structured steps:

Step 1: Group Formation and Landrace Listing.

Custodian farmers from each village were organized into focus groups. Each group collaboratively developed a comprehensive list of landraces currently cultivated and conserved in their village.

Step 2: Physical Layout of the Four-Square Grid.

A 2 m × 2 m grid was marked on the ground to create a visual four-quadrant structure. The two axes of the square represented:

Horizontal Axis (X-axis): Area under cultivation (large vs. small)

Visual markers were used to facilitate comprehension:

The top two quadrants represented landraces cultivated by many households

Stones/Icons were used for illustration:

Multiple house symbols/Multiple Stones in the top quadrants denoted many households

Step 3: Participatory Classification of Landraces.

During the session, the facilitator stood at the centre of the square, held up the name of a landrace, and asked participants whether the variety was cultivated by many or few households. Based on the group’s consensus, the facilitator moved to the corresponding vertical half of the square.

Step 4: Area-Based Categorization.

The facilitator then inquired whether the same landrace was grown on large or small plots of land. Based on the response, the landrace was placed into one of the four quadrants:

Many households – Large area

The process was repeated for each identified landrace until all were positioned accordingly in the matrix. This information was effectively visualized using a heat map, which indicated the presence, spread, and conservation status of each landrace (e.g., common, intermediate, or rare) in the sampled villages. A cut-off of 46.67% (equivalent to 7 out of 15 households) was used to classify landraces as either being cultivated by “many households” or “few households.” In each village, 15 landrace-conserving households were surveyed. During participatory discussions with custodian farmers, it was agreed that a threshold close to half of the sampled households would provide a meaningful distinction between commonly cultivated and sparsely cultivated landraces. Accordingly, landraces grown by seven or more households were categorized as being cultivated by many households, while those grown by fewer than seven households were classified as being cultivated by few households. According to Sthapit et al. (2006), “The cut-off point is a relative measure. The objective is to facilitate a consensus among custodian farmers to categorize their own diversity. If the group agrees that a 50% distribution (e.g., 7 or 8 out of 15) distinguishes common varieties from those held by only a few ‘custodians,’ this consensus provides the scientific basis for the classification.”

Similarly, cultivation area was categorized as “small” if less than 1 acre and “large” if more than 1 acre. The scoring for the Four-Square analysis has been discussed in Table 3. According to Gauchan et al. (2003), the area threshold is often set based on the minimum land required to meet a family’s food security for a specific period or the average plot size of the locality. In many smallholder systems, 0.5 hectares (approx. 1.2 acres) or 1 acre serves as the dividing line between a “niche” crop and a “mainstream” crop.

During the Four Square Analysis exercise, discussions with custodian farmers revealed a set of key characteristics that influenced their decisions to cultivate and conserve specific landraces. These characteristics were subsequently grouped into 10 major criteria: physiological traits, medicinal/nutritional value, adaptive capacity, commercial potential, social significance, cultural relevance, mystic-religious beliefs, agronomic benefits, food and culinary quality, and harvest and processing ease. To prioritize these criteria, a matrix ranking exercise was conducted involving selected custodian farmers who were actively cultivating and conserving landraces. Ten criteria, identified through earlier focus group discussions, were used for the ranking process. Locally available materials such as stones, pebbles, and sticks were employed as scoring tools. Each farmer was provided with 10 stones per landrace or cultivar and was asked to allocate them across the 10 criteria based on perceived importance-placing more stones on highly preferred criteria (10 indicating the highest preference and 1 the least). This participatory approach allowed custodian farmers to express their priorities in a tangible manner. The criteria receiving the highest total scores were assigned the top ranks, thereby reflecting the most influential factors guiding landrace conservation and cultivation. For the matrix ranking exercise, although the discussion was facilitated in a group setting, scores were elicited from each key informant individually. Each informant independently assigned scores to the ranking criteria based on their own knowledge and experience. These individual-level scores were then compiled to generate the final matrix rankings. Accordingly, the scores were treated as independent observations for statistical analysis.

