Replication Data for: Model-based evaluation of rainfed lowland rice responses to N fertiliser in variable hydro-edaphic wetlands of East Africa

In East Africa, rainfed lowland rice is primarily produced by smallholders in alluvial floodplain and inland valley wetlands. These wetlands differ in their dominant soil types and water regimes that vary seasonally, inter-annually and between field positions. Yield responses to mineral nitrogen (N) fertiliser thus likely vary between and within wetlands and years, modulating the profitability of N fertiliser use. Therefore, the locally-validated APSIM model was used to study yield responses to N fertiliser rates (0, 30, 60, 90, 120, and 150 kg ha-1) and supplemental irrigation at different field positions in a floodplain in Tanzania (fringe and middle positions) and an inland valley in Uganda (valley-fringe, mid-valley and valley-bottom positions) over 30-years. Average rainfed yield gains with mineral N and N use efficiencies were high, ranging between 1.7 and 4.5 Mg ha-1 and 27–70 kg kg-1 in the floodplain and between 1.0 and 3.2 Mg ha-1 and 18–34 kg kg-1 in the inland valley, depending on field position, N rate and year. Consequently, N fertiliser use was generally profitable in both wetlands, with value/cost ratios ≥ 4 and marginal rates of returns > 150%. Profitable N rates in all years were 30–120 kg ha-1 in the fringe and 30–90 kg ha-1 in the middle positions of the floodplain, and 60–150 kg ha-1 in the mid-valley and 90–150 kg ha-1 in the valley-bottom positions of the inland valley. In the valley-fringe position, N fertiliser use was comparatively riskier and profitable only in 77–90% of years at N rates of 60–150 kg ha-1. Supplemental irrigation may help boost N fertiliser use efficiencies and use profitability with average yield gains of > 1.5 and > 0.4 Mg ha-1 in the floodplain and inland valley, respectively, while simulated spatial-temporal water stress pattern may help guide efficient irrigation scheduling.

东非地区的雨养低地水稻（rainfed lowland rice）主要由小农户在冲积洪泛平原与内陆河谷湿地中种植。这类湿地的主导土壤类型与水分状况存在显著差异，且其随季节、年际尺度以及田块位置的不同而动态变化。因此，矿质氮肥（mineral nitrogen, N）的产量响应效应在不同湿地间、湿地内部以及不同年份均存在差异，进而调控氮肥施用的盈利性。为此，研究采用经本地验证的APSIM模型（APSIM），针对坦桑尼亚一处洪泛平原的两类田块位置（边缘区与中区），以及乌干达一处内陆河谷的三类田块位置（河谷边缘区、谷中地带与河谷底部区），设置0、30、60、90、120、150 kg·ha⁻¹共6个氮肥施用量梯度与补充灌溉处理，开展了为期30年的模拟研究。模拟结果表明，受田块位置、氮肥施用量与年份的共同影响，洪泛平原的雨养水稻在施用矿质氮肥后的平均产量增幅介于1.7~4.5 Mg·ha⁻¹，氮肥利用效率介于27~70 kg·kg⁻¹；内陆河谷的平均产量增幅介于1.0~3.2 Mg·ha⁻¹，氮肥利用效率介于18~34 kg·kg⁻¹。整体而言，两类湿地的氮肥施用均具备盈利性，产投比（value/cost ratios）≥4，边际回报率（marginal rates of returns）超过150%。洪泛平原边缘区与中区的最优盈利施氮量分别为30~120 kg·ha⁻¹与30~90 kg·ha⁻¹；内陆河谷谷中地带与河谷底部区的最优盈利施氮量分别为60~150 kg·ha⁻¹与90~150 kg·ha⁻¹。河谷边缘区的氮肥施用风险相对更高，仅在77%~90%的年份中，当施氮量为60~150 kg·ha⁻¹时具备盈利性。补充灌溉可有效提升氮肥利用效率与施用盈利性，洪泛平原与内陆河谷的平均产量增幅分别可达1.5 Mg·ha⁻¹以上与0.4 Mg·ha⁻¹以上；同时，模拟得到的时空水分胁迫格局可为高效灌溉调度提供科学指导。