Dongchuan Wang

Susceptibility assessment of earthquake-induced landslides using Bayesian network: A case study in Beichuan, China

Because of the uncertainties and complexities of the factors involved in causing landslides, it is generally difficult to analyze their influences quantitatively and to predict the probability of landslide occurrence. In this work, a hybrid method based on Bayesian network (BN) is proposed to analyze earthquake-induced landslide-causing factors and assess their effects. Our study area is Beichuan, China, where landslides have occurred in recent years, including mass landslides triggered by the 2008 Wenchuan earthquake. To provide a robust assessment of landslide probability, key techniques from landslide susceptibility assessment (LSA) modeling with BN are explored, including data acquisition and processing, BN modeling, and validation. In the study, eight landslide-causing factors were chosen as the independent variables for BN modeling. And this study shows that lithology and Arias intensity are the major factors affecting landslides in the study area. On the basis of the a posteriori probability distribution, the occurrence of a landslide is highly sensitive to relief amplitudes above 116.5m. Using a 10-fold cross-validation and a receiver operating characteristic (ROC) curve, the resulting accuracy of the BN model was determined to be 93%, which demonstrates that the model achieves a high probability of landslide detection and is a good alternative tool for landslide assessment.

Spatio-temporal pattern analysis of land use/cover change trajectories in Xihe watershed

Abstract Human-induced land use/cover change has been considered to be one of the most important parts of global environmental changes. In loess hilly and gully regions, to prevent soil loss and achieve better ecological environments, soil conservation measures have been taken during the past decades. The main objective of this study is to quantify the spatio-temporal variability of land use/cover change spatial patterns and make preliminary estimation of the role of human activity in the environmental change in Xihe watershed, Gansu Province, China. To achieve this objective, the methodology was developed in two different aspects, that is, (1) analysis of change patterns by binary image of change trajectories overlaid with different natural geographic factors, in which Relative Change Intensity (RCI) metric was established and used to make comparisons, and (2) analysis based on pattern metrics of main trajectories in the study area. Multi-source and multi-temporal Remote Sensing (RS) images (including Landsat ETM+ (30 June 2001), SPOT imagery (21 November 2003 and 5 May 2008) and CBERS02 CCD (5 June 2006)) were used due to the constraints of the availability of remotely sensed data. First, they were used to extract land use/cover types of each time node by object-oriented classification method. Classification results were then utilized in the trajectory analysis of land use/cover changes through the given four time nodes. Trajectories at every pixel were acquired to trace the history of land use/cover change for every location in the study area. Landscape metrics of trajectories were then analyzed to detect the change characteristics in time and space through the given time series. Analysis showed that most land use/cover changes were caused by human activities, most of which, under the direction of local government, had mainly led to virtuous change on the ecological environments. While, on the contrary, about one quarter of human-induced changes were vicious ones. Analysis through overlaying binary image of change trajectories with natural factors can efficiently show the spatio-temporal distribution characteristics of land use/cover change patterns. It is found that in the study area RCI of land use/cover changes is related to the distance to the river line. And there is a certain correlation between RCI and slope grades. However, no obvious correlation exists between RCI and aspect grades.

Comparative analysis of land use/cover change trajectories and their driving forces in two small watersheds in the western Loess Plateau of China

Abstract To prevent soil loss and achieve better ecological environments, soil conservation measures have been taken during the past decades in the western Loess Plateau of China. In this paper, a case study was taken in Luoyu valley and Lver valley, two sub-watersheds of Xihe watershed and comparison was carried out between them. The main object of this study is to monitor land use/cover changes in the two similar small watersheds utilizing SPOT5 imageries by object-oriented human–computer interactive classification method, further develop the method of spatio-temporal analysis of land use/cover change by using pattern metrics of change trajectories and relative land use suitability index ( R ) in smaller watersheds, and make comparisons between the two similar small watersheds, taking water and soil conservation measures into consideration. Results show that combining GIS and RS, this method can be perfectly applied to make comparisons between different small watersheds with similar geographical backgrounds. And land use/cover spatiotemporal dynamic change characteristics can be preferably expressed by pattern metrics of change trajectories and R values based on topographical data. Different emphases have been laid according to their own geological backgrounds in the two watersheds and human activities have different effects on the landscapes of the two watersheds. The main change pattern is from slope farmland to terrace (322, the largest in Luoyu valley) or to economic fruit forest (344, the largest in Lver valley). R value of every slope grade in both of the two watersheds drops with the rising of slope degree on the whole and it shows that there is still much to do for people in the two watersheds in consideration that all the R values are still lower than 0.7.

Impacts of the Wenchuan Earthquake on the Chaping River upstream channel change

The Wenchuan Earthquake, measuring magnitude 8.0 on the Richter scale, occurred on 12 May 2008 in Sichuan, southwest China and caused over 87 000 casualties. Geological disasters such as debris flows and landslides caused by the Wenchuan Earthquake were severe. Several high-risk dammed lakes formed in the Chaping River basin of An’xian County, Sichuan, China, which is located in the earthquake zone, caused great changes in river morphology and brought great danger to the people and properties in the river downstream. Channel change information from different periods was important to risk relief of dammed lakes and assessment of losses. A Satellite Pour l’Observation de la Terre 5 (SPOT 5) fused image of 10 November 2006 with 2.50 m spatial resolution, and unmanned aerial vehicle (UAV) images of 19 May 2008 and 23 December 2008 with 0.32 and 0.33 m spatial resolution, respectively, were applied to derive channel information in the Chaping River upstream from three periods (pre-disaster, during the disaster and post-disaster) using an object-oriented classification method with an edge-based segmentation algorithm and a support vector machine (SVM) classifier. Geological landslides and dammed lakes from different periods were analysed. Based on the channel information from different periods, channel changes were determined, and impacts of the earthquake on channel area, channel width, river morphology, and so on were discussed.

