Yield gap assessments at the field-level are key for developing management practices to increase crop production in a sustainable manner. Although rice is an important food crop worldwide, yield gaps remain less investigated in Latin American rice systems. In this study, we evaluated field-level factors for closing yield gaps (based on attainable farm yield, defined as the mean of top 10 % of farmers) for rice production in Uruguay using crop management and yield records from 2012 to 2017, covering approximately 70,000 ha per year (40 % of total rice area). The mean annual attainable yield gap ranged from 16 % to 22 % in fields with non-hybrid cultivars (90 % of planted area) and from 14 % to 22 % in fields with hybrid rice (10 % of planted area). Early seeding was identified as the most influential factor for reducing yield gaps in both systems, followed by N rate. Stand establishment was also important for closing yield gap in non-hybrid fields, while rotation with improved pasture was important in hybrid fields. When variables were categorized as input-related, manageable, or non-manageable, on average manageable factors (e.g. early planting and stand establishment) were more important than input-related factors (e.g. seed or fertilizer rate). This study highlights a simple, self-contained method using large field-level datasets to quantify yield gaps and develop strategies for improving agricultural productivity.

在田间进行单产差距评估是制定管理实践以可持续方式提高作物产量的关键。尽管稻米是世界范围内重要的粮食作物，但拉丁美洲稻米系统中尚未对单产差距进行调查。在这项研究中，我们使用2012年至2017年的作物管理和产量记录评估了乌拉圭稻米生产的缩小收成差距的实地水平因素（基于可达到的农业产量，定义为农民的前10％的平均值）每年70,000公顷（占水稻总面积的40％）。在非杂交品种（占种植面积的90％）的田地上，年均可达到的产量差距在16％至22％之间，在杂交水稻（占种植面积的10％）上，在14％至22％之间。早播被认为是减少两个系统产量差距的最有影响的因素，其次是N率。在非杂交田中，建立林分对于缩小产量差距也很重要，而在改良田中，改良牧场的轮作也很重要。当变量被分类为与输入有关的，可管理的或不可管理的时，平均而言，可管理因素（例如早期种植和林分建立）比与输入相关的因素（例如种子或肥料比率）更为重要。这项研究强调了一种简单的，自包含的方法，该方法使用大型田间数据集来量化产量差距并制定提高农业生产率的策略。平均而言，可管理因素（例如早播和林分建立）比投入相关因素（例如种子或肥料用量）更重要。这项研究强调了一种简单的，自包含的方法，该方法使用大型田间数据集来量化产量差距并制定提高农业生产率的策略。平均而言，可管理因素（例如早播和林分建立）比投入相关因素（例如种子或肥料用量）更重要。这项研究强调了一种简单的，自包含的方法，该方法使用大型田间数据集来量化产量差距并制定提高农业生产率的策略。