土壤湿度观测、模拟和估算研究

总结土壤湿度的观测手段和土壤湿度数据集建立的现状,详细阐述与土壤湿度模拟有关的方程离散化求解、物理和生化过程、陆面过程模式比较和陆面模式参数优化等方面的研究进展;综述估算土壤湿度廓线的数据同化方法,仔细比较集合卡曼滤波(EnKF)和四维变分(4-D Var)2类目前流行的同化算法,并对估算土壤湿度廓线的研究工作进行全面评估;最后,对土壤湿度观测、模拟和同化中需继续努力的方向进行了思索和展望。     

基于土壤水模型及站点资料的土壤湿度同化方法

基于非饱和土壤水模型和扩展卡尔曼滤波(Extended Kalman Filter)同化算法并结合陆面过程模型VIC发展了一个土壤湿度同化方案,并进行了理想试验及同化站点资料的同化试验。理想试验结果表明：扩展卡尔曼滤波方法能完整反演土壤湿度廓线,对土壤湿度的估计有较大改善;观测深度、观测层数和观测资料引入频率对同化结果有一定影响;加大观测频率,可以进一步改善同化效果。利用气象强迫驱动陆面模型VIC算出地表入渗条件而进行的同化站点资料的试验所得土壤湿度分布与观测资料基本吻合,反映了站点土壤湿度的月、季变化,表明该方案是合理的。     

评估两类模式对陆面状态的模拟和估算

针对夏季土壤变干过程,利用观测系统模拟试验,比较离线的陆面模式(LSM)和耦合大气边界层的陆面模式(SCM)对土壤温度、湿度和地表热通量等陆面状态的模拟,然后借助数据同化方法,评估2类模式对陆面状态的估算能力.结果显示:2类模式除对地表长波辐射和感热通量的模拟差别较大外,对其余量则较小;只同化表层土壤湿度观测时,LSM对土壤湿度和感热通量的估算好于SCM,对土壤温度的估算则相反,而对潜热通量估算的差距很小;同时同化表层土壤温度、湿度观测会使地表热通量的估算差距增大;最后对2类模式不同表现的可能原因进行分析讨论.上述数值模拟和同化结果:当用某一类模式的模拟结果或同化产品为另一类不同模式提供初边界条件时必须注意它们之间的差异,避免出现输入量引起的模式状态量间的动力不协调现象.     

评估集合卡曼滤波反演土壤湿度廓线的性能

集合卡曼滤波由于易于使用而被广泛地应用到陆面数据同化研究中,它是建立在模型为线性、误差为正态分布的假设上,而实际土壤湿度方程是高度非线性的,并且当土壤过干或过湿时会发生样本偏斜.为了全面评估它在同化表层土壤湿度观测来反演土壤湿度廓线的性能,特引入不需要上述假设的采样重要性重采样粒子滤波,比较非线性和偏斜性对同化算法的影响.结果显示:不管是小样本还是大样本,集合卡曼滤波都能快速、准确地逼近样本均值,而粒子滤波只有在大样本时才能缓慢地趋近;此外,集合卡曼滤波的粒子边缘概率密度及其偏度和峰度与粒子滤波完全不同,前者粒子虽不完全满足正态分布,但始终为单峰状态,而后者粒子随同化推进经历了单峰到双峰再到单峰的变化.     

