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.     

Retrieval of Snow Depth on Sea Ice in the Arctic Using the FengYun-3B Microwave Radiation Imager

Snow on sea ice is a sensitive indicator of climate change because it plays an important role regulating surface and near surface air temperatures. Given its high albedo and low thermal conductivity, snow cover is considered a key reason for amplified warming in polar regions. This study focuses on retrieving snow depth on sea ice from brightness temperatures recorded by the Microwave Radiation Imager(MWRI) on board the FengYun(FY)-3 B satellite. After cross calibration with the Advanced Microwave Scanning Radiometer-EOS(AMSR-E) Level 2 A data from January 1 to May 31, 2011, MWRI brightness temperatures were used to calculate sea ice concentrations based on the Arctic Radiation and Turbulence Interaction Study Sea Ice(ASI) algorithm. Snow depths were derived according to the proportional relationship between snow depth and surface scattering at 18.7 and 36.5 GHz. To eliminate the influence of uncertainties in snow grain sizes and sporadic weather effects, seven-day averaged snow depths were calculated. These results were compared with snow depths from two external data sets, the IceBridge ICDIS4 and AMSR-E Level 3 Sea Ice products. The bias and standard deviation of the differences between the MWRI snow depth and IceBridge data were respectively 1.6 and 3.2 cm for a total of 52 comparisons. Differences between MWRI snow depths and AMSR-E Level 3 products showed biases ranging between-1.01 and-0.58 cm, standard deviations from 3.63 to 4.23 cm, and correlation coefficients from 0.61 to 0.79 for the different months.     

Retrieval Snow Depth by Artificial Neural Network Methodology from Integrated AMSR-E and In-situ Data —— A Case Study in Qinghai-Tibet Plateau

On the basis of artificial neural network （ANN） model, this paper presents an algorithm for inversing snow depth with use of AMSR-E （Advanced Microwave Scanning Radiometer-Earth Observing System （EOS）） dataset, i.e., brightness temperature at 18.7 and 36.5GHz in Qinghai-Tibet Plateau during the snow season of 2002-2003. In order to overcome the overfitting problem in ANN modeling, this methodology adopts a Bayesian regularization approach. The experiments are performed to compare the results obtained from the ANN-based algorithm with those obtained from other existing algorithms, i.e., Chang algorithm, spectral polarization difference （SPD） algorithm, and temperature gradient （TG） algorithm. The experimental results show that the presented algorithm has the highest accuracy in estimating snow depth. In addition, the effects of the noises in datasets on model fitting can be decreased due to adopting the Bayesian regularization approach.     

基于多源数据的青藏高原雪深重建

青藏高原地形复杂,积雪时空分布异质性较强且大部分地区积雪较薄,而被动微波遥感因其空间分辨率低以及雪深反演中的不确定性,极大地限制了其反演青藏高原雪深的精度。本文尝试将多源遥感数据以及与积雪模型（SnowModel）相结合,来重建更高质量的青藏高原雪深数据。首先,利用MODIS积雪面积比例产品,根据构建的积雪衰减曲线以及经验的融合规则对低分辨率被动微波雪深进行了降尺度;然后,结合MODIS/被动微波融合雪深数据和SnowModel对研究区进行雪深数据同化实验;最后,利用地面站实测雪深数据对MODIS/被动微波融合雪深以及同化输出雪深的精度进行了分析和对比。结果表明,基于数据同化方法得到的雪深数据更接近地面观测雪深值,通过均方根误差以及相关系数的对比,同化雪深结果优于MODIS/被动微波融合雪深结果。     

