s:1970:"TI Pre-rice green manure production in rainfed environments : a simulation approach 
AU Garrity, D.P. 
AU Flinn J C 
AB The rainfed lowland ricelands prone to waterlogging in the pre-rice season are a major potential niche for the use of farm-grown green manure crops tolerant to saturated soils. But year-to-year variability in green manure growth and N accumulation would be large due to quite variable field hydrology in the dry-to-wet transition months. We developed a simulation model of the green manure-rainfed rice system that was used to estimate the yield levels and relative stability of pre rice green manures and the subsequent rainfed rice crops in three representative sites in southeast Asia. The average estimated green manure N yield over 25 years was higher in Los Bafios, Philippines (65 kg N/ha) than in Ubon, northeast Thailand (40 kg N/ha) or Tuguegarao, northern Luzon, Philippines (18 kg N/ha), with yield stability following the same trend. Simulation of sesbania planting was compared when established early (at 50 mm cumulative rainfall) or later (loo mm) at the beginning of the rainy season. N yields were higher for late planted crops (63 kg N/ha) compared to early establishment (51 kg N/ha) due to more favorable soil moisture. Estimated rice yields following the early and late establishment of sesbania were similar (2.9 and 3.0t/ha) indicating that delayed sesbania planting would riot adversely affect the rice. Optimal N-rates for rice were similar (85-86 kg N/ha) for both planting dates, consequently more purchased N is needed for earlier green manure establishment. The costs per unit of green manure N were higher than for urea N when full wage rates andintensive tillage were assumed. However, with zero tillage and/or a modest cumulative soil fertility benefit, the Costs of producing N with. green manures would be similar or lower than with urea-N. The implications of these finding for future research are discussed. 
";