Following the matrix ranking exercise, the mean preference scores for each of the 10 criteria were calculated to assess the overall importance attributed to them by custodian farmers. Matrix-ranking data in this study were derived from structured participatory scoring exercises in which respondents assigned numerical scores to predefined criteria. These scores were treated as quasi-continuous variables, a practice commonly adopted in agricultural extension and participatory research when ordinal rankings are converted into standardized numerical scales. Matrix-ranking data were analysed following the De Villiers (1996) (QuIK) approach, which permits the transformation of participatory rankings into quantitative scores suitable for ANOVA and DMRT under quasi-continuous assumptions. To statistically validate the differences among these criteria, the data were subjected to a one-way Analysis of Variance (ANOVA), followed by Duncan’s Multiple Range Test (DMRT) as modified by Kramer (1957). This post hoc test, originally proposed by Duncan (1955), is commonly employed after ANOVA when significant differences among group means are detected, enabling researchers to identify which specific groups differ. The procedure was conducted using the Statistical Package for Social Sciences (SPSS). Initially, ANOVA was performed to determine whether significant variation existed among the mean ranks of the 10 criteria. A significance level (α) of 0.05 was set for the analysis. DMRT then sorted the group means in ascending order and systematically compared each pair to assess whether the observed differences were statistically significant. The test calculated a critical range (least significant difference) based on the significance level, number of groups, and degrees of freedom obtained from ANOVA. Pairs of criteria whose mean differences exceeded this critical range were considered significantly different, while those within the range were grouped. The resulting groupings allowed for the classification of criteria into statistically distinct or overlapping categories, thereby offering empirical insight into the most influential factors that guide the cultivation and conservation of landraces. A further box plot was constructed to see which is the highly preferred criterion.

To visually interpret the variability and central tendency in the overall preference scores assigned to each criterion, a box plot was constructed. The box plot allowed for a clear comparison of the 10 criteria based on their median preference scores, interquartile ranges, and potential outliers. It provided a visual summary of how custodian farmers prioritised each criterion in relation to landrace cultivation and conservation. Criteria with higher medians and narrower interquartile ranges would be more consistently preferred, as indicated by their groupings labelled “a”, “b”, and “c” through the DMRT grouping. Criteria showing lower median preferences and broader ranges, suggesting greater variability in farmer responses.

To systematically assess the economic importance of conserved landraces, a Landrace Added Value Score (AddValQS) was computed, adapting the framework proposed by Raggi et al. (2021). This composite index quantifies economic value based on multiple factors reflecting market engagement, institutional recognition, and income generation. Each component was scored using a weighted system (see Table 4), reflecting increasing levels of economic engagement. While the original weighting structure was retained, additional components, sale to retail markets and value-addition activities were included based on participatory consultations with experienced custodian farmers and field experts involved in landrace cultivation and marketing during our pilot survey, a method recognized for enhancing the content validity of socio-economic indices (Munda, 2004). Although formal statistical sensitivity analysis was not conducted, consistency of AddValQS outcomes was cross-checked through qualitative triangulation with Focused Group Discussions and household-level data, ensuring contextual robustness of the index. This approach aligns with the mixed-methods framework outlined by O’Cathain et al. (2010), who describe triangulation as a rigorous technique for validating findings through the convergence of results derived from different methodological lenses.

Six key components were considered:

Annual Income from Landrace Cultivation: Percentage contribution of landrace cultivation to total annual farm income.

Factors Influencing Landrace Added Value: To determine the factors influencing the AddValQS, a multivariate analysis was conducted using binary categorical variables. The following independent variables were included:

Marginal Land Use (1 = <50% of farm under landrace, 0 = ≥50%)

These variables were regressed against AddValQS to identify significant determinants of economic value (see Table 5).

Custodian farmers in both western and eastern Odisha were predominantly middle-aged to elderly, with mean ages ranging from approximately 49 to 51 years, supporting the assertion that landrace conservation knowledge is largely held by experienced farmers. Regarding gender, women farmers were included in both custodian and non-custodian categories; however, consistent with local landholding and household headship patterns, the majority of identified custodians were male. This limitation is now explicitly acknowledged, and the manuscript clarifies that while women play a crucial role in seed selection, storage, and knowledge transmission, formal recognition as “custodian farmers” often remains male-biased due to socio-institutional norms. This is discussed in relation to earlier ethno-botanical findings (e.g., King et al. 2015; Panda and Pandey, 2021).

