针对MODIS 数据的地表温度非线性迭代反演方法

地表温度是气象、水文、生态等研究领域中的一个重要参数。构建了MODIS31/ 32 波段的热辐射传输方程, 讨论了方程的数值迭代解法, 提出了针对MODIS 数据地表温度的非线性迭代反演方法, 并介绍了大气透过率和地表比辐射率这两个中间参数的估计方法。误差及敏感性分析表明,提出的方法对大气透过率和地表比辐射率都不敏感, 反演精度优于传统的线性分裂窗算法。     

陆地资源卫星数据地表温度反演

针对目前陆地资源卫星(Landsat-8)地表温度反演过程中,地表比辐射率估计和敏感度分析中存在的不足,对这两方面进行改进,提出了一种基于Landsat-8数据的地表温度反演算法。该文主要从劈窗算法的推导、参数的估计、敏感度分析等方面进行研究。对于大气透过率的计算,首先用与其有相邻过境时间的MODIS数据反演大气水汽含量,然后通过中分辨率的大气传输模型(Moderate Resolution Atmospheric Transmission,MODTRAN)模拟大气水汽含量与透过率的关系,最后得到大气透过率。对于发射率的计算,通过分类和ASTER提供的光谱库获得。将大气辐射传输方程模拟的地表温度与此劈窗算法反演的地表温度做比较,结果表明平均精度达到0.82K。最后研究了大气水汽含量对地表温度的影响。结果显示,当大气水汽含量误差为0.1g/cm2,其对温度反演精度的影响最大不超过0.3K;当大气水汽含量的反演误差较大的时候,其对温度反演精度的影响较大。     

MODIS数据反演地表温度的参数敏感性分析

在利用MODIS卫星遥感数据进行地表温度反演过程中，有两个基本参数需要确定，即地表比辐射率和大气透过率，尽管采用了比较合理的参数估计方法，但仍会有一些不可避免的因素导致误差的产生。为了进一步研究可能的参数误差对地表温度反演精度的影响，我们对该算法的两个参数进行敏感性分析。结果表明，当31、32两个波段的参数估计都有中等误差时，可能的地表温度误差对大气透过率和地表比辐射率都不敏感，所引起的地表温度误差大约为0．6～0．8℃，算法能够得到较高精度的地表温度反演结果。     

热红外遥感地表温度反演研究现状与发展趋势

从Planck函数和热红外辐射传输方程出发,概述了热红外遥感反演地表温度的基本原理,总结了当前反演地表温度常用到的热红外遥感器及相应波段。将热红外遥感地表温度反演算法分为单通道、劈窗和多通道3大类,分析了每一类中较具代表性算法的原理、适用条件及精确度。从热红外遥感机理、发射率、环境辐射、混合像元、大气影响等方面概述了热红外遥感反演地表温度面临的主要问题,并对热红外遥感地表温度反演的发展趋势进行了展望。     

MODIS卫星数据地表反照率反演的简化模式

以内蒙西部地区的MODIS遥感图像数据和地表野外同步观测的光谱数据为例,在野外数据量较少且有定标数据的条件下反演地表反照率。使用6S大气1辐射传输模型进行大气校正,并通过MODTRAN4.0模型获取各波段地表入射光通量和窄波段的地表反照率;在窄波段反照率与宽波段反照率之间存在线性关系的前提下,以各波段的入射光通量占总入射通量的比例作为反演参数,实现窄波段到宽波段的反演。反演结果证明此方法简便可行。     

基于ETM+热红外遥感的武汉市城市地温反演研究

利用ERDAS及ArcGIS软件,通过影像预处理、影像解译,最终提取城区土地分类信息;由单窗算法,以大气辐射传输方程简化的单波段地表温度反演算法为基础,对武汉市2002年ETM+热红外遥感影像及相关气象资料进行地面亮温反演和不同土地利用类型的热效应定量评价研究.研究中采用了基于归一化差值植被指数和基于地表分类相结合的方法确定地表发射率;地温反演引入了热效应贡献度和区域热单元权重指数,对不同地表类型的热贡献度以及植被覆盖与地表温度的关系进行分析.     

