【作者】 刘云；

【导师】 宇振荣；

【作者基本信息】 中国农业大学 ， 生态学， 2003， 博士

【摘要】 本文以地面试验为基础，研究冬小麦主要生育期(拔节到灌浆期)冠层表面温度的变化特征及其与土壤含水量的相关关系，并建立了以冬小麦冠层表面温度为自变量估计土壤含水量的经验模型；在区域尺度(邯郸地区)上，筛选不同裂窗算法，并反演了邯郸地区冬小麦主要生育期冠层表面温度；以地面试验建立的冬小麦冠层表面温度与土壤含水量的经验模型为基础，将区域尺度(邯郸地区)反演的地表温度应用于作物旱情和长势监测。 研究结果表明：五个不同水分处理(淡水充分灌溉、淡水节水灌溉、咸水充分灌溉、咸水节水灌溉和不灌溉)冬小麦冠层表面温度、地温和气温表现为一致的日变化规律性：14:00左右各种温度达到最高，而早晨和傍晚温度较低，表现为明显的中间高两头低形式；各温度随生育阶段的发展也表现为一致规律性：随着生育阶段的发展气温逐渐增高；冬小麦冠层表面温度和地温也在逐渐增高，表现为拔节期和开花期是三个温度升高最快的发育阶段，但不同的水分处理冬小麦冠层表面温度具有一定的差异性(尤其14:00左右差异性最大)，表明14:00左右的冬小麦冠层表面温度能反映水分状况信息，同时也可以作为实施灌溉措施的参照指标。不同水分处理冬小麦冠气温差和地冠温差也具有相同的特征：14:00左右的差异幅度比较大，但不同水分处理各温差存在着差异：淡水充分灌溉、淡水节水灌溉、咸水充分灌溉和咸水节水灌溉的冬小麦冠气温差和地冠温差大部分小于零，表明这四种水分处理的冬小麦冠层表面温度小于气温，地温小于冠层表面温度，而不灌溉的冬小麦地冠温差和冠气温差基本上为正；不同高度处的冠气温差以冠层之上的冠气温差14:00的变化差异最大，认为冠层之上14:00的冠气温差表征作物水分状况最佳，从而为区域上应用冠气温差反映作物和土壤的水分状况提供了试验依据。 田间试验研究了不同水分处理冬小麦冠层表面温度与土壤含水量间的相关性表明：二者有很强的相关性，且二者的回归拟合方程基本上为二次或三次方程，但不同的水分处理表现为不同的差异性，且二者的拟合关系在时间(每天不同时刻)和空间(不同土壤层)上表现出差异性。进一步研究了不同水分处理三个不同高度冬小麦冠气温差与土壤含水量的相关性表明：二者也有很强的相关性，且二者的回归拟合方程基本上也为二次或三次方程，但不同高度处不同水分处理二者的相关性表现为不同的规律性，综合分析结果表明14:00各水分处理在冠层之上的冠气温差与土壤含水量具有相对较好的相关性，这为区域上利用地面试验建立二者间的回归拟合模型结合遥感资料估计土壤含水量奠定了试验基础。 村级试验尺度上的研究表明：分别在各生育阶段建立冬小麦冠层表面温度和冠气温差与土壤重量含水量间的回归拟合模型效果比较好，且用冠气温差所建立的二者的回归拟合关系要比用冠层表面温度建立二者的拟合关系的效果好。 通过地面试验比较9种裂窗算法，采用UL92裂窗法反演邯郸地区平原区的地表温度比较成功，尤其到冬小麦基本封垄的孕穗期开始，冬小麦的冠层表面温度基本上可以用地表温度来代替。采用UL92法反演的地表温度的估测结果与实测值比较接近，在冬小麦主要生育期(拔节到灌浆期)，反演的地表温度的平均误差为-0.27℃，标准误差为2.66℃。 将遥感方法反演的冬小麦冠层表面温度应用于作物的旱情和长势监测表明：区域尺度上，采 回用地表温度反映作物的水分状况是可行的。以地面试验所建立的冬小麦冠层表面温度与土壤含水量间的关系模型为基础，并将UL92裂窗法反演的邯郸地区的地表温度代人该模型，估计邯郸地区土壤含水量，是本研究的新思路，用供水植被指数模型验证，表明采用该方法基本能反映作物的旱情；采用供水植被指数（NDVTh）能详细反映作物的水分状况，且能表征作物的长势情况，但不如本文所提出的方法，它不能计算不同土层土壤含水量来反映作物的水分状况；采用NDVI与Ts的特征空间的思想表明该空间能及时反馈作物的长势情况和作物水分状况，分别以裸地与麦地两种土地覆盖类型为研究对象，采用Tk－yDVI矢量特征空间生态学内涵思想，证明该空间能反映冬小麦在主要生育期（拔节到灌浆期）的长势情况。更多还原

