ABSTRACT Brazil accounts for 32% of world soybean production and potential yield losses related to climate variability in the country can have an adverse effect on global supply. Recent soybean production declines in Brazil (e.g. The 15.6% soybean yield drop occurred during the 2011/12 growing season) were a consequence of extreme weather events caused by the effects of the El Nino Southern Oscillation (ENSO) phenomenon. ENSO is one of the most important atmospheric-oceanic phenomena responsible for climate and crop yield variability in several agricultural producing regions of Brazil. The greatest challenge related to adapting to ENSO climate cycles is to understand how to minimize the negative impacts and take advantage of favorable conditions to improve yields. Thus, the aim of this study was to determine the spatial and temporal impacts of sowing dates and ENSO phases on soybean yield and the frequency of extreme weather events (high, low temperature and dry period) during the vegetative and reproductive soybean phenological phases in southern Brazil. For that, soybean yields for 16 different sowing dates (from September to February) were simulated using a well-calibrated crop growth and development model (CROPGRO-Soybean) for 12 locations across southern Brazil and considering historical weather series of 56 years (1961-2016). Homogeneous regions related to ENSO impacts were defined by cluster analysis to facilitate the understanding and use of the information produced. Alternative ENSO phases were found to affect the frequency of extreme weather events and soybean yields across the region studied. Potential negative ENSO effects could be minimized by adjusting the sowing windows according to the expected climate cycle. Our results revealed that ENSO impacts on soybean yields in southern Brazil vary according to the location and can be categorized into three main groups: Group I, where the impacts of the El Nino phase on soybean yield tend to be positive, and its positive impact did not vary with the different sowing dates; Group II, where the negative soybean yield anomalies were more frequent in Neutral years; and Group III, where the positive and negative impacts of ENSO phases varied with sowing dates.

中文翻译：

---

厄尔尼诺南方涛动现象和播种日期对巴西南部大豆产量和极端天气事件发生的影响

摘要巴西占世界大豆产量的 32%，与该国气候变化相关的潜在产量损失可能对全球供应产生不利影响。巴西近期大豆产量下降（例如，2011/12 生长季大豆产量下降 15.6%）是厄尔尼诺南方涛动 (ENSO) 现象影响导致的极端天气事件的结果。ENSO 是导致巴西几个农业产区气候和作物产量变化的最重要的大气海洋现象之一。与适应 ENSO 气候循环相关的最大挑战是了解如何最大限度地减少负面影响并利用有利条件来提高产量。因此，本研究的目的是确定播种日期和 ENSO 阶段对巴西南部大豆生长和繁殖期大豆产量和极端天气事件（高温、低温和干旱期）频率的空间和时间影响。为此，使用经过良好校准的作物生长和发育模型 (CROPGRO-Soybean) 在巴西南部 12 个地点并考虑 56 年（1961 年至 2 月）的历史天气序列，模拟了 16 个不同播种日期（从 9 月至 2 月）的大豆产量。 2016）。与 ENSO 影响相关的同质区域通过聚类分析来定义，以促进对所产生信息的理解和使用。研究发现，其他 ENSO 阶段会影响所研究地区极端天气事件的频率和大豆产量。通过根据预期的气候周期调整播种窗口，可以最大限度地减少潜在的负面 ENSO 影响。我们的研究结果表明，ENSO 对巴西南部大豆产量的影响因地点而异，可分为三大类：I 组，其中厄尔尼诺现象对大豆产量的影响往往是积极的，其积极影响确实不随播期不同而变化；第二组，大豆产量负异常在中性年更为频繁；和第三组，其中 ENSO 阶段的正面和负面影响因播种日期而异。我们的研究结果表明，ENSO 对巴西南部大豆产量的影响因地点而异，可分为三大类：I 组，其中厄尔尼诺现象对大豆产量的影响往往是积极的，其积极影响确实不随播期不同而变化；第二组，大豆产量负异常在中性年更为频繁；和第三组，其中 ENSO 阶段的正面和负面影响因播种日期而异。我们的研究结果表明，ENSO 对巴西南部大豆产量的影响因地点而异，可分为三大类：I 组，其中厄尔尼诺现象对大豆产量的影响往往是积极的，其积极影响确实不随播期不同而变化；第二组，大豆产量负异常在中性年更为频繁；和第三组，其中 ENSO 阶段的正面和负面影响因播种日期而异。