利用MODIS数据反演南海南部海表温度及时空变化分析

南海海水表面温度对中国陆地的气候变化具有显著的影响。以南海南部海域为例,首先对MODIS基础数据进行几何校正及影像去云等预处理,利用辐射传输模型MODTRAN计算大气透过率,利用MODIS数据第31和32波段辐射亮度值计算亮度温度,采用劈窗算法反演南海南部海域海表温度,反演结果与产品及实测数据进行回归分析,采取决定系数（R ²）、误差平方和（SSE）及均方根误差（RMSE）进行拟合情况评价。决定系数（R ²）大于0.8,SSE、RMSE较小,其中反演结果与实测数据的SSE为1.025,RMSE为0.158,说明反演精度良好。研究表明:温度具有明显的区域和季节变化特征,秋冬较低,春夏较高,在空间上从离近岸向中心海域方向递减,海盆中心温度低。温度受气候的影响,与厄尔尼诺现象呈正相关,与拉尼娜现象呈负相关。     

基于AMSR-E的全球陆表被动微波发射率数据集

地表微波发射率表征了地物向外发射微波辐射的能力,星载被动微波发射率估算可在宏观、大尺度上对陆表微波辐射进行整体表达,是被动微波地表参数定量反演中重要基础数据,也是在大尺度上获取陆表微波辐射特征的一种途径。本数据集利用搭载在Aqua卫星上的高级微波扫描辐射计(AMSR-E)和中分辨率成像光谱仪(MODIS)的同步观测特点,采用MODIS的地表温度和大气水汽产品数据作为输入,基于考虑大气影响的发射率估算模型,生产了全球晴空条件下AMSR-E传感器运行期间(2002年6月~2011年10月)的陆表多通道双极化微波瞬时发射率。通过产品低频无线电信号影响、数据间比对、分布统计、不同地表覆盖条件的发射率特征、频率依赖和相关性研究等开展验证性分析,结果表明:瞬时发射率的动态大、细节表达丰富,月内日变化标准差在0.02以内,其时空变化、频率依赖和相关性等符合微波理论分析和自然物理过程理解。此套数据集还包括AMSR-E全生命周期的全球陆表逐日、侯、旬、半月及月产品,可用于开展星载被动微波遥感模拟、陆面模型以及陆表温度、积雪、大气降水/水汽/可降水量等反演研究。     

利用MODIS资料计算不同云天条件下的地表太阳辐射

利用MODIS气溶胶和云产品卫星数据与大气辐射传输模式RSTAR,进行了晴空和有云条件下地表太阳辐射计算,并与香河综合辐射站的地基辐射测量值相比较。分析表明,晴空下二者相关性较好,相关系数平方R~2值为0.95,均方根误差RMSE为38.8 W/m~2。有云条件下,计算结果较差于晴空条件下,R~2值为0.88,RMSE为88.2 W/m~2。观测显示,香河站云—气溶胶共存现象较多,而MODIS仅按单一层的云进行微物理参数反演,导致模式输入参数误差,给地表太阳辐射计算结果引入误差。为了分析云-气溶胶共存状态对计算地表太阳辐射的影响,利用RSTAR计算了不同光学厚度的云和气溶胶在特定波段卫星观测的辐亮度值,并对于特定波段卫星接收的辐亮度值,用不同垂直结构的云和气溶胶分别反演其光学和微物理参数,再利用反演的结果分别计算相应的地表太阳辐射。结果表明:相对于单一云层的反演结果,云下气溶胶光学厚度(AOD)为0.1时,由反演误差所导致的地表太阳辐射估算误差较小;而随着AOD增加影响明显增大,在AOD为1.2时,相对误差达17.79%~18.38%。对于污染较重的华北地区而言,分析云覆盖下的气溶胶对地表太阳辐射的影响,有助于提高有云条件下地表太阳辐射的计算精度。     