【Abstract】 Based on field experiments, the change characteristic of CST(canopy surface temperature) and the correlation among CST and SWC(soil water content) in major growth stages(from jointing stage to grain-filling stage) of winter wheat were studied, and the empirical models on estimating SWC with CST as independent variable were established; by comparison among 9 different split window methods using NOAA/AVHRR data in regional scale(Handan district), the optimal retrieval method UL92 for CST in Handan was derived, and LST(land surface temperature) retrieved using UL92 in regional scale was applied in monitoring drought and crop growth on the basis of empirical models between CST and SWC.The research results showed, the same daily change characteristic among CST, air temperature and SST (soil surface temperature) of winter wheat under 5 different water treatments(full-fresh-water-irrigation; fresh-saving-water-irrigation; full-salty-water-irrigation; salty saving-water-irrigation; no-irrigation): the highest temperature at 14:00 and the relative lower temperature in the morning and evening. During major growth stages, the similar change dynamic characteristic exists among the above 3 temperatures: with air temperature increasing gradually, CST and SST also increase gradually, and the 3 temperatures increase at the highest rate in jointing stage and anthesis stage. But the differences in CST change are distinct for different water treatments and indicate that CST at 14:00 can reflect crop water condition as a indicator to saving water irrigation capacity. The similar change trend also exists in temperature difference between canopy-air(Tca) and land-canopy(Tlc) of winter wheat in different water treatments: the highest temperature difference at 14:00, but the values of temperature difference under different water treatments differ from each other: Tca<0, Tlc<0 mostly with the order of T(LST)<T(canopy)<T(air) under the treatments of full-resh-water-irrigation, fresh -saving-water-irrigation, full-salty-water-irrigation, salty-saving-water-irrigation; Tca>0, Tlc>0 basically under no irrigation; the relative bigger fluctuation of Tea lies in the height of 50 cm above canopy at 14:00 between canopy-air temperature differences in three heights. From the above analysis, the Tea above canopy at 14:00 is the best indicator to crop water condition, which presents the support for application of canopy-air temperature differences in reflecting water condition of crop and soil in regional scale.The correlation between CST and SWC under different water treatments in field experiments was analyzed. The regression relationship results gave significant correlation of quadratic or cubic equations between CST and SWC and were different in different time and different soil layers, especially among different water treatments. The correlation between Tea in three heights and SWC under different water treatments was also studied and significant with quadratic or cubic regression equations basically, but there were the differences of correlation between Tea in different heights and SWC among different treatments. By comparison, Tea above canopy at 14:00 gave relative better correlation in different water treatments, which provided the experimental support for estimating SWC in regional scale usingregression models between Tea above canopy by field experiments and SWC combing with remote sensing data.The experimental results in the village scale showed: it was appropriate to establish respectively the regression models between CST, Tea and SWC according to growth stages of winter wheat, the regression results was better between Tea and SWC than between CST and SWC.By validating 9 split window methods by field experiments, UL92 can retrieve LST in Handan district more successfully than other retrieval methods, and especially from the beginning of booting stage, the retrieved LST was approximate to CST. During the major growth periods from jointing to g-ain-filling stage, the mean error -0.27℃ and standard error 2.66℃ were deriv更多还原