一种新的耦合模糊控制局地化的同化方法

由于在数据同化过程中远距离的观测与同化状态之间存在着虚假相关,局地化方法受到广泛关注。同时,在集合数目较少的同化情况下,观测数据难以得到有效利用,使得同化效果欠佳。因此,提出了一种新的模糊控制局地化同化方法,通过模糊控制算法判断观测点与状态更新点之间的距离,构造观测位置模糊权重。利用非线性Lorenz-96模型,比较分析模糊控制局地化同化(FLETKF)算法与模糊控制同化(FETKF)方法、局地化分析同化(LETKF)算法和集合转换卡尔曼滤波(ETKF)算法在非线性强迫参数变化时的性能,同时探讨了4种算法在不同强度下的优劣。研究结果表明,新方法能够获得更有效的观测权重,避免了远距离观测与状态变量之间的虚假相关,减小由于观测数据难以得到有效利用而带来的误差,在不同观测误差协方差情况下,随着集合数的增加,4种算法中FLETKF能够保持较好的鲁棒性,在观测误差协方差较大时,FLETKF方法的均方根误差(RMSE)相对FETKF方法的RMSE值减小98.2%,提高了同化精度,但在同化所需时间上,由于模糊控制局地化同化方法在判断观测点与状态更新点之间的距离,构造观测位置等价权重需要较长的额外时间,因此,并行计算的性能需进一步研究。     

高光谱大气红外探测器AIRS资料质量控制研究进展

同化卫星资料能够得到模式较好的初始场,目前资料变分同化是基于误差服从高斯分布这一理论,因此在同化高光谱大气红外探测器(AIRS)资料前,必须进行资料质量控制。从通道选择、异常值剔除、偏差订正、云检测和数据稀疏化5个方面对AIRS资料质量控制研究现状进行分析与讨论。归纳总结了基于信息熵分步迭代法、主成分累计影响系数法和主成分—逐步回归法3种通道选择方法。经分析比较认为信息熵分步迭代法使用得较为广泛,但所选通道之间存在"弱相关";主成分—逐步回归法能够获得信息量较大的通道组合,由于算法的原因,执行过程较耗时。探讨了莱茵达法则和稳健性较强的双权重法对异常离群值剔除,得到双权重法效果较好。介绍了离线和在线偏差订正方法,包括静态、自适应、回归法、变分、基于辐射传输模式、基于卡尔曼滤波偏差订正法和偏差订正的动态更新技术。对比发现静态法时效性较好;变分法能够解决数据漂移等问题;基于模式和卡尔曼法虽效果较好,但较耗时不适合业务化使用;综合而言,偏差订正动态更新技术的效果和时效性都较好。分析了晴空视场点、晴空通道、云辐射订正和不同仪器云产品的匹配4种云检测方法。从数值业务时效性角度出发,晴空视场点和晴空通道云检测法较为可行,但经过晴空视场点云检测后同化的资料量比晴空通道法少,会造成在气象敏感区如高层通道资料的丢弃,在一定程度上会影响分析场的质量。进一步分析了跳点跳线、box法和主成分分析法在AIRS资料稀疏化中的初步应用研究。从同化时效性和可操作性出发,得出box法可行;主成分分析法算法复杂度较高,但具有一定的应用前景。在综述质量控制部分基础上,进一步给出了该领域未来的相关研究方向。     

不同滤波算法在土壤湿度同化中的应用

为研究不同滤波算法在土壤湿度同化中的有效性,以及土壤湿度模拟结果对模型参数的敏感性,结合简单生物圈模型SiB2,设置敏感性实验,探求土壤饱和水力传导度对土壤湿度模拟结果的影响;并在此基础上,采用集合卡尔曼滤波（EnKF）、无迹卡尔曼滤波（UKF）和无迹粒子滤波（UPF）开展土壤湿度实时同化实验。结果表明：土壤饱和水力传导度能显著影响土壤湿度模拟精度;利用EnKF、UKF、UPF同化站点观测数据,均能改善土壤湿度模拟结果;3种同化方法在不同土壤层的同化效果不同,在土壤表层,EnKF的有效性优于UKF和UPF,在根域层和土壤深层,3种滤波方法有效性在降雨前后相差较大。因此,针对性地选择同化方法,是提高土壤湿度模拟精度的有效手段。     

