Lake ice thickness (LIT) significantly impacts water, climate, and socio-economic activities. Due to lack of observations, estimating the spatiotemporal variations in LIT is still a challenging task. Brightness temperature (TB) measured by microwave radiometers correlates well with LIT at certain frequencies, which potentially enables global and daily estimations of LIT by spaceborne passive microwave (PMW) remote sensing over decades. In this study, a generalized empirical linear regression algorithm to estimate LIT using PMW is proposed using TB as the sole data source. The method is based on statistical analysis of TB over 18 lakes. The analysis revealed a large variability of linear coefficients relating TB to LIT; however, the coefficients could be tied to TB at the Freeze-up End (FUE). When compared with LITs derived from in-situ measurements, radar altimetry, and published LIT data, the root mean square errors (RMSE) were 0.19 m, 0.16 m, and 0.15 m, respectively. This method was then used to estimate the LITs for 96 lakes in the Northern Hemisphere from 2002 to 2011, which was proven to be applicable for long-term LIT mapping and monitoring for large lakes, demonstrating improved generalizability.

As Arctic sea ice rapidly declines, accurate daily sea ice concentration (SIC) forecasts become crucial for Arctic research and operations. Emerging deep learning (DL) forecast models have the advantage of consuming fewer computational resources compared to numerical forecast systems. However, DL forecast models face challenges of lacking sea ice domain knowledge, low forecast accuracy in the marginal ice zone (MIZ), and poor explainability. This study proposes an explainable DL model, SIFNet, for daily Arctic SIC forecasts, which considers the domain knowledge of teleconnections and lagged correlations by integrating the convolutional block attention module (CBAM) and temporal feature convolution module (TFCM). To improve SIFNet’s accuracy in the MIZ, a MIZ-weighted loss function is employed for fine-tuning, leading to a notable 1.36% reduction in the mean absolute error (MAE) of MIZ forecasts (MIZ MAE). In the January 2020 to December 2023 test scenario, forecasting SIC for the future 49 days using the past 21 days’ data, SIFNet demonstrates forecast accuracies with the MAE of 4.69%, binary accuracy (BACC) of 95.16%, structural similarity index (SSIM) of 95.13%, and MIZ MAE of 18.22%. Compared to the other two DL models and nine numerical forecast systems, SIFNet’s forecasts achieve higher accuracies. Employing Layer-wise Relevance Propagation (LRP) technique, the LRP-|z| rule is constructed to improve SIFNet’s explainability. During Arctic minimum sea ice extent (SIE) periods, SIFNet forecasts confirm upward and downward surface solar radiation as predictors for minimum SIE. Similarly, over the past thirty years, precipitation phase is considered a potentially significant factor influencing SIC forecasts.

This paper presents a novel deep learning model for short-term sea ice forecasting that accounts for the effects of teleconnection and delay phenomena on sea ice changes. The proposed model is a convolutional neural network with an embedded attention mechanism that takes into account the past 21 days of Arctic environmental element data as input. It is capable of providing daily forecasts of sea ice concentration in the Arctic core area for the next 49 days. Notably, the model outperforms both the deep learning sea ice forecast model Icenet and the numerical sea ice forecasting models ArcIOPS and SEAS5, demonstrating its superior short-term forecasting capability for sea ice.

With the passive microwave remote sensing of snow, the high frequency (80-100 GHz) has the advantages of high resolution and fresh shallow snow detection, however, the application of high frequency for snow observation is often limited by atmosphere influence. Until now, few studies have quantified the atmospheric influence of high frequency in snow observations. The scattering and emission of the water vapor and cloud liquid water in the atmosphere cause the increment in Brightness Temperature (TB). To explore the influence of the atmosphere on snow cover observations, we utilize Nordic Snow Radar Experiment (NoSREx) data and the Microwave Emission Model of Layered Snowpacks (MEMLS) to quantify the influence of the atmospheric downward radiance in snow observation by using 90 GHz of the ground-based radiometer. The results show that, in Sodankylä of Finland, atmospheric contribution to ground-based radiometer at 90GHz 50-degree is up to 95.76K (H-pol) and 95.02K (V-pol), with an average of 25.36K (V-pol) and 25.95K (H-pol). The further calculated root mean square errors (RMSEs) of simulated TB with Tdown and observed TB are 41.40K (V-pol)/36.85(H-pol), and of simulated TB without Tdown are 41.25K (V-pol) and 36.72K (H-pol), respectively, the MEMLS slightly increased the simulation accuracy quantitatively of the snow emission without the influence of the atmosphere.