The study documented a total of 56 landraces of paddy and 10 millet landraces that were cultivated and conserved by custodian farmers in western and eastern agro-ecological regions of Odisha. A total of 30 paddy landraces were recorded in Western Odisha. The characteristics of these landraces were deeply rooted in custodian farmers’ traditional knowledge and were primarily associated with nutritional quality, stress tolerance, cultural importance, and culinary attributes. Landrace like Kalabati was noted for its sugar-free property despite having 78% carbohydrate content and was known to be persistently cultivated for over 6–7 years. Katia, Kala Maliphul, and Brahma Black were rich in iron and calcium, indicating their nutritional superiority. Landraces like Kusumkali and Sonagathi were believed to possess medicinal properties, particularly effective for joint pains and eye health, respectively. Talmuli was recognised for its resistance to Brown Plant Hopper, a serious pest of rice. Jalgudi was appreciated for its ability to survive in submerged conditions, highlighting its suitability in flood-prone areas. In terms of consumption, varieties like Bahal, Bhajna, Annapurna, and Kandhsagar were preferred due to their taste and cultural relevance, especially for traditional dishes such as Pakhala Bhat and Kanika. In Eastern Odisha, 16 paddy and 10 millet landraces were documented. These landraces reflected a blend of market relevance, agronomic resilience, and traditional use. Paddy landraces like Umuriachundi, Machhakanta, and Haldichudi had high market demand and were used in religious and cultural rituals. Dangarbasumati was notable for its high yield and resistance to insect-pest attacks, making it suitable for low-input agriculture. Landraces such as Dodikaburi and Paradhan displayed drought resistance, weed-suppressing abilities, and required minimal fertilizer input. Culturally significant landraces such as Sapuri and Gothia were used in marriage rituals and festivals. Among millets, Finger millet (Mandia) varieties like Batimandia, Dusseramandia, and Telenga Mandia were found to be climate-resilient, nutritious, and suitable for the preparation of local dishes such as Mandia Pitha and Pej. Seetadai (Little millet) and Santarah were specifically noted for their use in field meals and summer beverages, respectively. The detailed characteristics of landraces conserved in Western Odisha and Eastern Odisha are elaborated upon in Tables 6, 7, respectively. The graphical analysis of perceived traits among paddy landraces conserved in both regions of Odisha revealed that attributes such as market demand, suitability for home consumption, aromatic quality, and resistance to pests were most frequently valued by custodian farmers. A comparative trait-wise assessment between Eastern and Western Odisha showed that market demand and suitability for puffed rice were more prevalent in Eastern Odisha, whereas traits like black rice, quicker cooking time, and pest resistance were more prominent in Western Odisha. Notably, home consumption and nutritional value emerged as significant criteria across both regions. These insights were derived from Focus Group Discussions (FGDs) with custodian farmers. To validate these findings, matrix ranking was conducted and supported by rigorous statistical analysis using Analysis of Variance (ANOVA) and Duncan’s Multiple Range Test (DMRT), which helped identify and prioritize the most important criteria for the conservation of traditional landraces.

In Western Odisha, a diverse distribution pattern of landraces was observed across five villages: Kattipali, Baghdihi, Kaitara, Pandemal, and Katikhela, as shown in Table 8. Some landraces, such as Kalabati, were found to be widely cultivated across multiple villages, particularly in large areas but by fewer households, highlighting their potential for broader dissemination. In contrast, Katia, Brahma Black, and Jalgudi were grown in small areas and by fewer households, marking them as rare and potentially threatened. Other landraces, such as Talmuli, Kurumandal, and Kusumkali, showed intermediate distribution, being cultivated in large areas but by fewer households in selected villages. Additionally, a significant number of landraces were found to be restricted to single villages, including Madabaru, Megadabral, Karaini, Kala Maliphul, Bathrash, and Purutin, indicating high localisation and vulnerability. These landraces were mostly maintained by a few households over small patches, qualifying them as rare or isolated varieties in urgent need of conservation. To enhance the interpretability of data on the extent and distribution of landraces across villages in Western Odisha, a heat map was constructed (Figure 2). This visualization captures the intensity of landrace cultivation based on both the number of cultivating households and the area under cultivation. Darker shades indicate higher intensity, representing landraces grown by a greater number of households and/or over larger areas. For instance, landraces such as Kusumkali and Kalabati show a strong presence across multiple villages, highlighted by consistently darker cells. In contrast, landraces like Talmuli and Brahma Black appear sparsely distributed and with lower cultivation intensity (Tables 7–9).

