The endosperm starch of rice is classified as glutinous or non-glutinous by the amylose content. However, among growers and consumers, the opaque endosperm of glutinous rice distinguishes it from the translucent endosperm of non-glutinous grain. This distinction is confirmed by the glue-like stickiness of the cooked rice. In this study, most of the rice identified by farmers as glutinous contained 2.1-5% amylose, in contrast to the 0–2% amylose definition of glutinous rice in scientific and technical literature (Dela Cruz and Khush, 2000; TRKB 2024). Moreover, an even wider range of amylose content, from 0–15%, was reported for the rice grown and consumed as glutinous rice in Laos (Schiller et al. 2006). The prevalence of samples with > 10% amylose indicates preference for non-glutinous rice among Thailand’s highland minority groups, in common with the country at large but different from the Northern Thai population in lowland Chiang Mai, Chiang Rai and other upper northern provinces, who traditionally consume glutinous rice as the staple. The preference for glutinous rice among the Khmu farmers was found here, as it was so reported for a much larger population of the ethnic group residing in Laos (Schiller et al. 2006). The few glutinous rice varieties from non-glutinous rice consuming villages are used for special occasions, e.g. for making sweet rice dishes and snacks, the sticky rice cakes for the new year and other celebrations, or for brewing alcohol. The most common rice variety, grown by almost every household, in a non-glutinous rice consuming Karen village, was a special glutinous rice variety for brewing (Sirabanchongkran et al. 2004). Rice in Thailand, the majority of which is cultivated and consumed in the lowlands, has been dominated by the slender grain type since the 18th century (Watabe et al., 1970), and become the norm in the present (Calingacion et al. 2014; Prom-u-thai and Rerkasem 2020). The large-grain landraces from the highlands found in the survey are clearly differentiated from the predominantly slender-grain landraces of the lowlands (Rerkasem and Rerkasem 2002).
While most of the samples had colourless pericarp, which is milled into white rice, there is an apparent tolerance of red rice admixture that is considered a quality damaging and price lowering contamination in the milled rice trade (TMOC 2016). Also notable is the presence of local rice varieties with pigmented pericarp, especially those with the purple colour. In addition to their traditional usages, interest in the pigmented rice among consumers and industry is being stimulated by identification of bioactive compounds associated with the colour, such as anthocyanins and phenolics (Yodmanee et al. 2011; Bhat et al. 2020). With prices in the same range or higher than Hom Mali, Thailand’s premier quality rice retailed at triple or more the prices of ordinary rice, commercialization of purple rice production in the highlands is not surprising. Rice with red pericarp is also produced and marketed for premium prices in the lowlands with varieties such as Red Mali or Red Hom Mali, a natural mutant of KDML105 (Maksup et al. 2019; TRKB 2024), and Sang Yod, a local variety from the vicinity of Songkhla Lake in the South (Panomjan et al. 2016). Unlike purple or black rice these are commonly referred to by their variety names, and not ‘red rice’, possibly to avoid the sub-standard quality connotation of the term and to prevent confusion with ‘red rice’, a weed of the rice field in the Americas (Noldin 2000). The landraces with red pericarp from the highlands deserve a closer examination for their commercial potential.
Among the main staple crops, rice has the lowest concentration of Fe and Zn, which has contributed to the low intake and a deficiency in these micronutrients in a substantial share of the world’s rice consumers (Welch and Graham 2002; Bouis and Welch 2010). This study has established that rice grown for subsistence in the highlands is beneficial to the local consumers with a higher dietary Fe intake. In commercial production, on the other hand, rice with higher grain Fe and Zn content that can be produced in the field need to be kept segregated from non-enriched rice in the value chain without incurring additional cost to the target consumers, such as the urban poor (Prom-u-thai and Rerkasem 2020). No information is available on the benefit to the rice crop grown from seed with high Fe. The immobilization of Fe3+ in aerobic soil means that the impact of the seed Fe reserve is more likely to be significant in upland rice than in wetland rice, where Fe2+ is readily available in the waterlogged soil. Applying Fe to germinating upland rice seed has been found to stimulate germination and seedling growth (Wang et al. 2020).
Lowland grown rice in Thailand, from KDML105, RD6, along with other improved local varieties and modern varieties are largely moderate in their grain Zn, with 21–26 mg Zn kg− 1, and lower in some commonly grown modern varieties, at 17–18 mg Zn kg− 1 (Saenchai et al. 2012). With more than 60% of the highland rice samples in the moderate Zn range, plus another 20% in the lower Zn range, no sample exceeded 30 mg Zn kg− 1, the highland grown rice does not appear to be a better source of Zn than the lowland grown rice. The higher anthocyanin content observed in most of the farmers’ purple rice samples compared with the lowland grown Kam Doi Saket suggests a potential advantage for purple rice production in the highlands. In Laos, the largest number of rice landraces with purple pericarp were found in the upland rice agroecosystem under slash-and-burn cultivation (Appa Rao et al. 2006). Genotypic variation, with or without environmental effects, was suggested by the wide range of grain anthocyanin content, from 7 to 115 mg 100 g− 1 (data not shown) in samples identified as Bieisu in the survey. Bieisu is a generic name in the Karen language, where ‘biei’ refers to glutinous rice, and ‘su’ denotes a dark colour, similar to ‘Khao Kam’ or ‘Khao Dam’, which are generic names for rice with purple pericarp in Lao and Thai (Pusadee et al. 2024).