Sustainable development poses a good way for the world of Arctic, especially the temperature and vegetation has been experiencing a rising trend in the north land area. The research aims to understand the impacts of the gas flaring to the local environment through the time series analysis of vegetation changes around the gas flaring sites, and natural development sites of from 2000 to 2021 over Russia Siberia regions, the Timan-Pechora basin and West Siberia basin oil/gas rich area. The results shows that gas flaring sites has a profound to the vegetation changes at a statistical level, with the temperature analysis, obviously the flaring shows the positive contribution to the regional environment. The result indicates that the gas flaring.

Lake ice phenology is regarded as a good proxy for the past and present climates. Long time series passive microwave radiometry data are used to estimate lake ice phenology variations in the Qinghai–Tibet Plateau (QTP), and a contrasting pattern of phenology change trend is found that the time series trend of lake ice freeze-up or break-up time is obviously reversed for lakes in the QTP. The reason for this contrasting trend of lake ice phenology is discussed based on factors such as salinity, water volume change, and air temperature change. Lake ice phenology data are separated based on lake salinity for the climate study: lake ice phenology of lakes with low salinity can be used as air temperature and climate change indicator, whereas lake ice phenology of lakes with high salinity and a low water volume can be used as an indicator of water volume variation under climate change. Correlation analysis of air temperature and the lake ice phenology show that air temperature is the main driving factor behind lake ice phenology variations. The lake ice phenology results suggest overall rising air temperatures during the period 1987–2017 in all regions of the QTP.

Variable atmospheric conditions are typically ignored in the retrieval of geophysical parameters of the Earth’s surface when using spaceborne passive microwave observations. However, high frequencies, for example, 91.7 GHz, are sensitive to variable atmospheric absorption, even in winter’s dry conditions. In this article, the influence of variable atmospheric absorption on snow cover extent (SCE) mapping was quantitatively investigated. A physical method was derived to enable atmospheric correction for variable atmospheric conditions. The total column precipitable water vapor (TPWV) from Moderate Resolution Imaging Spectroradiometer (MODIS) was parametrized into transmittances in this correction method. The corrected brightness temperature at 19 and 91.7 GHz from the Special Sensor Microwave Imager Sounder (SSMIS) was applied to the threshold algorithm for snow mapping over China. Compared with the Interactive Multisensor Snow and Ice Mapping System (IMS) data in winter from 2012 to 2013, for Qinghai–Tibet plateau (QTP), a significant improvement after correction was obtained from February to March over ephemeral and shallow snow, where the largest daily improvement of accuracy is up to 20%. The accuracy (incl. precision, recall, and F1 index) improved on average is from 0.77 (0.60, 0.68, and 0.63) to 0.79 (0.69, 0.7, and 0.68) over the full winter time from December to March. Over forest-rich Northeast China, where snow in winter is thicker, small improvement was observed at the onset of the snow season and over snow margin area. It was evidenced that high frequency is a promising way of snow cover mapping with the proposed atmospheric correction method.

We discussed atmospheric correction of passive microwave over China in the snow covered area. SSM/I brightness temperature, radiosonde observation database and MODIS water vapor product (MOD05) were used to estimate the atmospheric influence. The traditional gradient SWE algorithm before and after the atmospheric correction was compared, and results showed atmospheric influence was bigger over snow cover area. It is about 4-5K over the whole China. In the QTP (Qinghai-Tibet Plateau) area, the atmospheric influence was less than other areas in China; also QTP always had thin snow, so the traditional gradient SWE algorithm could not perform well. The gradient algorithm at high frequency was tested, and the results showed even under the original brightness temperature difference at 19 GHz and 85.5 GHz, the gradient (19 GHz-85.5 GHz) performed good consistency with IMS products.