In Eastern Odisha, landraces such as Kalajeera, Haldichudi, and Umuriachundi were widely cultivated across multiple villages like Umuri, Sanokaudi, Nauguada, and Macchra. These were typically grown by many households and on large areas, making them common varieties with strong cultural and culinary relevance. Landraces such as Machhakanta, Sapuri, and Dangarbasumati exhibited intermediate characteristics being cultivated by fewer households or on smaller plots—suggesting a declining trend in popularity or area. Other landraces like Pokia Dhan, Pandkagura, Para Dhan, Mati Dhan, and Dodikaburi were found in very limited geographies and in very small areas, classifying them as rare and highlighting the need for targeted conservation efforts. The color intensity in the map corresponds to the level of conservation: darker shades indicate landraces that are more extensively cultivated or maintained by a higher number of households. Prominently, Kalajeera, Machhakanta, Haldichudi, Sapuri, and Umuriachundi showed higher conservation scores, particularly in Nauguada and Macchra villages. These landraces exhibited sustained conservation due to a combination of market demand, cultural relevance, and culinary value. In particular, Umuriachundi scored consistently high across four villages, signifying its widespread cultivation and cultural utility. Conversely, landraces such as Para Dhan, Mati Dhan, Gothia, Bhudei, and Dodikaburi had low conservation scores, often limited to a single village with minimal household participation. These varieties may be at risk of farm loss and represent candidates for urgent conservation attention (see Figure 3).

To better understand why certain landraces are conserved by many households while others are maintained only in small areas, matrix ranking was performed to assess custodian farmers’ preferences for conserving specific landraces. This approach also helps explain why some landraces are grown by many households but only on small plots, and why others are conserved by few households and on limited areas. Understanding these preferences is essential for identifying how such landraces can be protected and promoted. As shown in Table 10, in Western Odisha, paddy landrace Kusumkali, the most preferred criterion was its adaptive capacity. Custodian Farmers noted that Kusumkali was resistant to the attack of Brown Plant Hopper (Nilaparvata lugens), resulting in decreased yield losses compared to modern cultivars. This resilience resulted in high market demand for the black rice landraces, which led to prices ranging from Rs 60 to 250 per kilogram in retail shops, especially in villages like Baghdihi and Kattipali. Custodian farmers in Baghdihi were even able to obtain the Maximum Retail Price for Kusumkali in state mandis, emphasizing the commercial value as the second most preferred criterion. Similarly, due to its nutritional attributes, the Kalabati landrace had significant demand within local and state markets. According to custodian farmers, it is a sugar-free but rich source of carbohydrates, iron and calcium. The results align with Savitha and Kumari’s (2016) study, indicating medicinal paddy landraces were highly nutritious and rich in iron and zinc. Consequently, the foremost and third most prioritized criteria for conserving Kalabati were its commercial value and nutritious properties. The results are similar to the findings of Kinhoégbè et al. (2020), where the majority of the custodian farmers were conserving landraces due to their market value and nutritional value. Furthermore, the adaptability of this landrace to the local environmental conditions contributed to its consistent yield ranging from 20 to 30 quintals per acre. This proficiency in adapting to the climate and consistent yield positioned physiological ability as the second most preferred criterion, while adaptive capacity followed closely as the third. Custodian farmers engaged in Talmuli paddy cultivation made a noteworthy observation: Talmuli exhibited complete resistance to Brown Plant Hopper infestations, a contrast to the susceptibility observed in the modern cultivar. Furthermore, the Talmuli landrace showed a good yield ranging from 22 to 25 quintals per acre, unaffected by pest infestations. Sonagathi was preferred for its medicinal properties; it was used for eye treatment and to cure infections. The rice water of Sonagathi was used as eye drops for treating eye infections, and its consumption was also considered beneficial for maintaining optimal eye health. Bahal was cultivated for household consumption due to its good taste and, hence preferred for food and culinary purposes. In modern cultivars, the most preferred criterion was the commercial value in the market. The easy availability of modern cultivars in nearby input shops and local markets, as well as easy sale in state mandis on MSP, made it commercially valuable. The yield was also more than that of landraces; hence, the physiological ability of the modern cultivar was the second preferred criterion. If we see the overall preference in Figure 4 for cultivating and growing paddy landraces in western Odisha, the most favored criterion was the adaptive capacity of these landraces. This adaptability was evident in cases like Kusumkali, Talmuli, Sonagathi, etc., which resist pests, show adaptability to local conditions, minimise yield losses and drive high market demand. The high market demand was the second most preferred criterion for prioritization and conservation of landraces. Farmers value landraces that possess adaptive traits, such as drought resistance, disease tolerance, and stable yields, which allow crops to flourish in the local environment (Brush, 2008; Table 11).