The extraction of water information based on SPOT5 image using object-oriented method

The paper realizes water body extraction well using object-oriented classification method based on SPOT5 image. Firstly, segment image according to multi-scale image segmentation method based on edge-detection algorithm. Secondly, determine various characteristic parameters in land surface features according to object characteristics such as spectrum, shape and texture. And thirdly, use SVM (Support Vector Machine) method to achieve object classification through the establishment of the sample rules, and extract water body successfully. It takes SPOT5 image in Xiaojiaqiao section of Chaping River in Anxian County, Mianyang City, Sichuan Province, China as a case study. The results show that, compared with pixel-oriented supervised classification method, object-oriented classification method applied in water information extraction is more effective and its classification accuracy is higher.

Data integration and visualization: Dealing with massive and multi-dimensional marine spatial data

Marine observation has entered the era of three-dimensional observation at the present time. More and more spatial data is got form marine observing with no sophisticated visualization and analysis software for marine spatial data (MSD) either. Based on the analysis of data types and the acquisition methods in MSD, the paper proposed a data model for massive and multi-dimensional MSD. Then the procedure and implementation of MSD visualization is designed with the model. In addition, two applications for massive and multi-dimensional data visualization are presented.

Joint rectification of image series classification results based on trajectory analysis

There has been a common method to study land use changes based on series of remote sensing data. However, data on time nodes in the time series are almost the historical data, which is difficult in acquiring sample mark data for classification and result verification. Thus, the classification accuracy is severely limited. Taking a series of 4 remote sensing imageries of Xihe watershed, Gansu Province, Northwestern China as an example, the authors proposed a new method to jointly rectify image series classification results based on trajectory analysis. The object-oriented classification method was used to classify the remote sensing images, and results were output as vector data, which were then utilized to take trajectory analysis on the change process of every specific point in the study area. The olassification results were rectified by further investigation or expert querying on those patches with unreasonable trajectories. After joint rectification, compared with those with no joint rectification, the classification accuracy for the former three time node improved about 3%-8%, especially for the middle two periods of historical data, the classification accuracy improved up to 7% -8%.

Analysis of land cover changes using multisource and multitemporal images in loess hilly and gully regions

Land use/cover change (LUCC) is one of the most sensitive factors that show the interactions between human activities and ecological environments. In loess hilly and gully regions, to prevent soil loss and achieve better ecological environments, soil conservation measures have been taken during the past decades. This research attempts to detect land cover changes of Xihe ecological demonstration region by supervised classification to evaluate the effect of soil conservation measures. Support Vector Machines (SVM) classifier was applied for its ability to produce reliable classifications using small training samples. RBF kernel was finally employed for it nonlinearly maps samples into a higher dimensional space. Multi-source images of about one and a half decades were acquired for change detection, including Landsat TM and ETM+, CBERS2B CCD images. To minimize seasonal impacts, post-classification comparison method was employed for image processing. Unified land use classification plan was set up for all the images. The accuracy of classification was evaluated using more than 300 ground truth points mainly identified from high resolution Quickbird and CBERS2B HR images. The trend of land cover changes were detected and analyzed, and the relationship between land cover changes and human activities have been probed preliminarily. Result shows the ecological environments of the study area had deteriorated to a large extent in the first period, while were greatly bettered in the second period and controlling human activities or consciously exerting active human activities is an effective way to improve ecological environments.

Design and implementation of epidemiological field investigation method based on mobile collaboration

With the development of mobile technologies and the integration with the spatial information technologies, it becomes possible to provide a potential to develop new techno-support solutions to Epidemiological Field Investigation especially for the disposal of emergent public health events. Based on mobile technologies and virtual geographic environment, the authors have designed a model for collaborative work in four communication patterns, namely, S2S (Static to Static), M2S (Mobile to Static), S2M (Static to Mobile), and M2M (Mobile to Mobile). Based on the model mentioned above, this paper stresses to explore mobile online mapping regarding mobile collaboration and conducts an experimental case study of HFRS (Hemorrhagic Fever with Renal Syndrome) fieldwork, and then develops a prototype system of emergent response disposition information system to test the effectiveness and usefulness of field survey based on mobile collaboration.

Spatiotemporal analysis of land use/cover change patterns in the new coastal district of Tianjin, China

Land use/cover change is one of the most sensitive factors that show the interactions between human activities and the ecological environment. Since 1992, the government urged the development of the region of Bohai Gulf, the new coastal region of Tianjin has experienced great changes in land use/cover. Its urgent to detect the land use/cover change pattern to provide more explicit information on the further development of the new coastal district, which often requires to recover the history of land cover change and relates the spatio-temporal pattern of such change to other environmental and human factors, rather than merely relying on the change of areas or indices. This study makes Spatiotemporal analysis of trajectory land use/cover change patterns in the New Coastal District of Tianjin, after classifying the study area by Feature Extraction(FE) which first segments the whole area into objects and then classifies them based on the spectrum of the objects instead of pixels with Support Vector Machines (SVM) classifier. Multi-temporal and multi-source images of about one and a half decades were acquired for change detection, including Landsat TM, ETM+ and HJ1A images, and some land use maps were also used to supply training samples. First, all the images were geometrically corrected and registered on the same WGS84 coordinates. A unified land use classification plan was set up for all the images to classify land use type. The classification results were then utilized in the trajectory analysis of land use/cover change through the given three time nodes. Trajectories at every pixel were acquired to trace the history of land use/cover change for every location in the study area. Landscape metrics of change trajectories are also analyzed to detect the categorical change of different classes.