Atmospheric transmittance is an important factor for atmospheric correction in the inversion of land surface parameters. Under nonprecipitating conditions, microwave atmospheric transmittance in the X-, Ku-, and Ka-bands is mainly determined by oxygen, water vapor, and cloud liquid water content. In this letter, radiosoundings from 119 stations in China, performed twice a day from January 2011 to July 2014, were used in the Salonen-Uppala cloud detection algorithm to distinguish cloud layers from layered atmospheric profiles and to estimate the cloud liquid water content therein. The resulting atmospheric transmittances at frequencies of the Advanced Microwave Scanning Radiometer-Earth Observing System over China were estimated using Liebe's millimeter-wave propagation model and Mie theory. Atmospheric transmittance maps were obtained by interpolating the results from individual sites, driving a climatological database for over China, which may be used to correct for atmospheric influence in surface parameter retrievals. The simulated transmittances were validated through a series of on-site field experiments in the North of China. We compared simulated atmospheric brightness temperature with measurements performed in situ using a ground-based radiometer system. The correlation coefficients between the measured and simulated values were 0.97, 0.99, and 0.98, in the X-, Ku-, and Ka-bands, respectively, with a root-mean-square error of 0.6, 1.6, and 9.5 K, respectively.

The seasonal snow over the Tibetan Plateau is an important parameter for the regional energy and water cycle. The snow there experience a different trend compared to that of the Northern Hemisphere in the past decades. Because of the shallow and patchy snow with a fast precipitation process and cloud occurrence, the traditional snow algorithm (18GHz and 27GHz gradient brightness temperature) experiences an insufficient accuracy there when employed the global coefficients. In this paper, the atmosphere condition was assumed to play a key role. Through the simulation, and the realistic atmospheric balloon measurement, the transmittance difference between Plateau and low land area was calculated, and the correspondent cloud influence was analyzed. As a result, we need to consider the cloud and water vapor effect, the method over the Tibetan Plateau needs to consider these two components, i.e. the existence of clouds over snow cover areas, and the atmospheric water vapor content, which typically separates high elevation regions from low lying landscapes.

Central Asia and Xinjiang, China are conjunct areas, located in the hinterland of the Eurasian continent, where the snowfall is an important supplement water resource. The induced seasonal snow cover is vita factors to the regional energy and water balance, remote sensing plays a key role in the snow mapping filed, while the daily remote sensing products are normally contaminated by the occurrence of cloud, that obviously obstacles the utility of snow cover mapping. In this paper, based on the daily snow product from Moderate Resolution Imaging Spectroradiometer (MODIS), a cloud removing method was developed by considering the regional snow distribution characteristics with latitude and altitude dependence respectively. In the end, the daily cloud free products was compared with the same period of eight days MODIS standard product, and validated with the help of the in-situ observations, revealing that the methodology are feasible with ~ 91% Recall, ~81% Precision and 85% Balanced F when snow depth is greater than 3cm. The cloud free snow products are kept as the same accuracy, recall and precision rate as the MODIS source products, while could provide higher temporal resolution, and more details over Center Asia and Xinjiang Province of China.

Lake ice is an important component of the hydrosphere and cryosphere, especially over high-lat itudes regions of High Mountain Asia (HMA), the long-term changes of freezing and thawing time can directly reflect the changes of climate. Remote sensing technology provides unreplaceable complement for HMA where the lake ice phenology is almost no monitoring, except that of Qinghai Lake. Considering the microwave signature's sensitivity to the snow and ice phase changes, a method based on the freezing and thawing probability were proposed. The long time series of Swath Brightness Temperature (TB) data from satellite-based passive microwave sensor AMSR-E (Advanced Microwave Scanning Radiometer for EOS) and AMSR2 (The Advanced Microwave Scanning Radiometer 2) were employed to retrieval the lake ice status of four lakes with different areas, Qinghai Lake, Siling Co, Hala Lake and Aqqikkol Lake. The method was validated with the result from higher resolution of daily cloud-free MODIS (Moderate Resolution Imaging Spectroradiometer) image, the result shows that, for four selected lakes, the R-Square of start of freeze-up(FUS) is 0.80466, end of freeze-up(FUE) is 0.92197, start of break-up(BUS) is 0.83882 and end of break-up(BUE) is 0.96127, respectively.

