Climate and landscape mediate patterns of low lentil productivity in Nepal

practices and varieties (2012/13–16/17), and 3) simulations with the Stempedia model [49] to assess inter-annual variations in Stemphylium disease risk and to establish the risk-reducing value of modifying planting dates. Materials and methods This research uses different types of dataset and followed the standard ethics during the data collection. First, no personal identifiable information was recorded in the experimental trails data. Only farmers’ plots were used to establish experimental trials in the study areas, after tak- ing consent from the farmers to use their plots. Second, for the farm survey, the authors fol- lowed the donor’s standard ethics protocol regarding the human subject research during the data collection. The authors have also taken the consent from the farmers before the survey so that no any personal identifiable information would be disclosed during data sharing. Site description On-farm experiments were conducted over five seasons (2012–2016 harvest years) to evaluate the influence of agronomic interventions and varietal choice on lentil yields in two contrasting ecologies: the mid-hills (Surkhet and Dadeldhura districts) and the Terai plain (Banke, Bar- diya, Kailali, and Kanchanpur districts). The two districts in the mid-hills are located at 28.01˚ to 28.59˚ E latitude and 80.12˚ to 81.36˚ N longitude and 700–1800 m altitude, while the four districts in the lowland Terai are located at 27.51˚ to 29.28˚ E latitude and 80.01˚ to 82.8˚ lon- gitude and range of 100–700 m altitude. The experimental sites in the mid-hills are well-

drained (upland) with maize (April-August) and lentil (September to April) grown in the annual rotation, while the Terai sites are poorly drained (lowland) with rice (June-October) and lentil (October-March) grown in the annual cropping system rotation. On-farm experiments Three categories of on-farm experiments were conducted: (1) evaluation of different crop establishment methods: drainage modification (bed planting), zero tillage, minimum tillage (single pass), and hand broadcasting into conventionally-tilled soil, (2) varietal evaluations using recently released varieties by Government of Nepal, advanced genotypes (in the process of release) and locally available varieties grown by the farmers, and (3) integrated manage- ment, i.e., improved variety, fertilizer application (recommended rates of 20:40:20 kg N:P2O5: K2O ha-1), mechanized seeding contrasted to prevailing farmer practices of hand broadcasting of local seed with conventionally-tilled soil and no mineral fertilizer. Details of the year and location for these experiments are presented in S1 Table. In each of the districts experimental areas were located at 1 to 25 km distance. For each experiment, analysis of variance (ANOVA) was conducted to assess treatment and year effects along with their interaction. Household surveys The survey was conducted for two lentil producing years (2013–14 and 2014–15, hereafter referred as 2014 and 2015, respectively). In 2014, the survey was conducted in four major lentil growing districts: Banke and Bardyia in the Mid-western Terai region along with Kailali and

Kanchanpur in the Far-western Terai region. Based on the lentil acreage in each of the sub-dis- tricts (village development committees or VDCs), 10 VDCs were purposively selected to cap- ture major lentil growing areas. With these VDCs, a total of 193 lentil-growing households were randomly selected for the survey. In 2015, the survey was expanded to six districts by including two additional districts from the mid-hills, i.e., Surkhet (Mid-western region) and PLOS ONE Climate and landscape mediate low lentil productivity PLOS ONE | https://doi.org/10.1371/journal.pone.0231377 April 16, 2020 3 / 19