地表温度的多源遥感数据反演算法对比分析

 多源遥感数据的综合应用是提高地表温度反演精度的有效途径.MODIS数据和Landsat TM数据在我国同一地区获取的时间相差不大,可以获取近似同步的MODIS数据和TM数据.本文将基于MODIS数据反演的大气参数应用于TM影像的地表温度反演,分别对单窗口算法和普适性单通道算法进行了实验研究,应用气象站实测的地表温度数据对反演结果进行了检验,并对比分析了不同土地覆盖条件下两种算法的精度差异.结果表明:两种算法反演精度均较高,单窗口算法反演精度为0.76K,普适性单通道算法反演精度为1.23K;在不同的土地覆盖条件下,两种算法表现出明显的差异性,水体区反演结果差异最小,均值差异仅为0.02K,植被区差异最大,均值差异为0.62K.     

基于地理国情监测框架下长沙市望城区地理空间分析

利用望城区2013年LANDSAT8卫星遥感数据,反演了望城区地表温度;并结合望城区地理国情普查成果数据进行动态分析,结果表明望城区地理国情普查地表覆盖数据与利用LANDSAT8反演的地表温度之间有很好的相关性。该研究成果对于动态监测城镇土地利用变化,优化土地资源配置具有一定的参考意义。     

城市地表覆盖组分及植被多样性与热环境关联性分析——以贵阳市为例

本文以贵阳市为研究对象,研究了地表覆盖组分及植被多样性对地表温度的影响。首先基于Landsat8 OLI多时相影像数据在GEE平台上实现了研究区域地表覆盖精细分类;然后结合不同季节8天合成的MODIS温度产品数据,利用时空统计分析、相关分析等方法分析了研究区不同地表覆盖类型地表温度时空分布特征,地表温度与不同地表覆盖组分、地表覆盖多样性和植被覆盖多样性的相关性。结果表明：贵阳市建成区主要分布有常绿阔叶林、常绿针叶林等植被,常绿阔叶林在不同季节对地表温度的降温效应明显,而不透水面对地表温度具有明显的增温效应,尤其以夏季最为显著;地表覆盖多样性与地表温度之间具有较强相关性,其中植被覆盖多样性较植被覆盖率对地表温度的影响更为显著,而不透水面的增加会明显降低植被多样性的影响。因此,要发挥城市绿地对城市温度和热岛效应的调节作用,建议可以适当增加常绿阔叶林的绿化面积,同时在空间上要提升植被多样性水平,能够较大程度改善城市热环境。     

基于随机森林算法的近地表气温遥感反演研究

近地表气温是城市热环境的重要表征,是改变和影响城区气候的重要因素。为获得空间上连续的近地表气温,本文以北京市为研究区,利用Landsat5/TM数据计算分别得到地表温度、归一化植被指数、改进的归一化差异水体指数、地表反照率、不透水面盖度,并结合气象站点气温和高程作为输入参数建立随机森林模型反演近地表气温。结果表明,随机森林反演的近地表气温平均绝对误差（MAE）为0.80 ℃,均方根误差（RMSE）为1.06 ℃,与传统多元线性气温回归方法相比,平均绝对误差（MAE）和均方根误差（RMSE）分别提高0.06 ℃和0.09 ℃。研究表明,利用随机森林模型反演近地表气温是可行的,并且具有一定的优越性。此外,对随机森林模型的输入参数进行重要性分析,地表温度对气温反演模型的影响最大,其次为高程。     

基于马尔可夫链和最大后验准则的模拟静止气象卫星数据地表组分温度反演

地表组分温度比像元混合温度具有更强的物理意义和实用价值,是定量遥感反演的一个重要研究方向。本文以马尔可夫链和最大后验准则地表温度尺度转换方法,结合静止气象卫星数据高时间分辨率的特点,通过模拟静止气象卫星数据地表组分温度反演进行分析和验证。在研究过程中,地面被简化为由植被和土壤两组分组成,同时假设邻近像元的植被和土壤组分温度相同。鉴此,本文通过模拟构建20×20像元大小的静止气象卫星混合像元图像,并对各像元各时刻温度添加均值为0标准差为2K的随机误差,最终应用所提算法估算各像元各时刻的植被和土壤组分温度大小。精度分析结果表明,该算法能够较为精确地反演植被和土壤组分温度,且误差基本控制在2K以内。此外,本文还进一步讨论了算法的适用性及其对混合像元温度误差、植被覆盖度误差,以及邻近像元植被覆盖度变化范围的敏感度。分析结果再次证明,该方法对混合像元温度误差和植被覆盖度误差都具有较低的敏感性,在最大温度误差条件(均值为1.8K,标准差为5K)和最大植被覆盖度误差(均值为0.18,标准差为0.2)的条件下,各组分温度的估算精度分别能控制在3K和2K以内,满足精度要求。但是,由于组分温度初值的确定方法,对所计算窗口内植被覆盖度变化范围有较强的敏感性,反演结果与植被覆盖度变化范围相关,要求窗口内植被覆盖度变化范围足够大才能满足初值估算的精度要求。     

