2015 Vol. 40, No. 11
Weddell sea anomaly, appeared in west Antarctica, is an ionospheric anomaly. In this paper ground-based, GPS data and altimetry satellite Jason-2 data in west Antarctica are used to extract TEC in continental and oceanic areas respectively. The advantages of ground-based GPS are to obtain TEC time series over the station. TEC results of satellite altimetry and ground-based GPS are at the same precision level. Although the time resolution of satellite altimetry is lower, the advantages of satellite altimetry are to be observed in ocean areas. The two observations are complementary, which is valuable to depict the comprehensive characteristics of WSA in west Antarctic together. In view of space, WSA appeared mainly in Bellingshausen Sea, and also covered Bellingshausen sea and the Weddell sea, as well as up to 80° S of the west Antarctic continent. In view of time, the period of the appearance is from Oct. to Mar. next year, with the characteristics of low electron density in daytime but high electron density in nighttime, and the abnormal degree become larger in high solar activity.
The Grove Mountains area is a key study area in Chinese PANDA section of East Antarctica. Envisat ASAR images of 35 days interval were utilized to map the ice velocity in the Grove Mountains with DInSAR and offset-tracking methods. An ice velocity map of Grove Mountains displays complex motion features influenced by the widespread distribution of rocks and nunataks. We found two large ice streams bounding the Grove Mountains area, with a maximum velocity of 40 m/a. Some small ice streams flow between the nunataks in the area. GPS measurements and ice velocity MEaSUREs were applied to evaluate the accuracy, confirming the reliability of our results. Comparing winter ice velocities for the years 2006, 2007 and 2009, we detected no obvious interannual ice velocity changes in the Grove Mountains.
The Antarctic ice sheet plays a major role in studies of global climate change and sea-level rise, and digital elevation models (DEMs) are of great importance in geoscience studies of Antarctica. The interpolation method is one of the key factors to the accuracy of DEM. Five interpolation methods including inverse distance weighted, Kriging, radial basis functions, local polynomial, and Nearest Neighbor methods were used to develop DEMs for the Dome A area and the transect from Zhongshan station to Dome A using CryoSat-2 satellite altimetry data. A cross validation method was applied to evaluate the accuracy;the results indicate that Kriging method is the best of the five methods. Two DEMs for the two areas were generated by the Kriging interpolation method with CryoSat-2 data. The accuracy of two DEMs was assessed by GPS data from the 21st Chinese National Antarctic Research Expedition (CHINARE). The results show that the DEM of Dome A area has higher reliability with a bias of 1.248 m and standard deviation of 0.51 m. While in the steeper area in the Zhongshan station to Dome A transect, the accuracy of the DEM was low, the bias was 3.87 m and the standard deviation was 9.358 m.
In the study of atmospheric inversion based on radio occultation, the residual ionospheric errors have great influence on the accuracy and the second order errors are not ignorable. In this paper, MSIS90 atmospheric model, 3D NeUoG ionospheric model and IGRF11 geomagnetic model are all collected, and 3D ray tracing method was used to simulate the changes of the second order residual ionospheric errors in radio occultation. Under different solar activities, local time, azimuth, the second order errors were analyzed, especially for its global distribution. The results show that second order errors are usually in sub-centimeter level, but up to centimeter level under higher solar activity or in middle and low latitudes at 0° or 180° azimuth, and illustrate the structure of “three-peak” in globe. Therefore, in the occultation data process, the ionospheric monitoring must be strengthened and second order ionospheric errors must be eliminated effectively.
Radio occultation can provide the ionospheric electron density profile from the ground to the height of LEO orbit, which play an important role in the research of topside ionosphere. In this paper, COSMIC (Constellation Observing System for Meteorology Ionosphere and Climate) radio occultation data are used to obtain the electron density profile, and the ionospheric parameters, such as hmF2 (peak height of F2 layer), NmF2 (peak electron density of F2 layer), VSH (vertical scale height), are extracted. The changes of F2 layer in Antarctica are studied in the solar activity cycle, annual cycle and diurnal cycle, respectively. The vertical structure characteristics of top ionosphere in Antarctica, especially for Weddell Sea anomaly, are studied in detail. The results show that the daily averages of the whole Antarctic hmF2, NmF2 and VSH are 250 km, (1-4)× 1011el/m3, 100~200 km, respectively, during low solar activity, and up to 250-300 km, (1-8)× 1011el/m3, 150-250 km, respectively, during high solar activity. Weddell Sea Anomaly is mainly characterized by the electron density enhancement at the top and the electron density reduction at the bottom, which lasts four months, from late spring to early autumn.