In Eastern Odisha, paddy landrace Kalajeera was most preferred due to its commercial value in the market chain and the initiatives of Odisha Rural Development and Marketing Society (ORMAS), which have created a high demand for Kalajeera in the state and even the national market. Kalajeera rice, which is also known as the “Prince of Rice” is an aromatic landrace of rice with a unique black cumin-like appearance. With the assistance of ORMAS and the district administration of Koraput, Kalajeera rice from Macchra village is being marketed and sold in New Delhi. Additionally, it is packaged and available for purchase in local retail stores with retail price going up to Rs 300 per kg. Due to this, the landraces that used to be grown by a few custodian farmers in Machhara village have passed on to other villages because of the consumer demand, higher prices in the market and scope for value addition and processing. The second most preferred criterion was its cultural significance in the farming community. Kalajeera has been cultivated in Koraput from time immemorial due to its unique aroma and taste and significance in various festivals, where dishes from Kalajeera rice were being prepared. Hence, the second and third most preferred criteria for cultivating Kalajeera were cultural significance and food and culinary habits, respectively. Similarly, by efforts of MSSRF, Jeypore and ORMAS, systematic indigenous practices were performed both in Kalajeera and Umuriachundi, leading to their higher yield and easy processing of paddy and high demand in local, state and national markets. Hence, Umuriachundi cultivation was growing too due to its commercial value created by the efforts of external agencies. The quality of pop rice made from Umuriachundi also fetched good prices in the retail market and was preferred as an evening snack at home. The rice is also used in the Chaita Parva Purab festival, a three-day religious festival of the tribes of Koraput. Therefore, the second and third preferred criteria were food culinary preference and cultural value of the landrace. Para Dhan and Dangarbasumati were most preferred for their agronomic benefits in the field. Para dhan has weed suppressor properties while growing on the field and was also disease and pest-resistant, hence the two most preferred criteria were agronomic benefits (weed suppressing) and adaptive advantage (resistant to disease and pest. It is also preferred for the food and culinary habits of households, as the rice is suitable for making rice flakes. Dangarbasumati is known for its hardy nature (agronomic benefits) and natural resistance to insects and pests (adaptive advantage). The rice itself is soft and aromatic, making it ideal for preparing special dishes. The landrace of Sapuri was cultivated for various reasons. It was historically gifted to brides during weddings, symbolizing social status. Additionally, it is well-suited for producing popped rice during marriage ceremonies. The market demand for the Sapuri landrace surges during the wedding season due to these cultural practices. Moreover, Sapuri is known for its resistance to lodging. Consequently, social value, cultural value, and adaptive advantage emerged as the most favoured criteria for its cultivation. In modern cultivars, the most preferred criterion was the commercial value in the market, and the second one was their higher yield, but the input cost for cultivating these varieties is higher. Overall, the most favoured criteria for conserving landraces were food and culinary preference, closely followed by the commercial value in local and state markets. This highlights the intricate interplay of economic, cultural, and practical considerations in guiding custodian farmers’ choices for preserving these diverse landraces, as shown in Figure 5.

A total of 48.33% of custodian farmers grew landraces with a high level of economic significance. Approximately 30.67% of custodian farmers cultivated landraces with a medium level of economic significance. The remaining 21% of custodian farmers grew landraces with a low level of economic significance. The mean AddValQS for these landraces was 14.49, with a standard deviation of 6.65. AddValQS scores range from 3 to 26. In the economic analysis of landraces in western Odisha, it was observed that the majority of custodian farmers who cultivated these landraces exhibited a high level of economic significance (Table 12). This observation strongly indicated that these landraces held substantial economic and commercial value in the region. The average mean score for these landraces was notably high, and the range extended up to 26, with a maximum score of 27. This wide range of scores implied that these landraces found their place in both local and state markets. Specifically, landraces like Kalabati and Kusumkali in Western Odisha had gained recognition and protection under the Plant Varieties and Farmers’ Rights (PPVFR) Act. Furthermore, the value-added products derived from these landraces had a market presence, indicating that there was a demand for these products among consumers. Notably, black rice, a popular landrace, was directly sold in the retail market, further underlining its economic significance. These findings collectively illustrated the substantial economic value and market potential of landraces in western Odisha. The results indicated that the Added Value Quantitative Score (AddValQS) of landraces tended to increase when the primary use of the landrace was intended for the market. This suggests that landraces primarily cultivated for self-supply were less likely to attain higher levels of economic significance. In contrast, landraces geared towards commercial markets exhibited greater economic value (Table 10). It is worth noting that formal seed banks like the Anubhav seed bank in Kattipali village and Baghdihi, along with formal NGOs such as Abhinav Krishi in Bargarh, Odisha, played a significant role in driving these promotional initiatives. Their organized efforts contributed to the increased economic significance and market presence of the landraces under consideration.