基于Landsat TM影像不同地表温度反演算法的比较

地表温度是反映地表环境的一个重要参数,精确获取地表温度的方法对研究城市热岛效应、进行生态环境监测评价是必不可少的。利用遥感手段进行地表温度的反演是一种较新颖的方法,相对成熟的反演算法主要有单通道算法、辐射传输方程法及单窗算法。本文以山东省威海市Landsat TM数据为例,分别利用辐射传输方程法和单窗算法进行地表温度反演,通过遥感目视解译的方法提取研究区各类典型地物对应的地表温度,进行统计分析,结果表明,两者反演得到的地表温度具有一致的变化趋势,其中单窗算法中不同典型地物的地表温度整体标准差较小,如建设用地为0.530,起伏波动小,算法精度略优于辐射传输方程法。     

通用劈窗算法的NOAA-18(N)AVHRR/3数据地表温度遥感反演与验证

本文以NOAA-18(N)AVHRR/3数据,运用通用劈窗技术获得地表温度。首先,利用MODTRAN 4模拟不同地表和大气状况下热红外通道(Ch4,10.3~11.3μm和Ch5,11.5~12.5μm)的星上亮温,并建立模拟数据库。其次,按照地表温度、大气可降水汽含量、地表比辐射率和观测天顶角,对模拟数据库分组,确定出各分组的通用劈窗算法系数。然后,将构建的地表温度反演模型应用到NOAA-18(N)AVHRR/3数据,模型所需的地表比辐射率由NDVI阈值法确定,大气可降水汽含量是利用Li等(2003)提出的一种劈窗的协方差与方差比的方法来估算。反演结果表明:在观测天顶角小于30°或者大气可降水汽含量小于3.5 g/cm2时,地表温度反演的均方根误差小于1.0K;在观测天顶角小于45°并且大气可降水汽含量小于5.5g/cm2情况下,均方根误差小于1.5K。最后,利用美国通量站的实测数据对地表温度反演结果进行了验证,结果表明均方根误差小于1.8K。     

基于地表通量特征的城市不透水表面定量热红外遥感反演

城市不透水地表格局通过改变城市下垫面结构,引起地表反照率、比辐射率、地表粗糙度的变化,从而对地表辐射和能量平衡产生直接影响。不透水地表能增强地表显热通量,导致地表波文比升高,因此地表波文比的空间差异可推算城市人工不透水表面的分布。本研究选择北京市为实验区,应用Landsat TM卫星热红外遥感数据,采用PCACA模型及理论定位算法,对城市地表波文比进行反演,进而计算遥感地表波文比空间分布数据与城市不透水表面比例数据之间的相关关系,构建回归方程,实现北京市城区与近郊区人工不透水表面百分比分布的定量估算,最后以高分辨率遥感数据获取的城市人工不透水表面比例数据进行结果验证。结果表明,采用PCACA模型定量反演城市地表波文比数据,利用地表波文比数据与不透水表面比例数据之间的相关关系可实现城市人工不透水表面百分比数据的定量估算;波文比值不仅可在遥感像元水平定性判定不透水像元,还可对混合像元中的不透水比例进行较高精度的定量反演,其相关系数R²值为0.731。此方法有效地揭示了城市不透水下垫面对地表热通量影响的机制以及空间定量关系。     

