中国农科院环发所在黄河流域玉米水分利用效率阈值及关键驱动因子研究中取得新进展

由于农业灌溉用水短缺，以及气候变化下干旱风险的增加，确定作物水分利用效率（WUE）沿干湿梯度的制约因素非常重要。本研究将CMIP6五个高分辨率气候模式与DSSAT-CERES-Maize模型嵌套，并采用网格化偏差修正方法对作物模型进行了优化，阐明了三种气候情景（SSP126、SSP370和SSP585）下的干旱风险，揭示了递阶干湿梯度的WUE阈值及其关键驱动因子。结果表明，黄河流域未来干旱风险将持续存在，2030s西南区（V）干旱频率最高，西北区（III）干旱频率最低，SSP126情景下2080s西南区（V）可能会出现转湿趋势。修正后的作物模型能够有效模拟作物产量和蒸散量（ET），nRMSE分别降低4.00%和9.73%。在不同气候情景和未来时段中，各玉米生态区均存在 WUE 阈值（1.96-8.41 kg ha−1 mm-1），主要集中在轻度和中度干/湿条件下。东部（I和II）的WUE 主要受产量要素驱动，而西部（III、IV和V）则主要受ET驱动。研究成果对未来气候变化下，黄河流域不同区域实施差异化水分管理并优化农业技术途径提供了科学支撑。

Fig. 1. Five maize ecoregionsin the YRB of China.

Fig. 2. Land surface air temperature and precipitation in the YRB based on observations and CMIP6 simulations.(a), (b) are the observed (black line) and simulated changes in annual land surface air temperature (°C) and annual mean month precipitation (mm mth⁻¹) during the historical (grey line) period of 1985–2014 and future periods of 2030s, 2050s and 2080s under SSP126 (pink line), SSP370 (red line) and SSP585 (green line) across 5 CMIP6 models. The lines represent the ensemble mean time series, and the shading shows the uncertainty (25%-75%) in terms of the interquartile range across ensemble members.

Fig. 3. Drought frequency, SPEI temporal trends and their spatial changes in historical periods, and drought frequency in future periods. (a)The probability density function (PDF) of the drought frequency during the historical (1985–2014) period in each maize ecoregion. The dotted lines of different colors are the mean lines of drought frequency in each maize ecoregion. (b) The drought trend of the maize ecoregions in the YRB over the period 1985–2014. The red dotted line in (b) indicates that the SPEI trend is 0. In (b), the rose red symbols represent grid points that had significant trends (p < 0.05) in SPEI, while the blue symbols represent grid points with insignificant trends. (c) The spatial distribution of SPEI changing trend for the historical period (1985-2014) in the YRB. (d) The PDF of the drought frequency during 2021–2040, 2041–2070 and 2071–2100 (columns) and for the SSP126, SSP370 and SSP585 climate scenarios (rows) at the five maize ecoregions in the YRB.

Fig. 4. SPEI temporal trends and their spatial changes in the future. (a) The SPEI trend of the five maize ecoregions (Ⅰ, Ⅱ, Ⅲ, Ⅳ, Ⅴ) in the 2030s, 2050s and 2080s under SSP126, SSP370 and SSP585 scenarios in the YRB. The red dotted line in (a) indicates that SPEI trend is 0. In (a), the rose red symbols represent grids that had significant trends (p < 0.05) in SPEI, while the blue symbols represent grids with insignificant trends. (b) The spatial distribution of SPEI trend in the future periods (2030s, 2050s and 2080s) under three scenarios (SSP126, SSP370, and SSP585) in the subregions of the YRB.

Fig. 5. Comparison between observed and simulated yield at HWAM, ADAP, and MDAP for the representative stations in five regions. The solid black line is the 1:1 reference line and the two black dashed lines represent the upper and lower error bars.

Fig. 6. Comparisons and validations the model simulation values from five CMIP6 models and yield, ET estimates from the coupling observation-simulation correction based on the concept of yield gap against yield and ET measurement results of the 1985-2014 growing season at representative points in the five maize ecoregions in the YRB. Blue and orange solid circles represent data points after and before correction of the model simulation, respectively. The red dashed lines are 1:1 line and the boxes show the effect of the dense scatter points being enlarged.

Fig. 7. The relationship between SPEI and WUE. (a)The proportion of various dry and wet moisture gradients in each maize region in the YRB under the four scenarios of Historical, SSP126, SSP370, and SSP585, as well as the four periods of baseline (1985-2014), the future 2030s, 2050s and 2080s. Total represents the number of data points used in the analysis of moisture conditions for each maize region. (b)Responses of WUE to drought in the 2030s, 2050s and 2080s under SSP126, SSP370 and SSP585 scenarios for different maize ecoregions. All values are means ± standard deviation (sd). The error bars in all the subplots represent one sd and the shaded area indicates that the moisture status is in the normal category.Fig. 8. The dominance of yieldand ET over WUE. (a)The ratio of the partial correlation coefficient between WUE and yield to that between WUE and ET in the baseline (1985-2014) and future (2030s, 2050s, 2080s) periods for Historical, SSP126, SSP370 and SSP585 across five CMIP6 models. The light green square indicates that yield plays the dominant role, and the orange pink square indicates that ET plays the dominant role. (b) The relative importance (%) of yield (green) and ET (red) variables in influencing WUE across five subregions.

通讯作者：居辉，博士。中国农业科学院农业环境与可持续发展研究所研究员，博士生导师。中国气象学会理事，中国气象学会气候变化与低碳发展委员会委员，中国气候与气候变化标准化技术委员会委员，中国地理学会农业气象分会委员。长期从事气候变化对农业影响及其适应技术研究。担任政府间气候变化专门委员会（IPCC）第六次评估报告农业章节主要作者，IPCC第三次至第五次政府评审专家，中国《气候变化国家报告》农业章节首席作者。多次参加国际气候变化学术交流，曾赴10余个国家参加有关气候变化方面的学术研讨活动并与会做学术报告，其中韩国的国际气候变化研讨会为特邀的五名国际专家之一。受邀在联合国粮农组织（FAO）罗马总部、国际农发基金（IFAD）做“中国农业适应气候变化”相关特邀报告。主持国家重大基础科学计划、国家科技支撑计划、国家自然科学基金、国际合作等项目20余项，发表学术论文120余篇，主编专著2部，参编著作8部，获得省部级科技进步二等奖2项，三等奖1项，院级科技进步二等奖1项。

中国农业科学院居辉研究员主页: http://caas.teacher.360eol.com/teacherBasic/preview?teacherId=7130

居辉研究员百度百科: https://baike.baidu.com/item/%E5%B1%85%E8%BE%89/64281799?fr=ge_ala

第一作者：陈蔚，中国农业科学院农业环境与可持续发展研究所2023级博士研究生，在Journal of Cleaner Production、Carbon Balance and Management等环境科学领域主流期刊上发表了多篇SCI论文，主要研究方向为气候变化对作物水分利用效率的影响及其适应。

ResearchGate: https://www.researchgate.net/profile/Wei-Chen-633

ORCiD: https://orcid.org/0009-0001-4083-0813

干旱区气象生态拥有近1w活跃粉丝，是干旱区资源环境领域最大的自媒体平台，致力于帮助科研人员和团队宣传成果，促进学术交流。由于微信修改了推送规则，请大家将本公众号加为星标，一键三连，可以第一时间收到我们每日的推文！您的鼓励是我们最大的动力😊