Three years radiosoundings measurement at China were inputted in the Salonen-Uppala cloud detection algorithm to distinguish cloud layers, and to estimate cloud liquid water content therein, then atmospheric transmittances of China at frequencies of the Advanced Microwave Scanning Radiometer - Earth Observing System (AMSR-E) were estimated using MPM and Mie theory. The validation expressed a good agreement with the field experimental observations. A comprehensive temporal and spatial analysis has been executed, appeared that microwave atmospheric transmittances were much dependence on water vapor distribution, and transmittances were relatively stable from October to March, but fluctuated obviously in summer. Transmittances of northwest China were higher than those of south China, and the highest transmittances were in Qinghai-Tibet Plateau region. The correlation coefficient between microwave atmospheric transmittance and MODIS total column precipitable water vapor (MYD05) in January at 18.7, 23.7, 36.5 and 89 GHz was 0.904, 0.923, 0.847 and 0.879 respectively. The relationship between monthly averaged microwave atmospheric transmittances and MYD05 could be used to estimate the high resolution microwave atmospheric transmittances products. The results prepared to ensure atmospheric effect in the applications of passive microwave remote sensing.

Passive microwave remote sensing has all-weather work capabilities, but atmospheric media have different influences on satellite microwave brightness temperature under different atmospheric conditions and environments. In order to clarify atmospheric influences on Advanced Microwave Scanning Radiometer-Earth Observing System (AMSR-E), atmospheric radiation were simulated based on AMSR-E configuration under clear sky and cloudy conditions, by using radiative transfer model and atmospheric conditions data. Results showed that atmospheric water vapor was the major factor for atmospheric radiation under clear sky condition. Atmospheric transmittances were almost above 0.98 at AMSR-E's low frequencies (<18.7GHz) and the microwave brightness temperature changes caused by atmosphere can be ignored in clear sky condition. Atmospheric transmittances at 36.5GHz and 89GHz were 0.896 and 0.756 respectively. The effects of atmospheric water vapor needed to be corrected when using microwave high-frequency channels to inverse land surface parameters in clear sky condition. But under cloud covered conditions, cloud liquid water was the key factor to cause atmospheric radiation. When sky was covered by typical stratus cloud, atmospheric transmittances at 10.7GHz, 18.7GHz and 36.5GHz were 0.942, 0.828 and 0.605 respectively. Comparing with the clear sky condition, the down-welling atmospheric radiation caused by cloud liquid water increased up to 75.365K at 36.5GHz. It showed that the atmospheric correction under clouds covered condition was the primary work to improve the accuracy of land surface parameters inversion of passive microwave remote sensing. The results also provided the basis for microwave atmospheric correction algorithm development. Finally, the atmospheric sounding data was utilized to calculate the atmospheric transmittance of Hailaer Region, Inner Mongolia province, China, in July 2013. The results indicated that atmospheric transmittances were close to 1 at C-band and X-band. 89GHz was greatly influenced by water vapor and its atmospheric transmittance was not more than 0.7. Atmospheric transmittances in Hailaer Region had a relatively stable value at low frequencies(<18.7GHz) in summer, but had about 0.1 fluctuations with the local water vapor changes at high frequencies.