An AMSR-E data unmixing method for monitoring flood and waterlogging disaster

Spectral remote sensing technique is usually used to monitor flood and waterlogging disaster. Although spectral remote sensing data have many advantages for ground information observation, such as real time and high spatial resolution, they are often interfered by clouds, haze and rain. As a result, it is very difficult to retrieve ground information from spectral remote sensing data under those conditions. Compared with spectral remote sensing technique, passive microwave remote sensing technique has obvious superiority in most weather conditions. However, the main drawback of passive microwave remote sensing is the extreme low spatial resolution. Considering the wide application of the Advanced Microwave Scanning Radiometer-Earth Observing System (AMSR-E) data, an AMSR-E data unmixing method was proposed in this paper based on Bellerby’s algorithm. By utilizing the surface type classification results with high spatial resolution, the proposed unmixing method can obtain the component brightness temperature and corresponding spatial position distribution, which effectively improve the spatial resolution of passive microwave remote sensing data. Through researching the AMSR-E unmixed data of Yongji County, Jilin Provinc, Northeast China after the worst flood and waterlogging disaster occurred on July 28, 2010, the experimental results demonstrated that the AMSR-E unmixed data could effectively evaluate the flood and waterlogging disaster.     

Effects of snow cover on ground thermal regime: A case study in Heilongjiang Province of China

The important effects of snow cover to ground thermal regime has received much attention of scholars during the past few decades. In the most of previous research, the effects were usually evaluated through the numerical models and many important results are found. However, less examples and insufficient data based on field measurements are available to show natural cases. In the present work, a typical case study in Mohe and Beijicun meteorological stations, which both are located in the most northern tip of China, is given to show the effects of snow cover on the ground thermal regime. The spatial(the ground profile) and time series analysis in the extremely snowy winter of 2012–2013 in Heilongjiang Province are also performed by contrast with those in the winter of 2011–2012 based on the measured data collected by 63 meteorological stations. Our results illustrate the positive(warmer) effect of snow cover on the ground temperature(GT) on the daily basis, the highest difference between GT and daily mean air temperature(DGAT) is as high as 32.35℃. Moreover, by the lag time analysis method it is found that the response time of GT from 0 cm to 20 cm ground depth to the alternate change of snow depth has 10 days lag, while at 40 cm depth the response of DGAT is not significant. This result is different from the previous research by modeling, in which the response depth of ground to the alteration of snow depth is far more than 40 cm.     

1979-2010年青藏高原积雪深度时空变化遥感分析

积雪深度是表征积雪特征的重要参数,也是区域气候变化最敏感的响应因子之一。利用1979-2010年逐日中国雪深长时间序列数据集,采用GIS空间分析和地统计方法,分析了青藏高原积雪深度的时空变化规律及异常空间分布特征。结果表明：近32年来,青藏高原雪深呈显著增加趋势,增加速率为0.26 cm/10a,其中,昆仑高寒荒漠地带雪深增加最为明显,增加速率达0.73 cm/10a;20世纪80年代至90年代青藏高原雪深呈逐步增加趋势,21世纪初变化平稳;青藏高原4个季节雪深变化均呈现为上升趋势,尤以冬季增加最为明显,增加速率达0.57 cm/10a。青藏高原东南、西部和南部为雪深分布高值区;逐像元回归分析表明,高原雪深呈增加趋势的像元数占全区像元总数的67.1%,其中有91.3%为轻度和中度增加,主要分布在高原北部和西部;最大雪深变化基本维持在-0.1~0.1 cm/a（45.47%）之间,在昆仑北翼山地、柴达木山地、羌塘高寒地带南部等局部地区最大雪深有增加趋势,主要是轻度增加,面积比例为36.66%。果洛那曲高寒地带、青南高寒地带和羌塘高寒地带为青藏高原积雪深度异常变化敏感区。     

全球36 km格网土壤水分逐日估算

土壤水分是陆面生态系统和能量循环的核心变量之一,利用微波遥感技术获得的土壤水分产品的时间分辨率一般是2-3 d,因此精确地获得具有较高时间分辨率的土壤水分成了人们关注的焦点。本文尝试将SMAP (the Soil Moisture Passive and Active）土壤水分和MODIS光学数据相结合,利用广义回归神经网络进行全球36 km土壤水分的估算,提升SMAP土壤水分的时间分辨率。结果显示,广义回归神经网络估算土壤水分与SMAP保持了高相关性(r = 0.7528),但其却保留了较高的误差 (rmse = 0.0914 m³/m³)。尽管如此,估算的土壤水分能够很好地保持SMAP土壤水分的整体空间变化,并且提升了土壤水分的时间分辨率（1 d）。此处,本文研究了SMAP土壤水分与MODIS光学数据之间的关系,这对今后利用机器学习进行SMAP土壤水分降尺度研究提供了重要的参考价值。     

