In order to understand the distribution of sources and sinks of CO2,based on ground\|based observations and Level 3 monthly gridded CO2 retrieval products by AIRS from January 2003 to December 2012,the spatiotemporal pattern and multi\|years variation characteristics of tropospheric CO2 during different seasons in China was analyzed.The results show that ①AIRS retrievedCO2 products and ground observations showed good consistency,correlation coefficient was above 0.85,monthly average deviation was within 3 ppmv.CO2 showed a growth trend over time,both showed consistent seasonal variation,the satellite retrieval products lagged slightly.CO2retrieved from GOSAT is smaller than AIRS and their correlation is relatively low.②CO2 concentration was higher in northern China and lower in southwestern China,the average month growth rate was 0.177;③CO2concentration showed an increasing trend in spring and decreasing trend in autumn.The higher CO2concentration appeared in northeast plain、Mongolia and Xinjiang,annual mean concentration in these corresponding regions reached 389 ppmv,and annual mean increase was 2 ppmv.In addition,CO2 concentration in autumn showed a significant decreasing trend in Mongolia and southwestern China.
To improve the accuracy of estimation of soil moisture under model parameters with uncertainties,this paper develops a soil moisture assimilation scheme based on state\|parameter estimation method,in which the dual ensemble Kalman filter(DEnKF)is integrated into Simple Biosphere model(version 2)(SiB2)to assimilate surface soil moisture observations to simultaneously optimize model parameters(soil texture and organic matter)and update model states(soil moisture in three soil layers).A series of numerical experiments are conducted to assess the performance of the proposed scheme based on obtained observations from Arou station in the upper reaches of Heihe river basin in 2008.Results showed that DEnKF can optimize parameters and state variables simultaneously.Compared to EnKF,DEnKF can obtain more accurate estimation of soil moisture in surface and root zone than EnKF,especially when observation data are scarce.The proposed soil moisture scheme is easy to realize and can correct model bias,so it is suitable for operational data assimilation systems on regional and global scales.
Although DEnKF assimilation scheme naturally does not introduce any observation sampling error,but small ensemble will still bring the spurious correlation into the background error covariance matrix and lead to the filter divergence.Toward reducing the side\|effects of small ensemble on the data assimilation,this paper designs an experiment of Lorenz96 model with DEnKF assimilation scheme,and analyzes the impacts of the covariance localization method and the covariance inflation method on the background error covariance matrix,gain matrix and data assimilation results.The experimental results show that the covariance localization method can eliminate the spurious correlation of the background error covariance matrix and the gain matrix,and it can also increase the rank of background error covariance matrix,which is helpful for the filter converging to the real solution.However,the covariance inflation method cannot remove the spurious correlation of the background error covariance matrix and the gain matrix,and it can only improve the systematic underestimation of the background error covariance in each data assimilation cycle.It is noteworthy that an appropriate localization radius and an inflation factors are critical to improve the data assimilation analysis results.
Empirical Mode Decomposition (EMD) is a method of image up\|scaling with adaptive characteristic.Almostly continuous spectrum of hyperspectral image data is the basic of information extraction,but the scaling transform can lead to spectal distort.Thus it is important to analyse the spectral fidelity of hyperspectral image after EMD up\|scaling transform.CHRIS hyperspectral remote sensing image is taken as experiment data.The Correction Coefficient(CC),Error,Relative Error(RE) and Spectral Angle Mapper(SAM) are used to evaluate the spectral fidelity of image and the typical ground objects.The results of EMD and Wavelat transform are compared.The results show that:①The level 1\|10 up\|scaling images have well spectral fidelity,average spectral correlation coefficient is more than 0.979,error is less than 55,reletive error is less than 0.036,and spectral angle mapper is within 0.041.②With increaing of up\|scaling,the transform images spectral fidelity is slowly declining,and the spectral distortion of former 4 levels is more obvious than the others.③The spectral fidelity of seven typical groud objects which is chosed from image are well.Phragmites and River spectral are preserved best,but the Aquaculture water specral matching is not good.④Compared with Wavelet trandform on image spectral fidelity,the method of EMD is raletivly stability and has less spectral distortion with up\|scaling increaing.