In Eastern Odisha, 49.34% of custodian farmers were found to be cultivating landraces with a high level of economic significance. 45.33%, cultivated landraces with a medium level of economic significance. A smaller proportion, 5.33% of custodian farmers, focused on landraces with a low level of economic significance. These landraces showed an average AddValQS of 16.95, with a relatively low standard deviation of 5.97 (Table 12). Among these paddy landraces of Koraput (Eastern Odisha), varieties such as Kalajeera, Machhakanta, Haldichudi, Umuriachundi, and Sapuri were particularly in high demand within the market. Additionally, value-added products derived from Machhakanta and Haldichudi were successfully marketed under local farmer labels, with support provided by ORMAS (Odisha Rural Development and Marketing Society). This evidence suggested that these landraces held substantial economic value and commercial appeal. Their popularity and market success could be attributed to external support systems, such as ORMAS, which facilitated the marketing and promotion of these landraces and their related products. The variable “Use” and the variable “Multiplication Actor” were found to be statistically significant (p = 0.000). In particular, the Added Value Quantitative Score (AddValQS) of landraces increases when the use of the landrace was mainly intended for the market, meaning that it was rather difficult for a landrace used for self-supply to hold a geographical or a commercial label, or to have a higher economic significance (Table 10). These organized groups were actively involved in multiplying seeds, utilizing diversity blocks, and establishing demonstration plots within designated areas allocated for community cultivation under the seed bank.

The economic significance (AddValQS) of landraces reflected in Figure 6 can be explained by region-specific socio-economic and institutional dynamics. The strong negative effect of Use in both regions, particularly in Western Odisha, highlights a subsistence–market trade-off: landraces primarily retained for household consumption reduce marketable surplus and opportunities for value addition, thereby lowering economic value scores. The divergent effect of Promotion reflects differences in market maturity and institutional support. In Western Odisha, promotional efforts are largely conservation-oriented and have yet to translate into measurable economic gains, resulting in a negative association. In contrast, in Eastern Odisha, structured promotion by institutions such as MSSRF, along with GI recognition for landraces like Kalajeera, enhances market visibility and demand, leading to a positive effect on economic value. The effect of Multiplication Actor underscores the importance of local control over seed systems. In Eastern Odisha, increased involvement of external actors in seed multiplication may dilute household-level economic benefits, producing a negative effect, whereas in Western Odisha, community-based seed banks focused on conservation rather than commercialization exert a neutral influence on AddValQS. Other socio-economic variables exhibit limited or inconsistent effects, as the economic value of landraces is shaped more by niche markets, institutional recognition (e.g., PPVFR protection), and value-chain integration than by conventional farmer characteristics.

Farmer-reported data may be influenced by recall bias, subjective experience, and contextual variability, and therefore cannot fully substitute for controlled experimental trials. These limitations are acknowledged and were mitigated through pilot testing of instruments, triangulation across multiple data sources (custodian farmers, institutions, literature), and the use of statistical validation techniques to reduce subjectivity. Importantly, perception-based methods are treated not as proxies for experimental agronomy but as complementary evidence essential for understanding decision-making, adoption behaviour, and the socio-ecological rationale underlying landrace conservation. While a comprehensive statistical validation of agronomic traits using multivariate techniques (e.g., Principal Component Analysis) was undertaken as part of the larger study, the detailed results are not presented in this paper due to constraints of scope and the need to maintain thematic focus. Instead, this manuscript emphasizes farmer-perceived agronomic performance, which reflects real-world, on-farm outcomes and aligns with the study’s primary objectives of documenting landraces and analysing their socio-cultural and economic significance.

Given that paddy is a staple crop across India and traditional rice landraces persist in many states, the Odisha model-linking landrace characterization with economic valuation and value addition offers a transferable framework. Other states can adapt this approach by identifying region-specific landraces and integrating them into local agro-processing, nutrition, and sustainable farming initiatives. Understanding custodian farmers’ prioritization criteria enables extension agencies to promote landraces with high ecological adaptability and socio-economic value through targeted interventions, including training on post-harvest processing, branding, and quality standards. Tools such as AddValQS demonstrate how marketability, processing potential, and cultural demand can enhance conservation sustainability by reducing custodian farmers’ opportunity costs, thereby reinforcing the integration of biodiversity conservation with rural enterprise development.