光学信息分解被动微波土壤湿度方法

土壤水分是一个重要生态参量,以被动微波反演土壤水分,不受天气影响,且其算法成熟.但是星载被动微波数据的空间分辨率较低,可适合大区域尺度研究.本文将1km分辨率光学数据MODIS和25km分辨率被动微波数据AMSR- E2级土壤湿度产品结合,利用NDVI-Ts特征空间,去除植被影响,结合前人提出的裸土蒸散模型,将研究区被...     

2008年南方冰雪过程对生态环境影响的遥感监测——以湖南省资兴市为例

2008年1月中下旬至2月上旬,我国南方地区出现了大范围、长时间的强降雪及冰冻灾害。这次冰冻雨雪灾害,不仅对广大人民群众的生活和社会经济运行造成了极大破坏,而且对当地生态环境也产生了深刻影响。本文以湖南省资兴市为例,运用遥感技术和地面野外调查相结合的方法,监测了本次冰雪过程对当地生态环境的影响。监测发现:本次冰雪过程对植被覆盖破坏显著,植被指数呈较大范围降低趋势,降低区域的总面积占资兴市土地总面积的54.12%;森林生态系统受损严重,除灌木林外302.2万亩森林全部受灾,其中,针叶林比阔叶林严重,幼龄林比成熟林严重,外来种比乡土树种严重,人工林比天然林严重;冰雪过程后土壤含水量增加、边坡失稳,地质灾害危险程度明显增高,工程建设区和林缘陡坡地带以滑坡(崩塌)为主的地质灾害频频发生。     

Snow cover influences the thermal regime of active layer in Urumqi River Source,Tianshan Mountains,China

Snow cover is characterized by the high albedo, low thermal conductivity, and notable heat transition during phase changes. Thus, snow cover significantly affects the ground thermal regime. A comparison of the snow cover in high latitudes or high-altitude snowy mountain regions indicates that the eastern Tianshan Mountains (China) show a characteristically thin snow cover (snow depth below 15 cm) with remarkable temporal variability. Based on snow depth, heat flux, and ground temperature from 2014 to 2015 in the Urumqi River source, the spatialtemporal characteristics of snow cover and snow cover influences on the thermal conditions of active layer in the permafrost area were analyzed. During the autumn (Sept. - Oct.), thin and discontinuous snow cover can noticeably accelerate the exothermic process of the ground, producing a cooling effect on the shallow soil. During the winter (Nov. - Mar.), it is inferred that the effective thermal insulation starts with snow depth exceeding 10 cm during early winter. However, the snow depth in this area is generally below 15 cm, and the resulting snow-induced thermal insulation during the winter is very limited. Due to common heavy snowfalls in the spring (Apr. to May), the monthly mean snow thickness in April reached to 15 cm and remained until mid-May. Snow cover during the spring significantly retarded the ground warming. Broadly, snow cover in the study area exerts a cooling effect on the active layer and plays a positive role in the development and preservation of permafrost.     

Verification and recalibration of HY-2A microwave radiometer brightness temperature

HY-2A is the first one of the Chinese HY-2 ocean satellite series carrying a microwave radiometer(RM) to measure sea surface temperature,sea surface wind speed,atmospheric water vapor,cloud liquid water content, and rain rate. We verified the RM level 1B brightness temperature(T_B) to retrieve environmental parameters. In the verification, TB that simulated using the ocean-atmosphere radiative transfer model(RTM) was used as a reference. The total bias and total standard deviation(SD) of the RM level 1B T_B, with reference to the RTM simulation, ranged-20.6-4.38 K and 0.7-2.93 K, respectively. We found that both the total bias and the total SD depend on the frequency and polarization, although the values for ascending and descending passes are different. In addition, substantial seasonal variation of the bias was found at all channels. The verification results indicate the RM has some problems regarding calibration, e.g.,correction of antenna spillover and antenna physical emission, especially for the 18.7-GHz channel. Based on error analyses, a statistical recalibration algorithm was designed and recalibration was performed for the RM level 1B T_B. Validation of the recalibrated TB indicated that the quality of the recalibrated RM level 1B T_B was improved significantly. The bias of the recalibrated TB at all channels was reduced to 0.4 K, seasonal variation was almost eradicated, and SD was diminished(i.e., the SD of the 18.7-GHz channel was reduced by more than 0.5 K).     