Considering urgent request of buildings information for urban development and planning,updating of geographical situation information system,digital city and military reconnaissance,Semi\|supervised Discriminant Analysis (SDA) algorithm was employed for extraction of building areas in high resolution SAR image to improve the ability to target recognition and achieve fast extraction of building areas in city district.Radarsat\|2 images and TerraSAR\|X images were used as the experimental data.A variety of texture features of images were calculated based on gray level co\|occurrence matrix,and then the SDA algorithm was employed for feature extraction.The new feature was used as input of Otsu method to extract building areas.Finally,the post\|processing of image classification was performed.A comparison of the recognition result of building areas between SDA algorithm,Linear Discriminant Analysis (LDA) algorithm and Local Preserving Projection (LPP) algorithm was made.It is concluded that SDA algorithm has stronger generalization ability;SDA algorithm is applicable to feature extraction of high resolution SAR image under less prior class information,quickly and effectively extract the information of building areas in city district.
Grid computing as a distributed computing in the field of scientific computing provides a powerful computation force for Earth observation data processing.based on the analysis of data transmission and load balance in grid computing,the paper puts forward a data compression method based on Run Length Encoded(RLE) and Huffman Encoded and a task allocation strategy considering the "computing end member".This combined method can effectively solve the problems of network data transmission by irregular extraction and the task load balancing.On Remote Sensing information service Grid Node(RSSN) computing platforms,we take the Aerosol Optical Depth(AOD) retrieval in China area of 1 km resolution for example to validate and analyze the effectiveness of this method.
Compared with the impulse system,LFMCW SAR has obvious superiority in Miniaturization aspect.To satisfy the actual demand of radar miniaturization in motion platform such as unmanned aerial vehicle,this study designs a small size,low cost and low power miniature SAR system.This paper analyses the basic principle of LFMCW radar system’s high resolution imaging,then elaborates its design idea and key technology.Imaging results of flight test onboard are given in the end,which verifies the effectiveness and feasibility of this radar system.
The sensors with wide spatial coverage and high observation frequency are usually designed to have large field of view (FOV).However,the large FOV also results in some problems associated with the bidirectional effects due to sun/view angle differences.On the basis of kernel\|driven semiempirical BRDF model,a new method was proposed to normalize sun\|target\|sensor geometry of remote sensing image.The polynomial function of vegetation index was used as the coefficients of volume\|scatter kernel and surface scatter kernel in the kernel\|driven semiempirical BRDF model.Multi\|angle observation data acquired at difference dates was used to derive model coefficients which were applied to sun\|target\|sensor geometry normalization.Simulated data and observed satellite data were used to evaluated the proposed method.The experimental results show that the proposed method yielded good fits between the observed and estimated values,and indicate that this method is capable of normalizing satellite data to a standard sun\|target\|sensor geometry.
In order to analyze the urban building density from fully polarimetric synthetic aperture radar images,a novel method based on the polarization feature co\|occurrence matrix is proposed in this paper.Firstly,we introduce the selected polarization feature into co\|occurrence matrix,which is effective because it not only considers the polarization scattering mechanism but also introduces spatial arrangement information of building areas.Then,considering the local regional characteristics of building density,the histogram statistics feature is formed by K\|means unsupervised clustering based on the co\|occurrence matrix features in the certain image block.Finally,the histogram statistics feature vector quantization is adopted to realize grades of urban building density from fully polarimetric SAR image.The effectivity of the proposed urban building density analysis method is demonstrated by processing RadarSat\|2 fully polarimetric SAR images,experiments results show that the proposed method applies to both building orientation complex urban areas and the neat rows urban areas.
Due to the large searching space,the time complexity of traditional Ant Colony Algorithm (ACA) is very high.Thus,the ACA-based band selection algorithms require a long time to run,and always suffer from local optima.In comparison,the Polymorphic Ant Colony Algorithm (PACA) can significantly decrease the searching space and thus the time complexity.In considerations of this,this paper designed a PACA-based band selection algorithm (PACA-BS) for hyperspectral remote sensing imagery.Performance evaluation of algorithms was focused on the following aspects:computing time,separability of band sets,information amount and overall accuracy.Herein,Hyperion and AVRIS imagery were employed as data source.The results showed that the computing time of PACA-BS was markedly lower than ACA-BS.Furthermore,band sets derived by both algorithms possess similar separability,however,the band sets of PACA-BS ave larger information amount,and thus generate a higher overall classification accuracy.The PACA\|BS is thus proved to be a promising and optimized method for band selection of hyperspectral image.