Grown in the highland at Mae Wang, moderate levels of grain Zn were produced by the local variety Bue Bang and three improved local varieties (Exp. 1), including Sew Maejan, which was rated high grain Zn (45 mg Zn kg− 1) when grown with adequate Zn in the soil at CMU (Saenchai et al. 2012). The rice grain Zn content is highly variable due to the interaction between genotype and environment (Graham et al. 1997; Wissuwa et al. 2008). A modern high yielding rice variety with low Zn, CNT1, produced grain with 18 mg Zn kg− 1 when grown in soil with 2.1 mg kg− 1 extractable Zn compared with only 9 mg Zn kg− 1 when grown on soil with 0.5 mg kg− 1 extractable Zn (Phattarakul et al. 2012). Nitrogen fertilizer that generally increases yield affects the grain Zn concentration of rice varieties differently due to a close association between Zn and N nutrition (Khampuang et al. 2021). The grain Zn content of rice varieties with high grain Zn but lower yield potential was reported to decline when the yield was raised by N fertilizer, which increased the grain Zn concentration and yield simultaneously in modern high yielding rice varieties with low grain Zn (Jaksomsak et al. 2017). In the present study, however, the grain Zn content of neither Bue Bang nor KDML105 was affected by N which significantly increased yield.
The genetic potential for higher grain Zn of Bue Bang was demonstrated with the level of 57 mg Zn kg− 1 when grown with sufficient Zn in the soil in an on-farm experiment, whereas KDML105 never exceeded 33 mg Zn kg− 1. Further evidence of the genetic potential for higher grain Zn in certain local rice varieties in the highlands was previously provided in a study involving 20 landraces, with an average grain Zn content of 24 ± 3 mg Zn kg− 1 when grown on a low-Zn soil in farmers’ fields in the highlands, increasing to 32 ± 4 mg Zn kg− 1 when grown on a high-Zn soil in the lowlands (computed from data in Jaksomsak et al. 2015). Additionally high grain Zn levels ranging from 40 to 60 mg Zn ha− 1 were observed in a number of improved local upland rice varieties when grown at CMU in the lowland (Saenchai et al. 2012). Despite the average to low level of Zn found in the rice samples from the survey, a genetic potential for higher grain Zn is clearly present in the highlands of Northern Thailand. Individuals with significantly higher seed Zn content than the bulk samples have been identified by staining rice seeds with Zn-specific dithizone (Jaksomsak et al. 2015) and analysis for seed Zn content in single seed descent lines (Sreethong et al., 2020). The seed Zn supply is crucial to the survival and establishment of rice seedlings on soils with limited Zn availability (Slaton et al. 2001; Boonchuay et al. 2013). A subtle selection pressure for rice with high grain Zn may be associated with the practice of slash-and-burn cultivation of upland rice, which creates pockets of ash with raised alkalinity that limits Zn availability. The impact of this is not perceptible in the Zn content of rice samples determined by atomic absorption spectrometry, which typically provide average contents of some 50–100 seeds, ground and blended together before further sub-sampling for treatment with reagents and other procedures.
Rice with higher grain anthocyanin contents is sometimes produced from local varieties originating as upland rice in the highlands than those from the lowland but not always, as can be seen from results of the two on-farm experiments. The effect of water regime on anthocyanin content of rice grown in the highlands was significant in the upland varieties but not the wetland variety (Fig. 4). Under wetland cultivation anthocyanin content was higher in Bieisu than Kam Doi Saket only at Mae Wang, but not at a higher or lower elevation, at Omkoi and CMU (Fig. 6). Variation in the G x E effects on anthocyanin content in the grain of purple rice have also been reported by others. Under upland cultivation, at 330 m elevation the highest anthocyanin content was produced by an improved local variety originating in the highlands, while at 800 m it was achieved by a lowland variety (Rerkasem et al. 2015). However, it should be noted that the designation of ‘wetland’ and ‘upland’ for local rice varieties simply reflect the way in which the varieties have been cultivated by farmers. When grown in the lowlands, Bieisu and Kam Hom CMU both had consistently higher grain anthocyanin than the lowland variety Kam Doi Saket (Jaksomsak et al. 2021). Confirmation of the advantage of the highlands in the production of purple rice would yield economic benefits for farmers with limited options for cultivating higher-value crops, potentially leading to a Geographic Indication registration, similar to that applied to Sang Yod rice from the South (DIP, 2006; Petruang and Napasintuwong 2022).
In conclusion, the present study has demonstrated how a direct benefit to the local population with a higher dietary intake of Fe was realized with the rice produced from a local rice germplasm in the highlands of Norther Thailand. The highland grown rice does not appear to be a better source of Zn than the rice produced from mega-varieties grown in the lowlands, although a genetic potential for high grain Zn may be present among the local landraces. The highlands have also been identified as a rich source of pigmented rice germplasm and an environment favouring the production of premium quality purple rice with high anthocyanin content. Confirmation of the advantage of the highlands in the production of purple rice would yield economic benefits for farmers with limited options for cultivating higher-value crops. On-farm conservation efforts concrete evidence of direct benefits to local communities should be more effective and sustainable than rewards offered to the farmers from external funding.