遥感图像信息容量约束区间的选择与空间分异性

遥感图像信息容量是一种能量化表征地表复杂度的评价指标。计算时考虑了像元点所处的整个局部区域特征,其大小与图像灰度层次密切相关,灰度层次越丰富,信息容量的值越大。信息容量模型构建的核心问题是约束区间的选择和参数的确定,合理适宜的参数设置是保证信息容量特性的关键性技术。选取陕西省不同地貌类型区56个样区,以2007年ETM+和2008年SPOT5遥感图像为实验数据。采用了2种不同的约束区间的计算方法,即比较分析和数理统计的方法,分析了遥感图像信息容量约束区间的选择方法和空间分异规律。结果表明,信息容量在一定程度上能有效反映地表空间形态结构的复杂度,信息容量和分形维数、信息熵之间有较好的线性相关性,随着信息容量的增大,样区的分形维数、信息熵也在增大。信息容量的空间分异和陕北黄土高原的黄土地貌形态在空间上的变异是相关的,与陕西关中平原区的地表地物覆盖类型也是相关的,可作为地表形态结构复杂度定量评价指标之一。     

基于背景知识的全球长时间序列反照率反演

全球范围时空连续的长时间序列地表反照率,对气候模拟与陆面过程研究具有重要意义。针对现有地表反照率产品普遍存在大量的数据缺失、有效反演比例低和时间序列短的问题,本文以多年MODIS和AVHRR数据,通过构建背景知识库进行高时间分辨率的AVHRR和MODIS数据的BRDF参数反演,实现MODIS与AVHRR数据在像元尺度上的定量融合,生成了全球时空连续长时间序列的地表反照率产品。首先,通过假设不同年份同一时期的地表状态不变,利用多年同一时期的MODIS和AVHRR观测数据构造多角度方向反射率,基于BRDF模型反演得到窄波段反照率;然后,通过宽波-窄波转换,得到MODIS的宽波段反照率;最后,结合AVHRR长时间序列优势及MODIS数据多光谱的特点,对二者进行定量融合,生成具有高度一致性长时间序列地表反照率产品。验证结果表明,本文地表反照率产品在地表异质性较小时与SURFRAD地面实测反照率之间具有非常好的一致性,在无积雪覆盖时与MODIS反照率产品之间吻合良好。本文的地表反照率产品无时空缺失,且时间覆盖率得到了极大的提高,能支持气候模式模拟与陆面过程模型进行近30 a来的地气系统模拟研究。     

Flood and waterlogging monitoring over Huaihe River Basin by AMSR-E data analysis

In this paper, we investigated the potential of Advanced Microwave Scanning Radiometer for Earth Observation System （AMSR-E） for flood monitoring. The proposed approach was based on the polarized ratio index （PRI）, which was computed by using AMSR-E data at 37GHz, vertically and horizontally polarized brightness temperature values and the water surface fraction （WSF） got by using the PRI at 37GHz. Moderate Resolution Imaging Spectroradiometer （MODIS） data were used to validate the WSF values. The analysis of flood and waterlogging using the WSF map on July 6, 2003 indicates that the use of WSF for flood and waterlogging disaster assessment is feasible. Utilizing the correlation of WSF derived from AMSR-E and water area derived from MODIS, the water area of the Huaihe River Basin were computed by only using AMSR-E data in the summer of 2003, which overcame the influence of cloud on water estimation using MODIS data during flood.     

基于MODIS和AMSR-E反演京津冀地区大气水汽含量及比较

为了长时间、大范围获取水汽数值,利用2005~2008年光学遥感的MODIS近红外、红外水汽产品,以及微波遥感AMSR-E数据,2种方法反演水汽。微波AMSR-E亮温数据采用Merritt N.Deeter(2007)亮温极化差方法,选取18.7GHz和23.8GHz 2个波段,得到AMSR-E升轨、降轨大气水汽数值。以京津冀地区为研究区域,通过地统计相关性分析、时间序列分析、年际间变化分析,可知2种方法4种资料反演的大气水汽数值的R2都达到0.95,时间分布符合中国雨带移动规律,空间分布不均。MODIS数据反演值比AMSR-E值要低,得到2种方法反演水汽的各自优缺点。     

Retrieval of Land-surface Temperature from AMSR2 Data Using a Deep Dynamic Learning Neural Network