In this paper, global sea level variations and the influence on the sea level rise from Antarctic ice sheet and Greenland ice sheet ablation from 2003 to 2012 are estimated from satellite altimetry, GRACE and oceanographic data. From 2003 to 2012, global sea level rose with the rate at 2.72±0.07 mm/a and spatial distribution characteristics estimated from altimetry. Ocean mass changes estimated from GRACE accounted for 66% of the sea level change, with the rate at 1.80±0.10mm/a, and steric sea level changes accounted for 34% of the sea level change with the rate at 0.93±0.05 mm/a estimated from Argo and accounted for 19% of the sea level change with the rate at 0.51±0.03 mm/a estimated from Ishii. Meanwhile, global sea level, steric sea level and oceanic mass variations had significant seasonal characteristics, with the annual cycle amplitude 4.6±0.3 mm, 4.1±0.2 mm, 3.6±0.1 mm and 7.3±0.4 mm, respectively, estimated from altimetry, Argo, Ishii and GRACE. Using empirical orthogonal function (EOF) analysis, we found significant seasonal differences existed between the northern and southern hemispheres in global sea level, steric sea level, but did not exist in oceanic mass changes. The rates of Antarctica ice sheet and Greenland ice sheet ablation were -75.7±12.3 Gt/a and -124.1±2.9 Gt/a from 2003 to 2012, respectively. And the correspondent global sea level rising rates were 0.21±0.03 mm/a, 0.34±0.01 mm/a, which can only accounted for 12% and 19%, total up to only 31% of the ocean mass increase. So the melting of the polar ice sheet was not the most important factor in the increase of ocean mass from 2003 to 2012.
This study made comprehensive comparison of sea ice concentration derived from ship observations, optical and passive microwave images. Seven types of SIC derived from AMSR-E images and MODIS images were compared with RSO at 27 sample points between 2002 and 2007. The correlation between SIC derived from synchronous data observed at the same spatial scale had higher correlation coefficients than the others, whereas SICs observed at different times and scales displayed large inconsistencies. The correlation between SIC derived from data observed at the same spatial scale but at different times were affected by the spatial resolution. The RSO had considerably large differences with the PSO, the subjectivity in e visual estimation, weather conditions, quality of the image classification, and extraction method all attribute to their differences. Although the pseudo ship observation extraction method is helpful to generate large number of samples at the sea ice edges, error propagation must be controlled in the extraction process.
Sea ice thickness is an important parameter and indicator of climate change, sensitive to ecosystem in Polar Regions. For accurate forecasting of climate change, sea ice mass balance, ocean circulation and sea-atmosphere interactions, it is required to have long term records of sea ice thickness. Satellite altimetry provide useful technology for obtaining time series sea ice thickness information on hemispheric scale. So far, four mainstream algorithms based on satellite altimeter data have been used to estimate sea ice thickness effectively. Different algorithms and parameter values as selected by different researchers will lead to results unsuitable for comparison and large uncertainties. To solve this problem, this study compared sea ice thickness over the Arctic research area retrieved from these four mainstream algorithms based on CryoSat-2 satellite altimeter data. Our results demonstrate that: (1) when compared to each other, the retrieved mean sea ice thickness estimated from these four algorithms have similar spatial distribution, but with larger difference in the value, up to 0.476 m;(2) the sequence of estimated sea ice thickness for the Arctic spatial average based on four algorithms is Laxon03 algorithm, Yi11 algorithm, Laxon13 algorithm and Kurtz09 algorithm;(3) the difference of the mean sea ice thickness from these four algorithms was higher in Beaufort Sea than in Central Arctic or Greenland Sea;(4) The Laxon13 algorithm was the optimal algorithm, with the minimum bias and RMSE when compared to IceBridge sea ice thickness measurements. These results can provide useful reference and basis for further study to improve algorithms, so as to quantify dynamic changes of sea ice thickness more accurately.
Sea ice is the most active and unstable component in polar region that its change is significant for the investigation of regional and global climate and environmental change. In this research, the Arctic sea ice and new ice extent from 2004 to 2012 were extracted and their time series changes were analyzed by using multi-source remotely sensed data. The results indicate: there was an inverse relationship between the sea ice extent changes and the near surface air temperature variations;the new ice extent changes were inversely proportional to the sea ice extent changes;the regional new ice extent changes and the near surface air temperature variations had a close relationship.
