Majuro Supplemental data.xlsx

*The data file has been updated. The last column heading and footnotes have been corrected.<b>Site description and conditions</b>Majuro Atoll, the capital of the Republic of the Marshall Islands, is located in the Pacific Ocean near the equator. The maximum and minimum monthly average temperatures are 27.9 to 30.5 °C and 25.4 to 25.7 °C, respectively. The average monthly and annual precipitation is 192 to 337 mm and 3,236.5 mm, respectively. (National Oceanic and Atmospheric Administration. Retrieved March 11, 2015). Coral sand has low organic matter levels (OC 0−15 cm: 46.9 g kg−1, 15−45 cm: 10.8 g kg−1) (Deenik &amp; Yost, 2006) and high percolation rates (1.4−3.5 × 10­−­3 m s−1) (Hunt &amp; Peterson, 1980).The experimental field was located at Laura Farm (7° 8' 34" N, 171° 2' 9" E), which belongs to the Ministry of Resources and Development. The field was fallow more than six months before starting the experiments. Soil water-soluble NO3-N (0-5 cm layer; 1:2.5 soil-water extraction) was 4 µg g soil−1 by a nitrate ion meter (LAQUAtwin B-742, Horiba, Tokyo). Weeds in the field were collected, chopped, and fermented for four weeks under a roof for the first experiment (experiment 1).<br><b>Plot design and management</b>The plot size was 1.2 m × 6.0 m, and the experimental field consisted of two sets of six plots. The field size was approximately 12.5 m × 7.2 m. Sweet corn (Zea mays L.) was seeded in 2 rows at 0.5 m intervals just after land preparation. No irrigation, fertilizers, or chemicals were used in any experiment. Use of synthetic fertilizer and chemicals are forbidden in this atoll to protect the underground aquifer. Hand weeding was performed at two and five weeks after seeding (except in the treatment), and the weeds were left on the plot. Some remaining plots were used for comparison of the effect of no input in experiment 1, 2-1 and 2-2.<br><b>Experiment 1: A performance of plant residue as an alternative to copra cake</b>The experimental field was divided into four parts (each part consisting of three plots) with the same fertility based on an assessment of the original weed vegetation. The first part received fermented weed chips (6.3 Mg ha−1 in fresh weight; 5.0 Mg ha−1 in dry weight) and the second part received copra cakes (10.0 Mg ha−1 in fresh weight; 5.3 Mg ha−1 in dry weight), the conventional method used by local farmers. The third part received both weed chips and copra cakes (total 10.3 Mg ha−1 in dry weight). The fourth part did not receive an input of organic material. All the fields were cultivated with a cultivator and organic materials were mixed into soils. Sweet corn was seeded on November 26, 2013, and harvested on February 11, 2014 (77 days after seeding). Weeding was not conducted.<br><b>Experiment 2-1, 2-2: Effects of crop bed, mulching, and weeding</b>All the sweet corn stems and leaves from the previous experiment were fermented for six days until fungi began to propagate prior to use for this experiment. Eight plots were used for a 3 factors × 2 levels of factorial design: crop bed, mulching (using 1/4 of the crop residue), and first weeding (second weeding was conducted for all plots). The applied residue was 5.0 Mg ha−1 in fresh weight (3.7 Mg ha−1 in dry weight). Two control plots from the previous experiment that continued to grow sweet corn without any inputs were used as controls in this experiment. Sweet corn was seeded on February 17, 2014, and harvested on May 8 (80 days after seeding).This experiment was retried exchanging first weeding treatment plots (the other treatments were fixed). The applied residue was 2.0 Mg ha−1 in fresh weight (1.5 Mg ha−1 in dry weight) Sweet corn was seeded on May 16, 2014, and harvested on July 28 (73 days after seeding).<br><b>Experiment 3: Effects of fungi propagation, cutting, scattering, and incorporation</b>The experiment used an L12 orthogonal array design (Taguchi, 1986). The factors were: fungi propagation, cutting (approximately 10 cm), scattering and incorporation (covered by soil), or soil surface application. Plots were randomly assigned. The applied residue was 4.7 Mg ha−1 in fresh weight (2.1 Mg ha−1 in dry weight). Sweet corn was seeded on August 4, 2014, and harvested on October 21 (78 days after seeding).<br><b>Experiment 4: Evaluation of the linearity for the input and output</b>Crop residues from the previous experiment (0.4-2.0 Mg ha−1) were returned to each plot by the combination of scattering and incorporating. Namely, the quantity of applied residues was proportional to the preceding yield in each plot. Sweet corn was seeded on October 28, 2014, and harvested on January 9, 2015 (73 days after seeding).<br><b>Determinations and statistical analyses</b>Whole fresh kernels from each plot were weighed. The total biomass of sweet corn was calculated from the fresh kernel weight by assuming total dry matter/kernel fresh weight as 0.36 (Miura &amp; Watanabe, 2002). The p-values were calculated as non-paired, one-sided, and with an unequal variances condition t-test. Residual effects on total dry matter of the following experiment were examined by simple correlation (Pearson product-moment) analyses.