The paper introduced gaofen-1(GF-1) satellite images management and web service in Surveying and Mapping,and try to use PostGIS database technique in management of massive remote sensed imagery.It was implemented in Surveying and Mapping task,including metadata based image management,task\|oriented massive image storage,retrieval,browse,and product management.Auxiliary data was applied to schedule and quality management.The system providing web service through a combination of PostGIS database and tile map database.It expands the scope of Surveying and Mapping web services.
Aerosol Optical Depth is an important index in atmosphere monitoring.The distribution of AOD has a great practical significance for pollution control.Based on deep blue algorithm,AOD were retrieved from HJ-1/CCD over the Yangtze River Delta and validated by other AOD products.The retrieved results from HJ-1A and HJ-1B have good correlations with MODIS AOD products,and the retrieved values are generally higher.Compared with ground-based measurements from AERONET,the differences were between 0.008 and 0.364.Due to a lack of corresponding AERONET data,rigorous statistical analysis of the differences between them were ignored.Although there were systematic bias between the retrieved AOD and MODIS AOD products,monitoring AOD with HJ-1/CCD using the deep blue algorithm has a good application prospect.The retrieved data can reflect the distribution of AOD over Yangtze River Delta very well,which has a higher spatial resolution than MODIS AOD products.
With carbon monoxide data on surface and 500 hPa pressure surface in 2012 from Atmospheric Infrared Sounder(AIRS) observation,the spatial and temporal distribution of carbon monoxide in northeast China was investigated.The effects of domestic biomass burning and carbon monoxide advection from surrounding countries and regions on carbon monoxide in northeast of China were analyzed with fire data of MODIS integrated from Terra and Aqua and meteorological data from NCEP/NCAR. Results showed that carbon monoxide volume mixing ratio at surface was heavy during winter and spring,and was light in summer and autumn due to domestic Warm Supply,the seasonal differences of OH free radicals concentration,population density and industrial distribution.Carbon monoxide volume mixing ratio was heavy in the regions with high population density and industrial concentration district and vice versa.Meteorological data from NCEP/NCAR showed that there existed a transport path from Siberia and Mongolia towards northeast of China during most of months.Biomass burning could reflect the harvest time of crops and therefore the spatial and temporal distribution of crops could be reflected by the distribution of carbon monoxide.In this paper,we studied the effects from biomass burning on carbon monoxide concentration in northeast of China,which would be of great significance for environmental management.
Disorderly mining for rare earth ore not only wastes many rare earth resources,but also deteriorates the ecological environment in the mining areas and their surroundings.Taking Ganzhou city in Jiangxi province as an example,a set of feasible evaluation methodology for the vegetation Net Primary Productivity (NPP) loss resulted from the mining of rare earth ore was established.This methodology makes full use of the advantages of the remote sensing data at three levels of high,medium,and low spatial resolution.The key technologies in this methodology include the extraction of the reference area,the delimitation of the damaged vegetation range and the downscaling for the NPP data at low spatial resolution.based on this methodology,the vegetation NPP loss in Ganzhou in 2013 was assessed.The results showed that:①By 2013,the area of direct and indirect damage for vegetation caused by the mining of rare earth ore was 31.74 and 44.48 km2,respectively.The vegetation's indirect damage area presented an exponential decrease(R2=0.96,P<0.01) with the increase in the distance from the mining area.②The NPP loss resulted from the mining of rare earth ore was 3.87×1010 gC in 2013.The direct and indirect loss accounted for 77.81% and 22.89% of the total loss,respectively.These results indicated that the indirect loss should not be ignored in the ecological damage assessment.The methodology for the vegetation NPP damage assessment can provide a new probe to resolve similar problems in other mining area.The results can provide a basis for the ecological assessment,the pricing of the rare earth and the ecological environment management of mining area.