基于Landsat系列数据地表温度反演算法对比分析——以齐齐哈尔市辖区为例

Landsat热红外系列数据是地表温度反演的一项重要数据源。以齐齐哈尔市辖区为研究区域,基于2002、2008和2016年Landsat TM/ETM+/TIRS系列数据,分别采用单窗算法(MW算法)、单通道算法(SC算法)和辐射传输方程法(RTE算法)进行地表温度反演及对比分析,并利用MODIS地表温度产品对反演结果进行精度验证。结果表明:(1)基于Landsat系列数据,3种算法反演得到的地表温度的空间分布状况一致,总体上市区地表温度较高,水体区域温度最低;(2)基于ETM+数据,SC和RTE算法结果一致性较好,其中SC算法精度最高,MW算法在不同地物覆被区误差均较大;(3)MW算法基于TM数据反演精度最高,RTE算法次之,SC算法较差;(4)基于Landsat 8TIRS数据,SC算法精度最高,RTE算法误差较大。     

Evaluating NDVI-based emissivities of MODIS bands 31 and 32 using emissivities derived by Day/Night LST algorithm

Surface emissivity estimation is a significant factor for the land surface temperature estimation from remotely sensed data. For fully vegetated surfaces, the emissivity estimation is performed in a simple manner since the emissivity is relatively uniform. However, for arid land with sparse vegetation, the estimation is more complicated since the emissivity of the exposed soil and rock is highly variable. In this study, mean and difference emissivity for bands 31 and 32 of MODIS sensor have been derived based on NDVI values. First, the NDVI thresholds have been determined to separate bare soil, partially vegetated soil and fully vegetated land. Then regression relations have been derived to estimate mean and difference emissivity of the bare soil samples and partially vegetated surfaces. A constant emissivity is also used for fully vegetated area. Along with the correlations, standard deviations of the regression relations have been examined for a set of representative soil types. Standard deviations smaller than 0.003 in mean emissivity and smaller than 0.004 in difference emissivity are resulted in regression linear relations. Evaluation of the NDVI derived regression relations has been performed using the results of MODIS Day/Night Land Surface Temperature (LST) algorithm on a pair of MODIS images. Using around 45,500 pixels with different soil and land cover types, emissivity of each pixel in bands 31 and 32 have been estimated. The calculated emissivities have been compared with emissivities calculated by MODIS Day/Night LST algorithm. Biases and standard deviations of NDVI-based relations show relatively high agreement for mean and difference emissivity relations with Day/Night method results. It may be concluded that the proposed algorithm can be used as a rather simple alternative to complex emissivity estimation algorithms.     

Comparison of land surface emissivity and radiometric temperature derived from MODIS and ASTER sensors

This study compared surface emissivity and radiometric temperature retrievals derived from data collected with the MODerate resolution Imaging Spectroradiometer (MODIS) and Advanced Spaceborne Thermal Emission Reflection Radiometer (ASTER) sensors, onboard the NASA's Earth Observation System (EOS)-TERRA satellite. Two study sites were selected: a semi-arid area located in northern Chihuahuan desert, USA, and a Savannah landscape located in central Africa. Atmospheric corrections were performed using the MODTRAN 4 atmospheric radiative transfer code along with atmospheric profiles generated by the National Center for Environmental Predictions (NCEP). Atmospheric radiative properties were derived from MODTRAN 4 calculations according to the sensor swaths, which yielded different strategies from one sensor to the other. The MODIS estimates were then computed using a designed Temperature-Independent Spectral Indices of Emissivity (TISIE) method. The ASTER estimates were derived using the Temperature Emissivity Separation (TES) algorithm. The MODIS and ASTER radiometric temperature retrievals were in good agreement when the atmospheric corrections were similar, with differences lower than 0.9 K. The emissivity estimates were compared for MODIS/ASTER matching bands at 8.5 and 11 μm. It was shown that the retrievals agreed well, with RMSD ranging from 0.005 to 0.015, and biases ranging from −0.01 to 0.005. At 8.5 μm, the ranges of emissivities from both sensors were very similar. At 11 μm, however, the ranges of MODIS values were broader than those of the ASTER estimates. The larger MODIS values were ascribed to the gray body problem of the TES algorithm, whereas the lower MODIS values were not consistent with field references. Finally, we assessed the combined effects of spatial variability and sensor resolution. It was shown that for the study areas we considered, these effects were not critical.     