基于水文模型的蒸散发数据同化实验研究

流域蒸散发定量估算一直是水科学领域的研究前沿,水文模型和遥感反演是当前估算区域蒸散发的常用手段。研究通过数据同化,集成水文模型和遥感模型的优势,耦合遥感蒸散发到水文模型中以实现多源数据下的蒸散发数据同化。选择北京市沙河流域为研究区,分布式时变增益水文模型作为模型算子,基于集合卡尔曼滤波同化算法,利用双层遥感模型模拟的蒸散发同化水文模型,并基于地面通量站观测的日蒸散发进行验证。结果表明,同化结果与观测数据相比平均绝对百分比误差较同化前减少,精度进一步提升,且当遥感观测输入频繁时精度改善明显。研究证明基于水文模型的蒸散发数据同化系统,是一种可实现输出精度更高和时序连续的区域蒸散发的新型模式。该成果将进一步丰富创新蒸散发估算的学科内容,为准确理解区域水循环规律提供科学依据。     

陆面模式CLM4.5在青藏高原土壤冻融期的偏差特征及其原因

利用中国区域地面气象要素数据集制作的大气强迫场驱动通用陆面模式CLM4.5（Community Land Model version 4.5）对青藏高原区域进行离线模拟试验,模拟结果与D66、沱沱河（TTH）和玛曲（Maqu）3个站点的观测资料以及GLDAS（Global Land Data Assimilation System）-CLM2模拟结果进行了对比,并分析了陆面模式对冻融过程中土壤温度和湿度模拟的偏差及其可能原因。结果表明：CLM4.5对土壤温度模拟较好（平均RMSE≈3℃）,而GLDAS-CLM2计算的土壤温度偏高,偏差较大（平均RMSE>6℃）,且其偏差大于CLM4.5,尤其在冻融期;CLM4.5能较好地模拟出冻融过程中土壤湿度季节变化,但土壤湿度的模拟值与观测值存在一定偏差（平均RMSE≈0.1 mm³·mm^-3）,GLDAS-CLM2不能反映出土壤湿度在冻融过程中的变化特征。CLM4.5的模拟偏差主要来自大气强迫场,而GLDAS-CLM2的偏差除了大气强迫场的不确定性外,还来自于模式冻融参数化方案的不完善。大气强迫场中的气温和降水对土壤温度和湿度的影响在冻融期和非冻融期表现不同。在非冻融期,土壤温度的模拟主要受气温的影响（r>0.6）,气温偏差对土壤温度偏差的贡献率大于50%;土壤湿度的变化则主要受降水的影响,降水偏差对土壤湿度偏差的贡献率为20%~40%。在冻融期,受土壤水热相互作用的影响,气温和降水对土壤温度和湿度的作用效果减弱;土壤湿度的变化受气温影响显著,其贡献率为10%~20%。陆面模式中冻融参数方案的不完善是冻融过程中土壤温度和湿度偏差的重要来源之一。     

土壤水分同化系统的敏感性试验研究

利用1998年7月6日至8月9日青藏高原GAME-Tibet试验区MS3608站点的4cm、20cm和100cm的土壤水分观测数据同化SiB2模型输出的表层、根区和深层土壤水分,探讨了一个基于集合卡尔曼滤波和简单生物圈模型的单点土壤水分同化方案。分析和评价了集合大小、同化周期、模型误差、背景场误差以及观测误差对同化系统性能的影响。结果表明:①增加集合数目可以减小土壤水分同化系统的误差,但同时又降低了运行效率;②对于集合卡尔曼滤波,初始场的估计是否准确对同化系统性能影响不大;③模型误差和观测误差的准确估计可以提高土壤水分的估计精度;④利用数据同化的方法对土壤水分的估计有显著提高。     