It is more difficult to retrieve land surface temperature(LST) from passive microwave remote sensing data than from thermal remote sensing data, because the emissivities in the passive microwave band can change more easily than those in the thermal infrared band. Thus, it is very difficult to build a stable relationship. Passive microwave band emissivities are greatly influenced by the soil moisture, which varies with time. This makes it difficult to develop a general physical algorithm. This paper proposes a method to utilize multiple-satellite, sensors and resolution coupled with a deep dynamic learning neural network to retrieve the land surface temperature from images acquired by the Advanced Microwave Scanning Radiometer 2(AMSR2), a sensor that is similar to the Advanced Microwave Scanning Radiometer Earth Observing System(AMSR-E). The AMSR-E and MODIS sensors are located aboard the Aqua satellite. The MODIS LST product is used as the ground truth data to overcome the difficulties in obtaining large scale land surface temperature data. The mean and standard deviation of the retrieval error are approximately 1.4° and 1.9° when five frequencies(ten channels, 10.7, 18.7, 23.8, 36.5, 89 V/H GHz) are used. This method can effectively eliminate the influences of the soil moisture, roughness, atmosphere and various other factors. An analysis of the application of this method to the retrieval of land surface temperature from AMSR2 data indicates that the method is feasible. The accuracy is approximately 1.8° through a comparison between the retrieval results with ground measurement data from meteorological stations.     

光学与微波遥感的新疆积雪覆盖变化分析

利用2002-2013年冬季的MODIS光学遥感数据,以及AMSR-E、AMSR2与MWRI被动微波遥感数据,建立了新疆地区冬季每日积雪分布遥感反演模型。首先,将Terra与Aqua双星MODIS的积雪产品融合,初步去云并最大化积雪信息;然后,利用AMSR-E/AMSR2和MWRI被动微波数据进行每日雪盖提取;最后,利用被动微波遥感数据反演得到的每日雪盖结果对双星融合后依然有云的像元进行替换,得到每日积雪分布情况。据此模型提取了11年间冬季的积雪天数信息,结合气象台站观测数据,分析了新疆冬季积雪的年内和年际变化规律。结果表明,新疆地区积雪主要分布在北部新疆,积雪天数与地形关系密切,山区积雪天数较多,盆地及城市区积雪天数较少;积雪天数年内变化是从11月到次年1月随温度降低逐渐增加,从1月到3月积雪天数则逐渐减少。新疆地区积雪天数在这11年中存在一定的波动,积雪天数与该年的平均气温,以及月低于0℃的天数存在显著相关性,与降雪量关系不明显。新疆地区近年来积雪天数重心有向西向南移动的趋势,这可能与全球气候变暖导致多年积雪融化有关。     

Impact of land cover change on land surface temperature: A case study of Spiti Valley

Land surface temperature(LST) is the skin temperature of the earth surface. LST depends on the amount of sunlight received by any geographical area. Apart from sun light, LST is also affected by the land cover, which leads to change in land surface temperature. Impact of land cover change(LCC) on LST has been assessed using Landsat TM5, Landsat 8 TIRS/OLI and Digital Elevation Model(ASTER) for Spiti Valley, Himachal Pradesh, India. In the present study, Spiti valley was divided into three altitudinal zones to check the pattern of changing land cover along different altitudes and LST was calculated for all the four land cover categories extracted from remote sensing data for the years of 1990 and 2015. Matrix table was used as a technique to evaluate the land cover change between two different years. Matrix table shows that as a whole, about 2,151,647 ha(30%) area of Spiti valley experienced change in land cover in the last 25 years. The result also shows vegetation and water bodies increased by 107,560.2 ha(605.87%) and 45 ha(0.98%), respectively. Snow cover and barren land decreased by 19,016.5 ha(23.92%) and 88,589(14.14%), during the study period. A significant increase has been noticed in vegetation amongst all land cover types. Minimum, maximum and mean LST for three altitudinal zones have been calculated. The mean LST recorded was 11℃ in 1990 but it rose by 2℃ and reached to 13℃ in 2015. Changes in LST were obtained for each land cover categories. The mean temperature of different land cover types was calculated by averaging value of all pixels of a given land cover types. The mean LST of vegetation, barren land, snow cover and water body increased by 6℃, 9℃, 1℃, and 7℃, respectively. Further, relationships between LST, Normalized Difference Snow Index(NDSI), and Normalised Difference Vegetation Index(NDVI) were established using Linear Regression.     