Ozone profiles from surface to about 60 km are retrieved from GOME\|2 ultraviolet spectrometer,using GOME and OMI ozone retrieval algorithm.This study got tropospheric ozone products from these GOME\|2 ozone profiles.To evaluate the performance of GOME\|2’s tropospheric ozone data.This paper anlyzes tropospheric ozone products and characterization of ozone profile retrievals from GOME\|2 which are carried out:Statistical analyses showed that the tropospheric random\|noise errors are generally within 10% over China.GOME\|2 products were also compared with ozone products from Tropospheric Emission Spectrometer(TES) which applies thermal infrared observations.Statistical bias determination with GOME\|2 and TES showed that the monthly mean tropospheric column ozone of GOME\|2 are 5%~20% lower than TES in the tropics and southern middle latitudes;Tropospheric averaging kernels showed that GOME\|2 retrieved ozone values had almost the same sensitivity to ozone changes as TES under a clear sky;The GOME\|2 and TES tropospheric average kernels also showed that clouds screened both radiance signals below cloud top,while GOME\|2 had a stronger signal above cloud top as clouds enhance the solar backscatter of ultraviolet.
Satellite remote sensing provides a promising ways to estimate regional evapotranspiration (ET) in a spatially distributed manner.In this study,an improved Evapotranspiration Model for Land (EML) is proposed based on the trapezoid framework of VFC/LST space.In EML,a VFC/LST trapezoid space is theoretically defined for every pixel and then ET is estimated based on the Water Deficit Index (WDI).The proposed model was applied to the Soil Moisture\|Atmosphere Coupling Experiment (SMACEX) site.Regional scale evaluation with remotely sensed data set from Landsat 7 ETM+ was carried out to assess the performance of ETEML.The evaluation shows that EML is capable of estimating reliable surface heat fluxes.In this study,the Mean Absolute Difference (MAD) and root mean square deviation (RMSD) of Latent heat Flux (LE) estimates based on EML are 62.20 and 74.17 W/m2,respectively.The MAD and RMSD of sensible heat flux (H) estimates from ETEML are 43.37 W/m2 and 49.02W/m2,respectively.A further inter\|comparison between the Trapezoid Interpolation Model (TIM) and EML is undertaken and the results indicate EML eliminated the subjectivity and uncertainties related with TIM.Overall,results suggest the EML is capable of estimating reliable surface heat fluxes and greatly promote the application of the trapezoid framework\|based ET modeling approaches to heterogeneous surfaces.
The series of NASA algorithms (Chang algorithm,NASA 96 algorithm and Foster algorithm)are the simple,practical empirical algorithms of passive microwave remote sensing for snow depth and snow water equivalent inversion.A wide range of algorithm verification and improvement has been put forward by many scholars.In order to evaluate the applicability of the Series of NASA algorithms in Northeast China on space and time further,In this paper,we select a 10km×10km mixed pixel of passive microwave remote sensing in which farmland and forest as the main parts in Changchun Jingyuetan area.Continuous observation of snow parameters and meteorological data on time through the entire dry snow season (December 2014 to the next February)have been done,and combined with FY3B\|MWRI light temperature during this period.Valuating and analyzing the accuracy of the NASA series algorithm,and using the binary tree method to determine the snow cover type of the experimental area.The results show that Chang algorithm and NASA 96 algorithm perform better in snow depth Inversion in the former half of the period.With time going,the trend of overestimate becomes more obvious.Considering the influence of forest coverage,the inversion precision of NASA 96 algorithm is higher.Their largest overestimated value is 24.46 and 14.62 cm respectively.This may caused by changing snow properties,especially particle size increasing during the dry snow season.Foster algorithm can seriously overvalued snow water equivalent,the snow type classification system may not be suitable for Northeast China.In this article,the continuous observation data of snow lay the foundation for understanding the properties of snow in northeast region.Time series verification and analysis towards algorithms provide a reliable basis for the further improvement of snow parameter inversion algorithms.
High-resolution optical/ infrared data can be utilized to improve the low spatial resolution of microwave soil moisture data using downscaling methods.The Visible and Shortwave Infrared Drought Index (VSDI) derived from the information of shortwave infrared (SWIR),red and blue channels is regarded as an effective surface wetness indicator.A downscaling approach for AMSR\|E soil moisture based on VSDI is proposed in this study.Firstly,the VSDI was calculated based on SPOT\|VGT images with the spatial scale of 1 km.Secondly,the VSDI images were aggregated into AMSR\|E soil moisture data scale (25 km).Thirdly,the correlations between aggregated VSDI and AMSR\|E soil moisture at 25 km spatial scale were explored.It was found that S\|curve fitting function could provide insight into the relationship and the fitted residuals exhibited characteristics of randomly distributed in space.Finally,the predicted soil moisture and the fitted residuals at 1km spatial scale based on S\|curve model were used to downscale AMSR\|E soil moisture product.The downscaled soil moisture was validated by in situ observations and AMSR\|E soil moisture.The results suggest that the proposed downscaling method with good precision can be used to improve the spatial resolution of microwave soil moisture data.