冻结季沱沱河源多年冻土区活动层土壤水分含量分析

活动层含水量是表征多年冻土区气候、水文和生态过程的关键参数。长期以来,由于受多年冻土区活动层水分实测样点数量稀少的限制,各类基于遥感反演、模式模拟乃至数据融合和同化等手段生产的土壤水分空间数据均存在着较大的误差。2020年10—11月在青藏高原腹地（沱沱河源区）测定了 1 072组活动层土壤含水量数据并进行分析,探讨了该时段该区域活动层土壤水分的空间差异,并与全球陆面数据同化系统数据产品（GLDAS-Noah）和欧洲中期天气预报中心发布的第五代再分析资料（ERA5-Land）进行了对比分析。结果表明,在该区域平均厚度为2.72 m的活动层内,土壤质量含水量（总含水量）约为14.0%,活动层土壤含水量与植被发育情况存在正相关关系。除高寒沼泽草甸类型外,高寒草甸与高寒草原类型的活动层含水量随深度的增加呈现出先减小后增大的变化趋势。不同坡位类型的活动层含水量呈上坡位>下坡位>中坡位>平坡位,阳坡水分高于阴坡且两者活动层剖面水分变化相似。多年冻土区浅表层0~350 cm深度范围内的土壤含水量大于区内融区同深度的土壤含水量,两者土壤剖面水分分布均呈现出先增大后减小再增大的特征。该区域的GLDAS-Noah同化水分产品与实测数据对比的误差在10%以内,比ERA5-Land再分析土壤水分数据更为准确,但两种数据产品对土壤剖面上的水分垂直分布情况描述均与实测数据有较大差异。该研究结果可以为数据同化系统的模式冻融参数化方案优化及遥感水分产品研发提供科学依据。     

Validation of soil moisture simulation with a distributed hydrologic model (WetSpa)

The objective of this paper is to test the ability of a spatially distributed hydrologic model (WetSpa) to simulate soil moisture. The test is conducted in the Baron Fork river basin (Oklahoma, USA), where soil moisture has been recorded on a daily basis at the WEST site soil moisture monitoring station of the Oklahoma Mesonet. The model is calibrated by comparing hourly simulated and observed river flows at Eldon gaging station from October 1, 1995 to December 31, 1996. The model performance to predict hourly river flows is verified by comparing model estimates and observations from January 1, 1997 to December 31, 2000. Next, the model predictions of soil moisture are compared with observations of WEST site soil moisture monitoring station for the same validation period. The observations are well reproduced by the model with errors that are within range of the observation accuracy, although the predictions show somewhat more abrupt temporal fluctuations as the whole root zone is considered to react instantaneously to precipitation events. The good performance compared to other studies is especially attributed to the quality of the rainfall data provided by weather radars. It is concluded that the performance of the model to predict soils moisture is promising, although spatial patterns still have to be verified.     

海河流域不同下垫面土壤水分动态模拟研究

针对海河流域不同的下垫面类型,选取密云(果园林地)、大兴(城郊农田)、馆陶(平原农田)3个观测站,建立垂直方向上以含水率θ为因变量、含根系吸水项的非饱和土壤水分运动数值计算模型。该模型以一维Richards方程为基础(以下简称RE模型),采用实测的降水和蒸散数据作为模型的上边界条件,运用全隐式有限差分法,分别对不同生长期内的土壤水分进行数值模拟,得到时间序列的土壤水分廓线,并分别采用成熟软件HYDRUS-1D的模拟结果和各观测站实测土壤水分对RE模型进行交叉验证和直接验证。结果表明RE模型能够很好地模拟海河流域不同下垫面土壤水分动态变化过程,3个站模拟结果与实测土壤水分数据的均方根误差(RMSE)分别为0.03127,0.0359和0.0409 cm3/cm3。与HYDRUS-1D软件模拟结果(其与观测值的RMSE分别为0.03759,0.0647和0.0467 cm3/cm3)相比,RE模型模拟的土壤水分具有更高的精度,也显示出RE模型的可靠性。探讨3个站土壤水分的时空变异规律及其影响因子并以大兴站为例,通过优化RE模型参数,探讨犁底层对土壤水分模拟结果的影响,进一步改善RE模型的模拟精度。     