Using the brightness temperature data measured by an Advanced Microwave Scanning Radiometer (AMSR-E) and the in-situ measurement, this paper compared the accuracy and applicability of five snow-depth retrieval algorithms (Chang algorithm, GSFC 96 algorithm, AMSR-E SWE algorithm, the improved algorithm for Qinghai-Tibet Plateau and atmosphere-correction algorithm) in the regions of Xinjiang, Qinghai-Tibet Plateau, Inner Mongolia, Northeast China, Northwest China and the North China Plain. Generally, over the whole China area, the results of the study show that, the improved algorithm for Qinghai-Tibet Plateau shows a relatively higher accuracy, with a root-mean-square error (RMSE) of 9.16, 9.96 and 9.63cm and a mean relative error (MRE) of 59.77%, 52.79% and 48.47%. According to regional verification respectively, the accuracy algorithm for Xinjiang region is the GSFC 96 algorithm, the RMSE of which is 6.85cm~7.48cm; For Inner Mongolia, it is the QTP algorithm, the RMSE of which is 5.9, 6.11 and 5.46cm; and For Northeast China, it is also the Qinghai-Tibet Plateau algorithm, the RMSE of which is from 6.21cm to 7.83cm. None of the five algorithms is in a good quality for Northwest China and the North China plain regions. For the Qinghai-Tibetan plateau region, the author is not able to obtain the results of statistical verification due to a lack of actually measured data.

Land Surface temperature (LST) is a critical parameter in water and energy circle supporting the global climate research. The traditional optical/thermal remote sensing estimates LST under clear-sky conditions, while the microwave remote sensing provides the all-weather LST estimation capability. In this paper, we derivate a new LST retrieval algorithm based on the emissivity parameterization using the Advanced Microwave Scanning Radiometer - Earth Observation (AMSR-E), considering the atmospheric influence, to tackle the LST retrieval under the non-scattering rain and cloud condition. The new derived LST algorithm has a physical basis and the RMSE is 1.8K with the station validation over boreal forest area at Sodänkylä, Finland.

The land microwave emissivity derived from satellite measurement is a function of surface cover structure, water content, surface roughness and the atmosphere composition. In this paper, the microwave emissivity derived by AMSR-E has been compared with the estimated soil moisture and microwave vegetation index (MVI), MODIS NDVI over Tibetan Plateau, which aims to understand the intrinsic relationship among satellite products physically. The microwave emissivities display the more consistent change with MVIs than NDVI over mixed forest cover area from time series analysis for summer and winter. While over grassland area, NDVI has the good position correlation with soil moisture, and the same to emissivity in summer time. Soil moisture derived the good negative correlation with microwave emissivity in arid and semiarid areas but poor in vegetation covered area. All of these are consistent with the expected results statistically which can provide one way to validate the microwave emissivity dataset.

Over the western China, due to the influence from shallow snow, the passive microwave remote sensing algorithm show its imbecility when using the gradient brightness temperature (Tb) algorithm of 36.5Ghz-18.7Ghz. In this paper, we employee several ground time-series snow depth dataset and the corresponding satellite brightness temperature to evaluate the 36.5Ghz/l 8.7Ghz and high frequencies' (85Ghz or 89.0Ghz/18.7Ghz) ability for the shallow snow retrieval. From the analysis, we can get that the high frequencies has its potential for the shallow, which is suit for the snow situation over the high land, even with the atmosphere influence at high frequency. The result tell that the relative deep snow (>; 20cm) the Tbs at 36-18GHz are more reliable than that of high frequency, while over the shallow snow especially <;20cm), the pair 36-18 is insensitive, but the high frequency pair (89/85-18GHz) shows its obvious possibility.

Snow extent and snow depth (SD) are critical parameters in metro-hydrological models and are sensitive to the global climate change. Over the western China, due to the influence from shallow snow, changing seasonal permafrost and the sparse observation stations, the passive microwave remote sensing algorithm show its applicability when using the gradient brightness temperature (Tb) algorithm of 36.5Ghz-18.7Ghz. In this work, we employ one whole-winter Tb extracted from Advanced Microwave Scanning Radiometer - Earth Observing System (AMSR-E) L2A swath dataset and the ground measurements of snow depth (SD) to analyse the snow microwave emission and gradient algorithm ability. The time series analysis shows that the Tb differences (36.5-18.7) and (36.5-10.7) are sensitive to relatively deep snow (>20cm), while the Tb differences (89.0-18.7) are sensitive to the occurrence of the new snow, with a promising correlation with shallow snow (<;15cm) and quickly decreasing (melting) snow depth, which suggest that a high frequency Tb difference could potentially be a good snow monitoring signal for the shallow snow cover over western China.