Vegetation\|patches pattern is very popular in various ecosystems all over the world,such as arid,semiarid and coastal tidal flat environments.In the recent year,researches on patchy vegetation formation,temporal and spatial pattern successions have been paid more and more attentions.One of them was focused on patchy vegetation distribution mapping and temporal and spatial dynamics which was basis for studying the vegetation community succession machines and was a key indicator of long\|term vegetation change in the ecosystems.Patch vegetation community succession in the Yellow River Delta belongs to an allogenic primary succession.With studying this succession,it is easier to understand the law of the birth and development of vegetation under the natural environment and to discover the adaptation mechanism of the vegetation.This study was conducted in the Modern Yellow River Delta(MYRD).According to the visible degrees of quasi\|circular patches,the bare spot regions,quasi\|circular vegetation patch regions and hidden quasi\|circular vegetation patch regions were identified and derived from 1996,2005,2007,2010,2012 images of the MYRD by interactive visual interpretation and analysis,then the temporal and spatial dynamics of patchy vegetation were analyzed.The results from this paper indicated that:1)there were obvious gradients in space and dynamics of vegetation succession in the modern Yellow River Delta,and Starting in the sea,the patchy vegetation types are sequentially sea water,tidal flats,the bare spot regions,quasi\|circular vegetation patch regions,the hidden quasi\|circular vegetation patch regions along the elevation gradient;2)the diameter,vegetation cover percentage,average height of patchy vegetation and total salt content in soil could be considered as the intuitive indices for distinguishing between the new and the old patchy vegetation;3)The multispectral images with 10 m and 5~6 m spatial resolution were enough to classify these three types of quasi\|circular patchy vegetations,but still unsuitable for estimation of the area of patchy vegetation which would be overcome by the images with better than 1 m spatial resolution.The results from the present researches could provide a reference for deeply studying spatial pattern and succession machine of patchy vegetations in the MYRD in the future.
Accurate mapping health levels of Robinia pseudoacacia forests is the premise of forest health assessment and ecological restoration.Based on Multiple spectral bands,gray\|level co\|occurrence matrix (GLCM) texture information,and Local Getis\|Ord Gi statistical information extracted from High resolution IKONOS imagery,random forest classification method was used to identify and map health levels of Robinia pseudoacacia forests.Random forest method was applied to six combinations of the spectral,textural,and spatial features,and the contribution of predictive variables was ranked.The experimental results indicated that 19 m × 19 m is the best moving window size for GLCM texture information extraction,GLCM mean calculated from IKONOS Pan band which is the best textural feature;local Getis\|Ord Gi statistical information calculated from IKONOS band4 which has the most important role.Compared with the combination of all spectral,textural,and spatial features (total of 80 features),the best model based on random forests with a forward variable selection process which selected only16 variables of the original 80 variables and obtain the best predictive accuracy (overall accuracy of 93.14%,kappa coefficient of 0.894).Our results indicate that the spatial features extracted from high resolution imagery can improve classification accuracy of health levels of planted forests with a regular spatial pattern.Forecast sorting can utilize fewer predictors and obtain higher classification accuracy.
A very important and urgent problem for prospecting in Fangshankou of Beishan is to look for new ore body in an old mine.In this paper,the ASTER data were used to extract alteration mineral which containing Fe3+,Al-OH,Mg-OH,CO3-2-or Quartz,and then the results is verified by field inspection.By analyzing the remote sensing anomaly characteristics of a typical mineral deposit,the relations between remote sensing anomaly and mineralization were analyzed,then the remote sensing prospecting prediction areas were delineate.The results indicate that the accuracy of five kinds of alteration minerals are more than 80%,and these remote sensing anomaly results are in good agreement with geophysical and geochemical anomalies.Moreover,through comprehensive analysis and verification,this study considers that there is no correlation between a single alteration mineral which was caused by regional metamorphism or by rock itself and mineralization,but the alteration mineral assemblage exists in a symbiotic mode is often related to mineralization,such as the alteration mineral assemblage(Fe3++ Al-OH+ Mg-OH) in this paper has important instruction significance for magmatic hydrothermal mineralization (especially for gold mineralization).In addition,based on the similar analogy theory,we marked out five prospecting area of remote sensing,and some useful mineralization clues were found in those area by our field verification.So the study results will bring new inspiration for the geological prospecting work in research area.