土的含水率C-AHFO法现场长期分布式监测可行性研究

本文采用了一种碳纤维加热光缆的光纤含水率测试方法（简称C-AHFO法）,解决了AHFO法监测土的含水率长期（16个月）稳定性问题,并开展了室内标定和现场原位土的含水率监测试验,确定了土中含水率与温度特征值之间的关系,建立了数值模型。室内率定试验结果表明:含水率与温度特征值之间的关系可用分段函数表述。当含水率小于0.08 m3·m-3时,可采用线性关系表述;高于0.08 m3·m-3时可采用对数函数表述。经过16个月的现场监测表明:C-AHFO法在此时间范围稳定,可对土的含水率进行分布式监测,测试误差不超过0.033 m3·m-3,可准确地反映天气和季节变化过程中表层土（深度小于20 cm）的湿度变化,监测结果与数值模拟计算结果具有一致性。该方法的研发和应用为土的大面积实时含水率分布式监测提供了一种新的手段。     

基于指数滤波法估算沟道根区土壤水分

为更加有效获取黄土高原沟道地形根区土壤水分信息,基于表层（0~20cm）土壤水分观测值,采用指数滤波法在样带和沟道尺度对根区（20~80cm）土壤水分进行估算。结果表明,指数滤波法能够较准确估算沟道根区土壤水分,但随着土层深度增加估算误差逐渐增大。当有根区土壤水分初始值时,估算精度明显提高,但主要集中在时间序列前期。指数滤波法关键参数特征时长T的最优值（T_opt）随土层深度增加而增大,呈现每增加20cm,T_opt值较20~40cm增加1倍的变化规律。对于根区某一土层,不同空间位置（样带）和尺度T_opt交换使用没有明显降低土壤水分估算精度,表明指数滤波法对T_opt的空间敏感性较弱。需注意黄土高原区域尺度土壤性质具有极强空间异质性,指数滤波法在更大尺度上对T_opt的敏感性需进一步深入研究。     

Statistical analyses and controls of root-zone soil moisture in a large gully of the Loess Plateau

Soil moisture variability and controls are little known in large gullies of the Loess Plateau which represent complex topography with steep slopes. This study analyzed spatial–temporal variability of soil moisture at the 0–20, 20–40, 40–60, and 60–80 cm depths in a large gully of the Loess Plateau based on root-zone soil moisture measurements for 3 years (2009–2011). The result showed that mean soil moisture, standard deviation (SD), and coefficient of variation, were highly dependent on depth; the highest mean value was observed at the 20–40 cm depth, while the lowest one was at the 0–20 cm depth. The SD increased with mean soil moisture for various depths as soil moisture was relatively wet; however, a transition that SD decreased with mean soil moisture occurred when soil moisture was relatively dry. Positive correlations exist between moisture contents over different depths, and that the relationships of the neighboring layers are relatively high with R ² from 0.70 to 0.76. Correlation analysis, principle component analysis, and stepwise multiple regression analysis showed that soil particle size distribution and topography (slope and elevation) were the main environmental factors controlling soil moisture variability in the large gully.     

Soil moisture potential and water content in the unsaturated zone within the arid Ejina Oasis in Northwest China

Three soil profiles were selected in the Ejina Oasis, northwest China, to determine water content profiles and evolution of soil moisture potentials in the unsaturated zone within the arid area. The total soil moisture potentials have been monitored for about 3 months in 2001 at different depths in the soil profiles. The occurrence and movement of water in the unsaturated zone was analyzed using the zero flux plane (ZFP) method. It is shown that convergent ZFPs and divergent ZFPs may occur at depths between 0.5 and 3.0 m, and that the depth of the ZFPs was controlled by the root zone of plants growing on the land surface. Profiles of the total soil moisture potentials were observed to be coincident with those of the water contents at the three experimental sites. The total soil moisture potential showed a slight increasing trend and the ZFPs tend to vanish from summer to winter as the water extraction by roots decreased. Evapotranspiration through vegetation has an important bearing on the water content and the total potential in the unsaturated zone.     