Land subsidence in Beijing is supposed to be caused by over-exploitation of ground water, which is leading to a rapid decline of water levels, drying out clay layers that finally result in land subsidence. The Interferometric Point Target Analysis (IPTA) is an advanced method to monitor vertical motion of the land surface over time. IPTA identifies backscattering objects, named as coherent points or points targets, at the ground surface that persistently reflect radar radiation emitted by the SAR antenna. The core component of the IPTA technique is the iterative estimation of phase differences for all measurement points over the sets of the SAR data using a linear model. In this paper, IPTA technique was used to retrieve the phase history, extract the linear deformation information from interferometry phase and weaken atmosphere phase delay in Beijing. 20 ENVISAT ASAR images acquired between June-18-2003 and March-14-2007 have been selected. The intention of this article is to demonstrate how IPTA technique could be used to extract valuable information in Beijing area.

In this paper, the methods, primary PS-like method and Stanford Method for PS (StaMPS) are both studied and used to monitor the subsidence in Tianjin area.

Modeling of snow emission at microwave frequencies is necessary in order to understand the complex relations between the emitted brightness temperature and snowpack characteristics such as density, grain size, moisture content, and vertical structure. Several empirical, semiempirical, and purely theoretical models for the prediction of snow emission properties have been developed in recent years. In this paper, we investigate the capability of one such model to simulate snow emission during the peak snow season-a new multilayer version of the Helsinki University of Technology (HUT) snow model. Developed with a single layer, the original HUT model was easily applied over large geographic areas for the estimation of snow cover characteristics by model inversion. A single homogenous layer, however, may not accurately allow the simulation of vertically structured natural snowpacks. The new modification to the model allows the simulation of emission from a snowpack with several snow or ice layers, with the individual component layers treated as in the original HUT model. The results of modeled snowpack emission, using both the original model and the new multilayer modification, are compared with reference measurements made using ground-based radiometers deployed in Finland and Canada. Detailedin situmeasurements of the snowpack are used to set the model inputs. We show that, in most cases, use of the multiple-layer model improves estimates for the higher frequencies tested, with up to 38% improvement in rms error. In some cases, however, the use of the multiple-layer model weakens model performance particularly at lower frequencies.

In the above titled paper (ibid., vol. 48, no. 7, pp. 2781-2794, Jul. 10), there is an error in Section II-B, which is corrected here.

Modeling of snow emission properties on microwave frequencies is necessary in order to understand the complex relations between the snowpack microwave emission and its characteristics, such as density, snow grain size, moisture content and snowpack vertical structure. With a reliable model, snowpack characteristics could be derived from passive microwave observations using model inversion, potentially improving retrieval accuracy of snow parameters when compared to traditional empirical inversion algorithms. In this study, we present a summary of recent activities aiming at experimental validation of the semi-empirical HUT snow emission model. The activities consist of comparisons of modeled brightness temperatures against tower-based and airborne reference radiometer data. Model inputs are derived form intensive field measurements of snowpack characteristics. Forward modeling of snow emission on the satellite scale, and a comparison with satellite observations is presented. Furthermore, a recent update of the model, enabling simulation of multiple snow layers and special cases such as snow covered lake ice, is experimented.

In satellite passive microwave measurements, the sky brightness temperature is a function of frequencies, sensitive to parameters such as water vapor content, liquid water (cloud and precipitation), oxygen, hydrometeors and atmospheric temperature. In order to investigate the atmospheric influence to the retrieval of snow parameters quantitatively, firstly, we combined the HUT (Helsinki University of Technology) snow emission model (except the atmosphere parameterization) and an atmosphere model to do theoretical simulation estimations. We indicate that the C and X band atmospheric influence could be ignored, while the atmosphere is a non-negligible absorber and emitter of microwave radiation at frequencies higher than 19 GHz. We also launched a 13-day experimental measurement in winter time over Sodankyla¿, Finland, with synchronous satellite (AMSR-E) and tower-based radiometer measurements, together with extensive in-situ atmospheric measurement dataset. The evaluation result indicates that the atmosphere plays a relative positive contribution (about 20K for 36.5GHz and 89.0/94.0GHz). The difference between satellite observation and point experiment comparison suggests conducting more physical model work with atmosphere contribution.