Ocean tide loading(OTL), which is usually computed from an ocean tide model, is a factor that cannot be ignored in high accuracy GPS data processing. Due to the lack of highly accurate altimetry data and the presence of sea ice, the polar region is one of the regions where ocean tide model has the lowest accuracy in the world. To show the intermodel discrepancies on polar regions and the effect of OTL on GPS baseline solutions, the OTL computed by seven different models (FES2004, DTU10, ETO11a, GOT4.7, HAMTIDE11a, OSU12, and TPXO.7.2) at Zhongshan and six other IGS stations in the coastal area of East Antarctica were calculated, and the effects of these seven models on GPS baseline solutions were compared using GAMIT software. The results show that, the effect of OTL on GPS baseline solutions was subjected to the station locations and baseline directions, and can reach centimeter level and 10-8 for absolute and relative accuracy, respectively. The effects of OTL must be considered during high accuracy GPS data processing. The effects of OTLs computed from different ocean tide models on GPS baseline solutions for the coastal area of East Antarctica are not significant and can be disregarded.
Pseudo-stochastic pulse can compensate for the deficiencies of a dynamic model in precise orbit determination. Because of the great number of pseudo-stochastic parameters, the efficiency of the conventional least squares estimation may be depressed. To resolve this problem, a method based on the theory of parameter pre-elimination and back substitution, is presented. This proposed method was validated by experiments, and the result shows that parameter estimation using this method is the same as that of the conventional least squares, but with improved efficiency.
This paper introduces a method to retrieve the vertical and the horizontal wavelength, the potential energy and the momentum flux of the gravity waves in the atmosphere from single temperature profiles. Using GPS RO dry temperature profiles from the COSMIC mission, we obtain the global distributions of the gravity wave parameters during the summer and winter of 2007 to 2008, regarded as the characteristics of the mesoscale atomosphere gravity wave activities. Our results show that the distribution form of the gravity wave is in good agreement with those found in the latest references. Our study shows that at the altitudes of 20~30 km, the atmospheric gravity wave activities are apparently stronger in NH winters than in NH summers. Gravity wave activities have zonal distribution characteristics. In the high latitudes of the hemisphere during summer, gravity wave activities have significant seasonal variations. Gravity wave activities are affected by topography and convections. The distributions of the momentum flux and potential energy derived from COSMIC RO data show similar characteristics.
Three methods including offset centre of gravity retracker(OCOG), threshold method and Brown-Hayne model are evaluated in this paper. Retracking experiments were conducted for Bohai, Yellow Sea, East China Sea, and South China Sea areas using Jason-2 20 Hz Sensor geophysical data recorder(SGDR). The results show that the iterative least square method has an accuracy of 2 cm, comparing favorably to the maximum likelihood estimator (MLE) method issued by CNES in the four sea area. A 50% threshold value yields superior performance than other values applied to the China seas area with an accuracy of 3~6 cm and bias of 4~10 cm in contrast to the MLE method. The OCOG method delivers less accurate results but this may be due to the initial value in the Brown-Hayne model. The correction quantity with respect to the normal tracking points increases when close to continents and islands, especially in the South China Sea, as the biggest correction is 1m;radar waveforms are affected when satellite tracks cross the Xisha archipelago.
An exact and simple method for measuring distance on polar stereographic nautical charts used for polar navigation is investigated. Equations for thumb lines and great-circle lines were deduced;shape analysis results indicate that great-circle distance is preferable for distance measuring in polar navigation. Formulas for calculating the distance of straight-line segments on polar stereographic charts based on their geometry are given as distance difference analysis results indicate that measuring distance by straight segments is feasible. Moreover, an exact and simple distance measurement method based on equidistance circles is presented based on the fact that small-circles on earth can be projected into circles on polar stereographic charts. The proposed method can satisfy the polar navigation application requirements for polar stereographic charts.
Calibration of the stereo-camera system is the basis of lunar rover navigation and cartography. The scientific instruments on the Chang'e-3 lunar rover include a Navcam, Pancam, and Hazcam, Self-calibration bundle adjustment with restrictive constraints based on high accuracy control field was employed to calibrate the stereo-camera system of Chang'e-3 lunar rover. The experimental results reveal that this approach can improve the accuracy and stability of calibration.
The Moon is the brightest body in the night sky, and it is of great value to be used to determine the azimuth. However, the key problem is how to solve the center of the apparent Moon accurately. This paper puts forward a new method called spherical circle fitting algorithm to determine the center of the apparent Moon. It has such many merits as strictness, briefness and little calculation in contrast to the existing algorithm. Simulated observation data is processed to understand the relationship between the attitude of the phase of the Moon and fitting accuracy. An experiment based on real observations was used to testify the correctness of the algorithm, and a comparison is made between the spherical circle fitting algorithm and the old algorithm. The results show that while the inner precision is almost the same, the outer accuracy of our algorithm improves by 0.5″. These results suggest that spherical circle fitting algorithm is more suitable for engineering applications.