Comparative analysis of microwave brightness temperature data in Northeast China using AMSR-E and MWRI products

With such significant advantages as all-day observation, penetrability and all-weather coverage, passive microwave remote sensing technique has been widely applied in the research of global environmental change. As the satellite-based passive microwave remote sensor, the Advanced Microwave Scanning Radiometer-Earth Observing System (AMSR-E) loaded on NASA’s (National Aeronautics and Space Administration of USA) Aqua satellite has been popularly used in the field of microwave observation. The Microwave Radiation Imager (MWRI) loaded on the Chinese FengYun-3A (FY-3A) satellite is an AMSR-E-like conical scanning microwave sensor, but there are few reports about MWRI data. This paper firstly proposed an optimal spatial position matching algorithm from rough to exact for the position matching between AMSR-E and MWRI data, then taking Northeast China as an example, comparatively analyzed the microwave brightness temperature data derived from AMSR-E and MWRI. The results show that when the antenna footprints of the two sensors are filled with either full water, or full land, or mixed land and water with approximate proportion, the errors of brightness temperature between AMSR-E and MWRI are usually in the range from −10 K to +10 K. In general, the residual values of brightness temperature between the two microwave sensors with the same spatial resolution are in the range of ±3 K. Because the spatial resolution of AMSR-E is three times as high as that of MWRI, the results indicate that the quality of MWRI data is better. The research can provide useful information for the MWRI data application and microwave unmixing method in the future.     

Spatio-temporal variation of land surface temperature and temperature lapse rate over mountainous Kashmir Himalaya

In this study, Land Surface Temperature(LST) and its lapse rate over the mountainous Kashmir Himalaya was estimated using MODIS data and correlated with the observed in-situ air temperature(Tair) data. Comparison between the MODIS LST and Tair showed a close agreement with the maximum error of the estimate ±1°C and the correlation coefficient 0.90. Analysis of the LST data from 2002-2012 showed an increasing trend at all the selected locations except at a site located in the southeastern part of Kashmir valley. Using the GTOPO30 DEM, MODIS LST data was used to estimate the actual temperature lapse rate(ATLR) along various transects across Kashmir Himalaya, which showed significant variations in space and time ranging from 0.3°C to 1.2°C per 100 m altitude change. This observation is at variance with the standard temperature lapse rate(STLR) of 0.65°C used universally in most of the hydrological and other land surface models. Snowmelt Runoff Model(SRM) was used to determine the efficacy of using the ATLR for simulating the stream flows in one of the glaciated and snow-covered watersheds in Kashmir. The use of ATLR in the SRM model improved the R2 between the observed and predicted streamflows from 0.92 to 0.97.It is hoped that the operational use of satellite-derived LST and ATLR shall improve the understanding and quantification of various processes related to climate, hydrology and ecosystem in the mountainous and data-scarce Himalaya where the use of temperature and ATLR are critical parameters for understanding various land surface and climate processes.     

基于伪纯像元的精度评价策略及其应用

土地覆被是地球科学研究中的重要参量,评价土地覆被数据的制图精度是保障数据合理使用的前提。本文提出了一种基于伪纯像元的精度评价策略（伪纯像元策略）,即当低空间分辨率栅格窗口内对应的高空间分辨率数据中优势类别（面积最大的地类）的占比高于伪纯像元纯度阈值（代表像元纯度,取值范围：35%~100%,步长为5%）时,以此栅格窗口为基准生成土地覆被类型为优势类别的伪纯像元用于精度评价。以澜沧江-湄公河（澜湄）流域为试验区,选择GlobeLand30为参考数据,并基于混淆矩阵精度评价方法对比分析了伪纯像元策略与重采样法（最近邻法和众数法）在CCI-LC（300 m）和MCD12Q1（500 m） 2套全球土地覆被数据精度评价中的差异。结果表明：① 伪纯像元策略在35%~100%纯度阈值下对CCI-LC和MCD12Q1在澜湄流域评价的精度分别为72.76%~55.26%和71.44%~45.41%,比重采样法评价的单一精度（众数法：71.21%和70.54%、最近邻法：71.48和69.87%）能更好地反映像元纯度对土地覆被数据精度的影响;② CCI-LC的总体精度高于MCD12Q1,且2套数据的精度差随纯度阈值的增大而增加,CCI-LC和MCD12Q1在35%、100%纯度阈值下的精度差分别为1.32%和9.85%;③ 2套数据中耕地、有林地、草地和水体的分类精度均相对较高,而灌木林地（精度接近0）和裸地的分类精度均较低;④ 2套数据与GlobeLand30的空间不一致区域多出现在土地覆被类型高度异质化的混合像元区域,且随纯度阈值的增大,评价样本栅格更趋均质,混合像元对评价精度的影响也会递减。伪纯像元精度评价策略适用于跨空间分辨率土地覆被数据的精度对比,为评价全球土地覆被产品在区域尺度的适用性及适用范围提供了新的检验策略。