东北三省春玉米旱灾动态风险评估

干旱演变趋势的不确定性决定了旱灾风险的动态变化特征，如何采用有效的方法探究风险随时间的变化是评估旱灾动态风险的关键。由于干旱发生发展缓慢，长时间尺度数值预报产品精度较低，用于干旱预报有一定的局限性。本研究选用天气发生器随机生成未来逐日气象数据的大量样本预测干旱演变趋势，驱动作物模型评估不同趋势下作物产量因旱损失及动态风险。选取东北三省为研究区，在2012—2017年田间试验数据的基础上，率定及验证APSIM作物模型，利用BCC/RCG-WG天气发生器随机生成未来气象数据，驱动APSIM作物模型模拟春玉米产量因旱损失，计算期望产量因旱损失率，以典型干旱年2000年为例，实时动态评估东北三省春玉米生育期内（5月1日—9月18日）旱灾动态风险。结果表明：（1）APSIM作物模型模拟春玉米播种到开花日数、生育期日数以及产量与试验观测结果决定系数R²均大于0.5,标准均方根误差NRMSE均在10%以下，表明APSIM模型在东北三省模拟春玉米生长效果较好；BCC/RCG-WG天气发生器生成100个气象要素样本与1961—2020年各气象数据年均值R²均在...

Dynamic drought risk assessment of spring maize in three provinces in Northeast China

Uncertainty of drought evolution and trends determines the dynamic characteristics of drought risk. How to use effective methods to characterize the changes of risk is the key to evaluate the dynamic drought risk. Due to the slow occurrence and development of drought, the accuracy of long-term numerical forecasting products is low, which has certain limitations in drought prediction. In this study, weather generator was selected to randomly generate a large number of samples of future daily weather data to describe the drought evolution which was used to drive crop models to evaluate crop yield loss due to drought and dynamic drought risk. Three provinces located in Northeast China were selected as the study area. The BCC/RCG-WG weather generator developed by the National Climate Center was employed to generate future meteorological data randomly. Driven by the meteorological data, the APSIM crop model was used to simulate spring maize yield loss due to drought which was calibrated and validated based on field test data during 2012-2017. Expected yield loss rate due to drought was calculated to evaluate the dynamic drought risk during the spring maize growth period in the three provinces of Northeast China in a typical drought year 2000 in real time. The results showed that R² between simulated and measured spring maize sowing to flowering days, growth period days, and yield were all greater than 0.5. The Normalized Root Mean Square Error (NRMSE) were all below 10%, indicating that APSIM model had a good effect in simulating spring maize growth in the three provinces of Northeast China. The R² of annual average meteorological elements generated by the BCC/RCG-WG with observed meteorological elements were above 0.9 during 1961-2020, and the average relative errors of monthly average rainfall, maximum temperature, minimum temperature, sunshine hours, relative humidity, and average wind speed were 9.1%, 9.9%, 14.5%, 6.1%, 14.7% and 21.6%, respectively. In 2000, the dynamic drought risk during the spring maize growth period the three provinces of Northeast China showed a trend of first increasing and then decreasing. The drought dynamic risk during the spring maize jointing to tasseling stage was the highest. The average expected yield loss rate in the three provinces of Northeast China was 0.23 on 10^th July. Spatially, there were high drought dynamic risk in northern Heilongjiang, western Liaoning, and midwest Jilin Province. There was a low drought risk in spring maize seedling stage and sufficient water in the later stage had a strong compensation effect. While there was a high drought risk when drought occurs at jointing stage and water compensation effect was weak when water was sufficient at later stage.