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。     

MODIS_LST与AMSR-E_BT的相关性及地表温度反演

本文以2007年和2008年MODIS每日地表温度（LST）数据及AMSR-E地表亮温（BT）数据为研究对象,结合土地覆盖类型数据,统计分析MODIS_LST与AMSR-E_BT在不同土地覆盖类型、频率和极化方式条件下的相关性。结果表明,频率在18.7、23.8和36.5 GHz的AMSR-E-BT与MODIS_LST的相关性较大,且在垂直极化通道上的相关性较在水平极化上大;不同土地覆盖类型,与MODIS_LST相关性较大所对应的AMSR-E微波通道不同。同时,考虑混合像元问题对相关性的影响,对25种不同地物类型组合下MODIS_LST与AMSR-E-BT的相关性进行统计分析,发现混合像元中地物类型越多,则二者相关性越小。最后,采用多元线性回归分析法,根据不同土地覆盖类型建立反演回归模型,对部分研究区域MODIS-LST进行反演,误差平均在±3.15 K以内,与不考虑下垫面覆盖的模型比较,反演MODIS_LST精度平均提高了1.5 K。     

基于全连接和LSTM神经网络的空气污染物预测

在空气污染日益严重的情况下进行空气污染物的预测工作是十分必要的。针对城市的空气污染物预测,提出了一种基于神经网络的混合模型方法:通过全连接神经网络方法,结合长短期记忆网络(Long Short-Term Memory,LSTM)方法,将历史空气污染物数据与大气数据进行空间与时间上的挖掘分析。运用全连接和LSTM两种神经网络方法混合的形式,与传统的单一模型方法相比,不仅能摆脱单一模型特征空间的局限性,还能提高预测的精度,具有更大的应用性和操作性。最后,以武汉市为例通过实验证明该混合模型较单一模型在空气污染物预测上具有更高的精度。     

Spatial distribution of snow depth based on geographically weighted regression kriging in the Bayanbulak Basin of the Tianshan Mountains,China

Snow depth is a general input variable in many models of agriculture,hydrology,climate and ecology.This study makes use of observational data of snow depth and explanatory variables to compare the accuracy and effect of geographically weighted regression kriging(GWRK)and regression kriging(RK)in a spatial interpolation of regional snow depth.The auxiliary variables are analyzed using correlation coefficients and the variance inflation factor(VIF).Three variables,Height,topographic ruggedness index(TRI),and land surface temperature(LST),are used as explanatory variables to establish a regression model for snow depth.The estimated spatial distribution of snow depth in the Bayanbulak Basin of the Tianshan Mountains in China with a spatial resolution of 1 km is obtained.The results indicate that 1)the result of GWRK's accuracy is slightly higher than that of RK(R~2=0.55 vs.R~2=0.50,RMSE(root mean square error)=0.102 m vs.RMSE=0.077 m);2)for the subareas,GWRK and RK exhibit similar estimation results of snow depth.Areas in the Bayanbulak Basin with a snow depth greater than 0.15m are mainly distributed in an elevation range of 2632.00–3269.00 m and the snow in this area comprises 45.00–46.00% of the total amount of snow in this basin.However,the GWRK resulted in more detailed information on snow depth distribution than the RK.The final conclusion is that GWRK is better suited for estimating regional snow depth distribution.     