本数据文件已更新，最后一列的标题及脚注均已修正。
<b>场地概况与环境条件</b>
马朱罗环礁（Majuro Atoll）是马绍尔群岛共和国的首都，坐落于赤道附近的太平洋海域。其月均温极值分别为27.9~30.5 ℃（最高）与25.4~25.7 ℃（最低），月均降水量与年降水量分别为192~337 mm与3236.5 mm。数据源自美国国家海洋和大气管理局（National Oceanic and Atmospheric Administration），检索于2015年3月11日。
珊瑚砂的有机质含量较低（0~15 cm土层：46.9 g·kg⁻¹；15~45 cm土层：10.8 g·kg⁻¹）（Deenik & Yost, 2006），且入渗速率较高（1.4~3.5 × 10⁻³ m·s⁻¹）（Hunt & Peterson, 1980）。
试验田设于劳拉农场（Laura Farm），坐标为北纬7°8′34″、东经171°2′9″，隶属于马绍尔群岛资源与发展部。试验开始前，该地块已休耕超过6个月。
采用硝酸根离子检测仪（LAQUAtwin B-742，堀场（Horiba），东京）测得，0~5 cm土层中土壤水溶性硝态氮（NO₃-N）含量为4 μg·g⁻¹（土水比1:2.5浸提）。
为开展首项试验（试验1），研究人员采集田间杂草并切碎，于棚下发酵4周。
<b>小区设计与田间管理</b>
试验小区规格为1.2 m × 6.0 m，试验田共设两组，每组6个小区，试验田总面积约为12.5 m × 7.2 m。
甜玉米（Zea mays L.）于整地完成后，按0.5 m的间隔种植2行。所有试验均不进行灌溉、施肥及施用农药。为保护该环礁的地下含水层，本试验田禁止使用合成肥料与农药。
除特定处理组外，分别于播种后2周和5周进行人工除草，杂草留存于小区内。
剩余部分小区用于对照试验，以评估无外源投入对试验1、2-1及2-2的影响。
<b>试验1：植物残体替代椰干饼的效果评估</b>
基于原始杂草植被的评估结果，试验田被划分为4个肥力均一的区块，每个区块包含3个小区。
第一区块施用发酵杂草碎料（鲜重6.3 Mg·ha⁻¹，干重5.0 Mg·ha⁻¹）；第二区块施用椰干饼（鲜重10.0 Mg·ha⁻¹，干重5.3 Mg·ha⁻¹），此为当地农户的常规施肥方式；第三区块同时施用杂草碎料与椰干饼，总干重施用量为10.3 Mg·ha⁻¹；第四区块不添加任何有机物料。
所有区块均使用中耕机整地，并将有机物料翻混入土壤中。甜玉米于2013年11月26日播种，2014年2月11日收获（播种后77天），试验期间未进行除草作业。
<b>试验2-1、2-2：作床、覆盖与除草的效应</b>
本试验所用物料为前一试验收获的甜玉米茎叶，经6天发酵至真菌开始繁殖后使用。
本次试验采用3因素×2水平的析因设计，共设置8个小区，因素分别为作床处理、覆盖处理（施用1/4作物残体）与首次除草处理，所有小区均需开展第二次除草作业。本次试验的残体施用量为鲜重5.0 Mg·ha⁻¹（干重3.7 Mg·ha⁻¹）。
本试验的对照小区选自前一试验中未施加任何外源投入、持续种植甜玉米的2个小区。甜玉米于2014年2月17日播种，2014年5月8日收获（播种后80天）。
本次试验进行了重复，交换了首次除草处理的小区（其余处理组保持不变），该重复试验的残体施用量为鲜重2.0 Mg·ha⁻¹（干重1.5 Mg·ha⁻¹）。甜玉米于2014年5月16日播种，2014年7月28日收获（播种后73天）。
<b>试验3：真菌繁殖、切碎、撒施与翻混的效应</b>
本次试验采用L12正交表设计（田口，Taguchi, 1986），试验因素包括：真菌繁殖处理、切碎处理（切段长度约10 cm）、撒施翻混处理（覆土）与土表撒施处理。小区采用随机区组排列。
本次试验的残体施用量为鲜重4.7 Mg·ha⁻¹（干重2.1 Mg·ha⁻¹）。甜玉米于2014年8月4日播种，2014年10月21日收获（播种后78天）。
<b>试验4：投入与产出的线性关系评估</b>
将前一试验收获的作物残体（施用量为0.4~2.0 Mg·ha⁻¹）通过撒施与翻混结合的方式返还至各小区，即残体施用量与各小区前一茬作物的产量成正比。
甜玉米于2014年10月28日播种，2015年1月9日收获（播种后73天）。
<b>测定与统计分析</b>
称取各小区全部甜玉米鲜籽粒的重量。基于总干物质/鲜籽粒重量比值为0.36的假设（Miura & Watanabe, 2002），由鲜籽粒重量推算甜玉米的总生物量。
采用非配对、单侧、方差不齐的t检验计算p值。采用简单相关（皮尔逊积矩相关）分析，评估前茬试验对后茬试验总干物质的残留效应。