Preliminary estimate of calibration of the moderate resolution imaging spectroradiometer thermal infrared data using Lake Titicaca

The absolute radiometric accuracy of Moderate Resolution Imaging Spectroradiometer (MODIS) thermal infrared (TIR) data was evaluated with in situ data collected in a vicarious calibration field campaign conducted in Lake Titicaca, Bolivia during May 26 and June 17, 2000. The comparison between MODIS TIR data produced by the version 2.5.4 Level-1B code and the band radiances calculated with atmospheric radiative transfer code MODTRAN4.0 based on lake surface kinetic temperatures measured by five IR radiometers deployed in the high-elevation Lake Titicaca and the atmospheric temperature and water vapor profiles measured by radiosondes launched on the lake shore on June, 15 2000, a calm clear-sky day, shows good agreements in bands 31 and 32 (within an accuracy of 0.4%) in the daytime overpass case. Sensitivity analysis indicates that the changes on the measured atmospheric temperature and water vapor profiles result in negligible or small effects on the calculated radiances in the atmospheric window bands (bands 20–23, 29, and 31–32). Therefore, comparisons for these bands were made for cases when lake surface temperature measurements were available but no radiosonde data were available and in subareas of 10×16 pixels where there was no in situ measurement but MODIS brightness temperatures in band 31 vary within ±0.15 K by using the validated band 31 to determine lake surface temperatures through the MODTRAN4.0 code. Comparisons and error analysis show that the specified absolute radiometric accuracies are reached or nearly reached in MODIS bands 21, 29, and 31–33 and that there is a calibration bias of 2–3% in bands 20, 22, and 23. The error analysis also shows that the radiosondes cannot provide accurate atmospheric temperature and water vapor profiles to estimate the calibration accuracies in the atmospheric sounding bands (bands 24–25, 27–28, and 34–36) at the specified 1% level and that the calibration accuracy in the ozone band 30 cannot be estimated without in situ measurements of ozone.     

融合遥感先验信息的叶面积指数反演

借助植被辐射传输模型，利用遥感观测数据估算LAI是一种较为可靠和稳健的反演方法。然而，地表的复杂性、遥感观测的有限性以及自相关性导致遥感数据包含的信息量不足，不能完全支持LAI等地表参数的估算，易造成“病态”反演。在遥感反演过程中引入先验知识能够有效地解决该问题。研究基于遥感数据提取LAI先验信息，并将其用于代价函数的构建，利用PROSAIL辐射传输模型和遗传算法，分别在500 m和250 m尺度反演LAI。将高空间分辨率LAI分别升尺度到500 m和250 m，验证对应尺度LAI结果，评价引入先验信息对于提高LAI反演精度的作用。研究表明，引入先验信息有助于提高不同分辨率下LAI反演精度，且先验信息的质量一定程度上也影响着LAI反演结果。与未加入先验信息的LAI反演结果相比，以MODIS LAI产品作为先验信息反演的500 m尺度LAI结果精度R²由0.55提高至0.65，RMSE由1.29下降至0.38。在250 m尺度，以500 m LAI反演结果作为先验信息反演的叶面积指数，其精度优于以MODIS LAI产品为先验知识的估算结果，验证精度R²增加了0.08，RMSE减少了0.18。研究使用的先验信息主要来自遥感数据本身，没有地面实测数据的参与，在此基础上发展的多分辨率LAI反演方法具有估算大区域尺度LAI的应用潜力。     

