水稻低温冷害是东北地区主要的农业气象灾害之一.与耗时、费力的传统灾情监测和调查方式相比,遥感技术能以宏观视角高效地监测灾害发生的范围及程度.为顺应立体化农业气象服务的发展趋势,探索利用Terra\Aqua MODIS反射率和地表温度数据进行水稻延迟型冷害动态监测的可能性,本文在遥感估算全天候平均气温分布、遥感识别水稻种植区及其关键生育期的基础上,构建了基于相对累积生长度日距平rAGDDa的水稻延迟型冷害指标,并开展了东北地区2000年—2012年水稻延迟型冷害动态监测的应用研究.结果表明:(1)MODIS估算的累积生长度日(MODIS_AGDD)与其对应台站估算值之间存在高度相关,除了5月上旬及9月下旬外,两者差值的多年平均值随时间变化幅度基本保持在55℃·d左右;(2)rAGDDa与其对应台站估算值的相关性比累积生长度日距平(AGDDa)的更高,采用rAGDDa指标监测东北地区水稻生长季热量条件的年际差异更有效;(3)基于MODIS_rAGDDa指标的监测结果与用气象行业标准指标(ΔT_5-9)监测得到的延迟型冷害分布在众多灾害年均有较好的空间一致性,通过对比水稻减产率分布图,认为以MODIS_rAGDDa≤-5%判断水稻延迟型冷害的发生具有一定可行性;(4)分阶段统计的rAGDDa指标能反映出水稻各生育阶段的低温累积效应和高温补偿效应,可用于水稻延迟型冷害动态监测业务服务.

Rice chilling damage remains one of the major agricultural meteorological disasters in northeast China. Remote sensing technology can easily monitor the disaster on a relatively large scale compared with traditional methods; this is the development trend for the agro-meteorological service system. Although several studies have focused on freeze injury monitoring using remotely sensed data, few applications exist for chilling damage.
In this study, Terra and Aqua Moderate Resolution Imaging Spectroradiometer (MODIS) land surface temperature and reflectance data were used to produce mean air temperature time series by employing multiple regression models and to identify the planting areas and key development stages of single rice. A novel index based on relative accumulated growing degree day anomaly (rAGDDa) was accordingly proposed to dynamically monitor the distribution and degree of rice delayed-type chilling damage in northeast China from 2000 to 2012.
Results are as below. (1) High correlation is obtained over the 0.01 significant level of accumulated growing degree day derived from MODIS data (MODIS_AGDD) and meteorological data at each 8-day period, with a stable multi-year average difference of approximately 55 ℃·day throughout the entire rice growth season, except for early May and late September. (2) Compared with MODIS_AGDD, MODIS_rAGDDa yields a higher R value with a meteorological estimation over the 0.05 significant level in each year and can more effectively eliminate the threshold difference in the temperature index of chilling damage caused by the discrepancy in geographical and heat conditions. (3) Dynamic monitoring based on MODIS_rAGDDa is illustrated for three rice growth stages, such as early (from transplanting to heading), late (from heading to maturation), and whole (from transplanting to maturation) growth stages. Results achieved by MODIS_rAGDDa index during the whole growth stage of rice are broadly consistent in spatial distribution with those produced by meteorological standard index (summation of monthly mean air temperature from May to September anomalies) in several disaster years, especially in 2009 when a wide-range disaster occurred. Meanwhile, the spatial distribution of average reduction rate in rice yield indicates that MODIS_rAGDDa less than -5% can be taken as the indicator for monitoring rice delayed-type chilling damage in northeast China.
MODIS_rAGDDa estimation for different rice growth stages can reasonably show the low-temperature accumulation and high-temperature compensation effects. It is considered a suitable index applied to the business service of dynamical monitoring of rice delayed-type chilling damage in northeast China.