pH influence on sorption characteristics of heavy metal in the vadose zone

Sorption is an important process in the modelling and prediction of the movement of heavy metals in unsaturated clay barriers. This experimental study investigates the effect of pH changes in the acidic range on the sorption characteristics of heavy metals such as: lead, copper and zinc in an unsaturated soil. A series of one-dimensional coupled solute and moisture leaching column tests, using different heavy metal solutions, were conducted on an unsaturated illitic soil at varying pH values. Variations of volumetric water content (VWC) with distance were measured for different time durations, and concentrations of heavy metals in the liquid and solid phases were analysed. Partitioning coefficient profiles of contaminants along the soil column were determined for each individual layer in the soil.

Results from column leaching tests showed that the sorption characteristics of heavy metals are controlled by many factors which should be taken into consideration, i.e. the VWC, time of wetting, soil pH, and the influent heavy metal concentrations. Simplification of K_d as a constant and of the VWC as a linear function cannot be considered a good assumption and may lead to an improper evaluation of the sorption phenomena and also to serious errors in predicting contaminant transport through unsaturated soils.     

Analysis of soil boundary conditions of flash floods in a small basin in SW Hungary

Flash floods are one of the most significant natural hazards of today. Due to the complexity of flash flood triggering factors, to prevent or mitigate flood triggered losses, numeric model based flood forecasting models are capable tools to predict stream water levels. The main goal of the current research was to reproduce two flow peaks with the HEC-HMS rainfall-runoff model and test the model sensitivity for various input parameters. To obtain sufficient input data, we monitored soil depth, maximum infiltration rate, soil moisture content, rainfall, time of concentration and flow. To obtain input data, parameters were calculated, measured in the Sás Valley experimental watershed (SW Hungary) or optimized with the built in function of the HEC-HMS. Soil moisture was monitored in the 1.7 km2 pilot catchment over the period between September 2008 and September 2009. HEC-HMS had a good performance reproducing the two events, however simulated flow time series are highly influenced by the antecedent soil moisture, infiltration rate and canopy storage. Outflow modeled data were verified for two flood events (June 4, 2008 and July 9, 2009). The HEC-HMS was over-sensitive for input soil moisture and with increasing input rainfall and increasing outflow, larger simulation errors were observed.     

Research and Application of Cosmic-ray Fast Neutron Method to Measure Soil Moisture in the Field

It is very difficult to measure large region soil moisture. In recent years, measurement of surface soil moisture by the cosmic ray fast neutron probe has gradually attracted more attention. By this passive, non-invasive and intermediate scale measurement, soil moisture at a horizontal scale of around 300 m can be observed, which makes this method available to fill the gap between little scale of traditional point measurement and large scale of remote sensing in the measurement of soil moisture. In this paper, Cosmic-ray Soil Moisture Observing System (COSMOS) and Wireless Sensor Network (WSN) were used to observe field soil moisture in Zhangye Oasis. The results of the COSMOS soil moisture well reflected the variation trend of soil moisture at the field scale. There were some regular changes with the cosmic ray soil moisture during the irrigation period. Combined with irrigation data in cosmic-ray probe footprint, a bimodal change was showed in the cosmic-ray soil moisture figure during irrigation period. This was because that the order of irrigation of the three communities was different in cosmic-ray probe footprint. WSN data were used to validate this result and we found that root-mean-square error between cosmic-ray soil moisture and SoilNET average soil moisture was very large during irrigation periods because of the impact of cosmic-ray measurement theory. While root-mean-square error would decrease from 0.037 m³/m³ to 0.028 m³/m³ after we eliminated the data in irrigation period. So COSMOS can be used to measure field soil moisture under high heterogeneity condition and provide truly ground data for the validation of remote sensing