Evaluation and improvement of the MODIS land surface temperature/emissivity products using ground‐based measurements at a semi‐desert site on the western Tibetan Plateau

Current MODerate‐resolution Imaging Spectroradiometer (MODIS) land surface temperature (LST, surface skin temperature)/emissivity products are evaluated and improvements are investigated. The ground‐based measurements of LST at Gaize (32.30° N, 84.06° E, 4420 m) on the western Tibetan Plateau from January 2001 to December 2002 agree well (mean and standard deviation of differences of 0.27 K and 0.84 K) with the 1‐km Version 004 (V4) Terra MODIS LST product (MOD11A1) generated by the split‐window algorithm. Spectral emissivities measured from surface soil samples collected at and around the Gaize site are in close agreement with the landcover‐based emissivities in bands 31 and 32 used by the split‐window algorithm. The LSTs in the V4 MODIS LST/emissivity products (MYD11B1 for Aqua and MOD11B1 for Terra) from the day/night LST algorithm are higher by 1–1.7 K (standard deviation around 0.6 K) in comparisons to the 5‐km grid aggregated values of the LSTs in the 1‐km products, which is consistent with the results of a comparison of emissivities. On average, the emissivity in MYD11B1 (MOD11B1) is 0.0107 (0.0167) less than the ground‐based measurements, which is equivalent to a 0.64 K (1.25 K) overestimation of LST around the average value of 285 K. Knowledge obtained from the evaluation of MODIS LST/emissivity retrievals provides useful information for the improvement of the MODIS LST day/night algorithm. Improved performance of the refined (V5) day/night algorithm was demonstrated with the Terra MODIS data in May–June 2004.     

融合遥感先验信息的叶面积指数反演

借助植被辐射传输模型,利用遥感观测数据估算LAI是一种较为可靠和稳健的反演方法。然而,地表的复杂性、遥感观测的有限性以及自相关性导致遥感数据包含的信息量不足,不能完全支持LAI等地表参数的估算,易造成“病态”反演。在遥感反演过程中引入先验知识能够有效地解决该问题。研究基于遥感数据提取LAI先验信息,并将其用于代价函数的构建,利用PROSAIL辐射传输模型和遗传算法,分别在500 m和250 m尺度反演LAI。将高空间分辨率LAI分别升尺度到500 m和250 m,验证对应尺度LAI结果,评价引入先验信息对于提高LAI反演精度的作用。研究表明,引入先验信息有助于提高不同分辨率下LAI反演精度,且先验信息的质量一定程度上也影响着LAI反演结果。与未加入先验信息的LAI反演结果相比,以MODIS LAI产品作为先验信息反演的500 m尺度LAI结果精度R²由0.55提高至0.65,RMSE由1.29下降至0.38。在250 m尺度,以500 m LAI反演结果作为先验信息反演的叶面积指数,其精度优于以MODIS LAI产品为先验知识的估算结果,验证精度R²增加了0.08,RMSE减少了0.18。研究使用的先验信息主要来自遥感数据本身,没有地面实测数据的参与,在此基础上发展的多分辨率LAI反演方法具有估算大区域尺度LAI的应用潜力。     

Optimal modalities for radiative transfer-neural network estimation of canopy biophysical characteristics: Evaluation over an agricultural area with CHRIS/PROBA observations