基于微波数据与光学数据集成的机器学习技术在作物产量估算中的应用

各类光学植被指数已成功地应用于各种植被监测与作物产量估算中,但这些指数易受大气状况的影响。由星载微波辐射计得到的植被光学厚度数据（VOD）与植被密度、含水量密切相关,数据可全天候获得,在农业遥感监测中呈现着巨大的潜力。作为来自不同传感器的遥感数据,微波遥感数据与光学遥感数据可以提供不同波长范围内的植被信息。为了更准确地进行作物产量估算,本研究提出将微波遥感数据与光学遥感数据共同应用于冬小麦单产估算中。研究选择L波段微波辐射计SMAP卫星的VOD数据与MODIS的标准归一化植被指数NDVI、增强型植被指数EVI、叶面积指数LAI、光合有效辐射分量FPAR数据作为研究变量,分别使用BP神经网络、GA-BP神经网络和PSO-BP神经网络建立冬小麦产量估算模型。结果表明： 3种神经网络回归模型的P值均小于0.001,通过了显著性检验。GA-BP神经网络回归模型的估算值与真实值在3种神经网络回归模型中表现了最高的相关性（R=0.755）与最低的均方根误差（RMSE=529.145 kg/hm²）,平均绝对误差(MAE=425.168 kg/hm²)和平均相对误差（MRE=6.530%）。为了分析多源遥感数据的结合在作物产量估算中的优势,研究同时构建了仅使用NDVI和LAI,使用NDVI、EVI、LAI、FPAR等光学数据进行冬小麦产量估算的3种GA-BP神经网络回归模型作为对比。结果表明,使用微波遥感数据与光学遥感数建立的GA-BP神经网络回归模型较上述3种作为对比的GA-BP神经网络回归模型的相关系数R值分别提高了0.163,0.229与0.056,均方根误差RMSE分别降低了122.334、158.462和46.923 kg/hm²,使用多源遥感数据的组合可以很好地提高作物产量估算的准确性。     

利用GPS的SNR观测值进行雪深监测研究

在分析多路径与信噪比关系的基础上,给出基于SNR观测值的GPS-MR技术探测雪深的基本原理。利用板块监测PBO网P360站2013年174 d～2014年151 d的GPS实测数据进行算例分析,其结果与实测雪深数据吻合较好,相关系数大于0.97。     

土壤水分微波反演方法进展和发展趋势

土壤水分是连接地表水循环和能量循环的关键参量,精确获取该参量对于理解气候变化、地表水文过程、地气间能量交换机理等具有重要意义。微波遥感由于其较为合适的探测深度和坚实的理论基础在观测地表浅层土壤水分上具有很大优势,结合反演方法可以获取空间连续的土壤水分含量,有助于更加客观认知土壤水分的时空演变机理。随着微波遥感数据的不断丰富,多种微波遥感土壤水分反演方法相继涌现,为了更好地了解其发展和趋势,本文总结了当前土壤水分微波反演常用的卫星遥感数据并分析其发展趋势,后从主动微波反演、被动微波反演和多源协同反演3个方面梳理了各类土壤水分微波反演方法的原理、发展和优缺点,最终总结出目前微波遥感土壤水分反演方法的发展趋势：即土壤水分微波反演方法的时空普适性逐渐增强、面向高时空分辨率的土壤水分微波协同反演方法快速发展以及土壤水分微波反演方法的智能化水平不断提高。     

3种土壤冻融判别算法在青藏高原的分类精度评价

本文比较了基于AMSR-E被动微波数据的3种土壤冻融判别算法在青藏高原相关地区的分类精度。3种算法分别是：双指标算法、决策树算法、判别函数算法。本文选取了来自青藏高原那曲、玛曲、阿里3个地区土壤温湿度观测网的地表温度数据,并结合AMSR-E被动亮温数据,对上述算法在以上地区的分类精度分别进行了比较评价。结果表明：不论是白天还是夜间,相较于干旱区微波信号来自深层土壤的难以准确探测,在青藏高原半湿润半干旱区算法可取得相对较好的判别准确率;双指标算法相较于其他2种算法,在观测区具有较高的分类精度,且夜间分类精度高于白天;实测数据存在资料代表性不普遍即网格所包含站点信息量不够的问题,这也是后续工作中提高分类精度值得关注的着手点。