Climate and landscape mediate patterns of low lentil productivity in Nepal

significantly higher than the long-term average and is reflected in the survey data with 27% of farmers reporting that their lentil fields were affected by waterlogging. The average land area under lentil production was 0.26 ha with a grain yield of 90 kg ha-1. In the region, 57% of farm- ers cultivated lentil in a mixed cropping system with mustard, wheat and lathyrus, and 7% of farmers practiced relay seeding of lentil with rice. The majority of farmers (91%) were using a local variety of lentil, with very low use of purchased fertilizers with average application rate for nitrogen, phosphorus, and potassium of 3.23, 6.31, and 0.65 kg ha-1, respectively. Sixty-five percent of farmers grew lentil in lowland areas which are dominated by clay to silty loam soils and vulnerable to waterlogging during periods of high rainfall. About 76% of farmers reported that their lentil field suffered from diseases (Table 1). In relatively wet winters included in the survey, lentil productivity in upland ecologies with better drainage (397 kg ha-1) was significantly higher than in lowland ecologies (329 kg ha-1) (Table 2). On the other hand, although still modest, investment in fertilizer was significantly higher in lowland ecologies (US$ 12.8 ha-1) than in upland (US$ 2.4 ha-1). The total labor cost involved in lentil cultivation was significantly higher in the upland ecologies (US$ 201 ha-1) than in lowland (US$ 117 ha-1). The higher labor cost in upland ecologies is associated with limited mechanization options. However, the non-labor capital cost was significantly higher in

lowland (US$ 57.8 ha-1) than in upland ecologies (US$ 46.6 ha-1). The total variable cost (US$ 295 ha-1 in upland vs. US$ 237 ha-1 in lowland) and gross margin (US$ 53.5 ha-1 vs US$ 22.2 ha-1) were significantly higher in the upland than in lowland ecologies. Despite having signifi- cantly higher yields and gross margins, the benefit–cost ratio is significantly lower in upland environments due to higher production costs. The spatial distribution of rainfall and average temperature during lentil production season for surveyed years are presented in Fig 1. There was a high within and across seasonal variabil- ity in rainfall among sampled districts. Winter rainfall variability in 2014 ranged from 19 mm to 340 mm, while the variability widened to 86 mm to 657 mm in 2015. There was less variabil- ity in average temperature during surveyed years (Fig 1). Long-term trend analysis for seasonal rainfall shows a positive slope indicating winters are becoming wetter (S1 Fig). Similar results were reported in earlier studies conducted in western Nepal [40,75]. Table 2. Lentil production characteristics differentiated by production ecologies. Variables Upland (N = 278) Lowland (N = 515) Sig. Overall (N = 793) Lentil yield (kg ha-1) 397.3 329.2 ��� 353.1 Seed cost ($ ha-1) 44.5 49.8 ns 47.9 Fertilizer cost ($ ha-1) 2.4 12.8 ��� 9.1 Labor cost ($ ha-1) 201.3 116.8 ��� 146.5 Non-labor capital ($ ha-1) 46.6 57.8 �� 53.9 Total variable cost ($ ha-1) 294.7 237.3 ��� 257.4 Gross Revenue ($ ha-1) 348.2 259.5 ��� 290.6

Gross Margin ($ ha-1) 53.5 22.2 ��� 33.2 Benefit cost (B:C) ratio† 1.30 1.35 ��� 1.33 ��� indicates significant at 1% level, �� indicates significant at 5% level, and “ns” indicates non-significant. Across production ecologies comparison are based on Mann-Whitney test. Exchange rate: 1 US $ = NRs 98.6 and NRs. 102.7, respectively, for year 2014 and 2015 [86]. †The benefit cost ratio was calculated by dividing gross revenue with the total variable cost. https://doi.org/10.1371/journal.pone.0231377.t002 PLOS ONE Climate and landscape mediate low lentil productivity PLOS ONE | https://doi.org/10.1371/journal.pone.0231377 April 16, 2020 8 / 19