Neural networks trained over radiative transfer simulations constitute the basis of several operational algorithms to estimate canopy biophysical variables from satellite reflectance measurements. However, only little attention was paid to the training process which has a major impact on retrieval performances. This study focused on the several modalities of the training process within neural network estimation of LAI, FCOVER and FAPAR biophysical variables. Performances were evaluated over both actual experimental observations and model simulations. The SAIL and PROSPECT radiative transfer models were used here to simulate the training and the synthetic test datasets. Measurements of LAI, FCOVER and FAPAR were achieved over the Barrax (Spain) agricultural site for a range of crop types concurrently to CHRIS/PROBA satellite image acquisition. Results showed that the spectral band selection was specific to LAI, FCOVER and FAPAR variables. The optimal band set provided significantly improved performances for LAI, while only small differences were observed for the other variables. Gaussian distributions of the radiative transfer model input variables performed better than uniform distributions for which no prior information was exploited. Including moderate uncertainties in the reflectance simulations used in the training process improved the flexibility of the neural network in cases where simulations departed slightly from observations. Simple neural network architecture with a single hidden layer of five tangent sigmoid transfer functions was performing as good as more complex architectures if the training dataset was larger than ten times the number of coefficients to tune. Small sensitivity of performances was observed depending on the way the solution was selected when several networks were trained in parallel. Finally, comparison with a NDVI based approach showed the generally better retrieval accuracy of neural networks.     

Derivation of land surface temperatures from MODIS data using the general split‐window technique

Fast Atmospheric Signature Code (FASCODE), a line‐by‐line radiative transfer programme, was used to simulate Moderate Resolution Imaging Spectroradiometer (MODIS) data at wavelengths 11.03 and 12.02 µm to ascertain how accurately the land surface temperature (LST) can be inferred, by the split‐window technique (SWT), for a wide range of atmospheric and terrestrial conditions. The approach starts from the Ulivieri algorithm, originally applied to Advanced Very High Resolution Radiometer (AVHRR) channels 4 and 5. This algorithm proved to be very accurate compared to several others and takes into account the atmospheric effects, in particular the water vapour column (WVC) amount and a non‐unitary surface emissivity. Extended simulations allowed the determination of new coefficients of this algorithm appropriate to MODIS bands 31 and 32, using different atmospheric conditions. The algorithm was also improved by removing some of the hypothesis on which its original expression was based. This led to the addition of a new corrective term that took into account the interdependence between water vapour and non‐unitary emissivity values and their effects on the retrieved surface temperature. The LST products were validated within 1 K with in situ LSTs in 11 cases.     

Retrieval of land surface albedo and temperature using data from the Indian geostationary satellite: a case study for the winter months

The shortwave and longwave radiation budget at land surfaces is largely dependent on two fundamental quantities, the albedo and the land surface temperature (LST). A time series (November 2005 to March 2006) of daily data from the Indian geostationary satellite Kalpana‐1 Very High Resolution Radiometer (K1VHRR) sensor in the visible (VIS), water vapour (WV) and thermal infrared (TIR) bands from noontime (0900 GMT) observations were processed to retrieve these quantities in clear skies for five winter months. Cloud detection was carried out using bispectral threshold tests (in both VIS and TIR bands) in a dekadal time series. Surface albedo was retrieved using a simple atmospheric transmission model. K1VHRR albedo was compared with Moderate Resolution Imaging Spectroradiometer (MODIS) AQUA noontime albedo over different land targets (agriculture, forest, desert, scrub and snow) that showed minimum differences over agriculture and forest. The comparison of spatial albedo over different landscapes yielded a root mean square deviation (RMSD) of 0.021 in VHRR albedo (9% of MODIS albedo). A mono‐window algorithm was implemented with a single TIR band to retrieve the LST. Its accuracy was also verified over different land targets by comparison with aggregated MODIS AQUA LST. The maximum RMSD was obtained over agriculture. Spatial comparison of VHRR and AQUA LSTs over homogeneous and heterogeneous landscape cutouts revealed an overall RMSD of 2.3 K. An improvement in the retrieval accuracy is expected to be achieved with atmospheric products from the sounder and split thermal bands in the imager of future INSAT 3D missions.