The system error equations in the SINS transverse coordinate method are deduced for the transverse coordinate system. According to the solutions from the static base system error equations, the impact of system initial errors and inertial measuring unit errors on the transverse coordinate method is analyzed. System errors are simulated, and compared with system errors from the geographic coordinate computation method. Theoretical analysis and simulation results show transverse velocity errors and horizontal attitude errors are a type of periodic oscillation;position errors, azimuth error angle, wander azimuth error and heading errors increase with time. The transverse coordinate method can solve the problems arising when system errors increase and calculations output singular values in the geographic latitude and longitude coordinate calculating method.
Polarimetric interferometry SAR (PolInSAR) coherence is important information when solving the coherence optimization problem such as when executing forest parameter inversion. The distribution of coherences, the set of coherence with different polarimetric bases, is related to the field of values in a complex coherence matrix. It is discussed the specialty of the field of values of complex coherence matrix with different formation, and the relationship between the coherence with different polarimetric states and the field of values of complex coherence matrix. Using simulated SAR data and real SAR data, the influence of the sizes of window and scatters type on the field of values of a complex coherence matrix are evaluated. The results indicate that the field of values of coherence matrix can provide a new method to obtain optimal polarimetric bases and optimal coherence.
Each pixel of a tensor image is usually characterized by a 3-order positive definite matrix. Currently, existing methods of denoising tensor images usually regard the data as multi-channel images, which is likely to destroy the structure of positive definite matrices of the image and some information will be missed possibly. This paper addresses the problem of denoising complex tensor images. More precisely, we extend the anisotropic diffusion model, also known as P-M model, from scalar or vector images to complex tensor ones. The proposed method can be applied to remove speckle noises in PolSAR images. In contrast with existing denosing algorithms, our method is better at suppressing the effects of speckles while preserving edges.
Constructing the appearance model of object is a key problem that affects visual tracking performance. To solve this problem, we propose an online learning discriminative model for visual tracking;a robust tracking algorithm with this model with Bayesian estimation. Firstly, we segmented the initial tracking area to generate training samples, and obtained a discriminative model via clustering of the training samples. Then, we computed a likelihood map of the predicted tracking area of the current frame using the discriminative model. Finally, we estimated the object state via maximum a posterior estimation and updated the discriminative model online. The proposed algorithm updates the appearance model via online learning, which improves adaptability for large variations in appearance. Experimental results indicate that the proposed algorithm can cope well with complex change of object's appearance, demonstrating an especially robust performance when tracking an object undergoing scaling, illumination change, occlusion, and non-rigid deformation. Both qualitative and quantitative comparisons show the superiority of the proposed algorithm to now current state-of-the-art approaches.
Computing efficiency and the type of generator are a limitation and barely sustain the construction of Voronoi diagrams based on polyline/polygons. Thus, a new decomposition and combination algorithm for polyline/polygons is proposed in this paper, the main procedure transforms polyline/polygons into discrete feature points. According to cross connection characteristics, vicinity feature points pairs are constructed, whose cross parts of Voronoi diagrams can be substitute for the equal distance boundaries of polyline/polygons. Based on the area differences in iterative discrete computational sub-Voronoi regions, before and after, the generator is alternative in the iterative procedure that making sub-Voronoi region being gradually adjusted to the required precision. From empirical results and analysis, it can be concluded is that the time complexity of the proposed algorithm is O(n2) at high precision. Lastly, generated effect of polyline/polygon Voronoi diagram in the case of ample data using this algorithm is revealed.
Fast and accurate solution of a corner reflector's radar cross section is the important foundation of synthetic aperture radar radiometric calibration. Since a corner reflector used in radiometric calibration is usually electrically large, its backscatter field is appropriately analyzed with a high frequency approximation method;a fast and accurate solution is proposed in this paper. The incident field and its corresponding illuminated area are determined by a ray tracing of a incident wave and reflected wave, then the backscatter field of each illuminated area is computed by the Gordan surface integral method respectively and added up for the total RCS of a corner reflector. Simulation results agreed well with the results from the references and from the electromagnetic computing software FEKO. So the method proposed in this paper was demonstrated to be accurate and efficient, and especially suitable for fast radar cross section computation of corner reflectors composed of large size plates.
Service discovery is a hot topic in the field of services computing. As compared to classic methods that focus on service contents, this paper presents a new approach from the viewpoint of interconnectivity, to find important services and the corresponding composable services based on previous records of service composition. Empirical results show that (1) 95% of mashups on the website ProgrammableWeb contain important Web APIs discovered by our approach, and (2) the rate of success for analyzing the composability of a given set of Web APIs runs up to 98%. Thus, this approach would be helpful for effectively searching and recommending services on the Internet.