Finding the target contour information from a remote sensing image is one of the fundamental steps for image analysis. Conventional target contour extraction methods are usually based on the statistics information of the image. In this paper, using the maximum return value of the normalized scattering matrix derived from full-polarized Synthetic Aperture Radar (PolSAR), the relationship between the contour of targets and their corresponding dominate scattering type is preliminary researched. Then a novel target contour information extraction method based on the physical scattering mechanism of terrain targets is proposed, which is more effective and adaptable due to the scattering mechanism of terrain targets do not depend on the radar backscattering intensity, but its proportion among different polarizations. After applying to E-SAR airborne data, the results show that this method has a good capability to extract the target contour information.

The Moderate Resolution Imaging Spectra-radiometer (MODIS/Aqua) and the Advanced Microwave Scanning Radiometer - EOS (AMSR-E) are two sensors aboard on satellite Aqua. Atmospheric parameters retrieved from MODIS/Aqua, such as the layered atmosphere temperature, humidity and pressure profile and land surface temperature (LST), are used to help the estimation of AMSR-E instantaneous microwave emissivity in clear sky conditions over land. As an example, a two-week (from 12-08-2006 to 25-08-2006) instantaneous emissivity over land has been calculated globally for 6.9-, 10.7-, 23.8-, 36.5-, and 89.0-GHz using both polarizations, ascending and descending orbit respectively. The calculated AMSR-E emissivities agree well with the other study [6] through comparison, and can provide more details. The frequency dependency and correlation analysis show the promising emissivity prediction with different channels. The time series analysis over different land covers reveal that the variation of emissivities do not exceed 0.05 in average and it is almost zero for polarization difference (PD) change, which all induce that the time extrapolation of emissivities could tackle cloud contamination issues (time series gap).

It is very difficult to retrieve the land surface temperature (LST) from passive microwave remote sensing because a single multi-frequency thermal measurement with N bands owns n equations in N+1unknowns (N emissivities and LST) which is a typical ill-posed inversion problem. However, the emissivity is mainly influenced by dielectric constant which is a function of physical temperature, salinity, water content, soil texture, and other factors (the structure and types of vegetation). These make it very difficult to develop a general physical algorithm. This paper intends to utilize the multiple- sensor/resolution and neural network to retrieve land surface temperature from AMSR-E data. MODIS LST product is made as ground data which overcomes the difficulty of obtaining large scale land surface temperature data. The retrieval result and analysis indicate that the neural network can be used to accurately retrieve land surface temperature from AMSR-E data.

This paper demonstrates a study to the atmospheric influence on the passive microwave Brightness Temperature (BT) in Tibetan Plateau area at clear-sky condition. The absorption and emission of dry air and water vapor are considered as the main contribution of atmosphere at the fact of cloud-free. We choose the day of Dec. 07, 2005 as an example, and calculated the atmosphere absorption factor and effective atmospheric temperature which are based on an updated atmospheric microwave absorption model. With the help of MODIS-Aqua land surface products (MYD11_L2) and MODIS atmospheric profile (MOD07_L2) products, which can decide a real atmospheric status, a simplified radiative transfer equation (RTE) is employed to estimate the AMSR-E frequencies surface emissivity over Tibetan Plateau. As a result, the surface actual microwave brightness temperature is obtained through the product of retrieved emissivity and MODIS LST, it can be found that the atmospheric contribution to the brightness temperature add up to about 5.56K at 89.0GHz and average 0.54K at 23.8GHz somewhere Tibetan Plateau in the cloud-free winter days, and the space variation of atmospheric effect to microwave BT has been further discussed.