The paper selected 50km buffers of the middle and lower reaches in the Ussuri River as the study area.Using the object\|oriented classification method extracted the wetland information from remote sensing images.Supported by GIS technology,the paper selected Total Class Area(CA),Percentage of Landscape(PLAND),Largest Patch Index(LPI),Patch Area Mean(AREA\|MN),Number of Patches(NP),Patch Density(PD)et al.to analyze the dynamic characteristics of landscape metrics during 1989\|2013.Results indicated that the wetland information could be accurately extracted by object\|oriented classification method,and the "salt and pepper" was avoided;During 1989\|2013 period,wetland of the region changed dramatically,and the main characteristics of change was that paddy fields,reservoirs/ponds,canals/ditches,shrub swamps increased and rivers,lakes,forest marshes,herbaceous marshes reduced;Herbaceous marshes significantly reduced and the most were converted into paddy fields;The main wetland type was herbaceous marshes in 1989,and the paddy fields was over herbaceous marshes of the area by 2013;The dominating regions of wetland changes were distributed in Naoli River basin and Muling River basin in China.Fragmentation of natural wetland landscape gradually deepened,constructed wetlands tended to gather.Human disturbance is the main driving forces of wetlands landscape pattern changes.Difference of wetlands landscape change was significant in the territory in China and Russia.Because a lot of herbaceous marshes transformed into paddy fields,natural wetlands lost seriously and constructed wetlands increased significantly in China and Russia.The changes of wetlands were insignificant,and natural wetlands was preserved better in Russia.Farmland cultivation was a major cause of differences of wetland landscape fragmentation between China and Russia.
Using the 7 years remote sensing data（MSS data of 1974;TM data of 1990,2000;ETM+ data of 2006,2009,2012;OLI data of 2014）calculate five indicators:LUCC,composite vegetation index,water and soil loss,land desertification ,soil moisture rate and build land degradation to establish evaluate model of Fuxian Lake basin.Making some studies of land degradation evaluation of Fuxian Lake Basin:using the Analytic Hierarchy Process (AHP) construct judgment matrixes of the five indicators and calculate weights of the five degradation indicators from judgment matrixes by hierarchical analysis software;land degradation dynamic Monitoring and evaluation based on GIS spatial analysis methods,finally put forward the comprehensive measures to prevent and solve land degradation.The result shows that:during 1974~1990,the area of land degradation of Fuxian Lake Basin in creased sharply,during 2000~2006 the area of land degradation declined slightly;between 2009 and 2012,the total area of land degradation dropped significantly,the area of land degradation was 200.27 km2,which was the lowest in history;during 2012~2014 the degradation area increased slightly.
Remote sensing is an effective approach to monitoring near\|surface atmospheric particulate matter.In this paper,the vertical,humidity and wind\|speed correction were conducted for the MODIS (moderate resolution imaging spectrometer) AOD (aerosol optical depth) data over the Wuxi city by the vertical distribution of extinction coefficients obtained from laser radar,surface relative humidity data and wind\|speed data.Validation was performed using PM10 and PM2.5 concentration monitoring data in 7 ground sites over the study area.The results showed that there were good correlations between the corrected MODIS AOD remote sensing products and the ground monitoring data.The coefficients of determination between corrected MODIS AOD and PM10 was 0.452,and between corrected MODIS AOD and PM2.5 was 0.449.These results suggested that corrected MODIS AOD data should be used to monitor air quality over the Wuxi city.The results also showed that the vertical correction results using the laser radar data were better than those using the visibility data.In addition,the correlations between corrected MODIS AOD and ground monitoring data were highest in summer,followed in autumn and spring.The corrections were lowest in winter.
