Most paddy rice fields in Asia are comprised of small parcels of land, and the weather conditions during the growing season are usually cloudy. This study develops a geographic information system (GIS) object-based post classification (GOBPC) that combines low-cost remotely sensed and GIS data to precisely map paddy rice fields in the intensively cultivated but fragmented growing areas which are characteristic of Asia. FORMOSAT-2 multispectral images have an 8-meter resolution and a one-day recurrence, making them ideal for mapping such areas. Multitemporal images are examined to distinguish the different growth characteristics between paddy rice and other types of ground cover. The pixel-based hybrid classification technique is used with both the unsupervised and supervised approach to distinguish the paddy rice fields from their surroundings. In addition to the pixel-based approach, we also use GOBPC to deal with over-fragmented parcels of land and to reduce the incidence of misclassification caused by speckle or mixed pixels (mixels) in the images. A comparison is made with the pixel-based technique. The Kappa index of agreement obtained with the GOBPC reaches 0.095 to 0.291, and there is a statistically significant improvement in the user and producer accuracy for all the classes (z>1.96) with McNemar's test in the four study areas. The proposed GOBPC approach is shown to be useful in highly fragmented rice growing areas and may have the potential for other agricultural applications.
Land degradation has become an important issue in western China recently. Oasis ecosystem is sensitive to
environmental disturbances, such as abnormal /extreme events of precipitations, water supply from upper watersheds,
fluctuations of temperatures, etc. Satellite remote sensing of terrestrial ecosystems provides temporal dynamics and
spatial distributions of landscape green covers over large areas. Seasonal green cover data are normally important in
assessing landscape health (ex. desertification, rate of urban sprawl, natural disturbances) in arid and semi-arid regions.
In this study, green cover data is derived from vegetation indices retrieved from MODIS sensors onboard Terra. The
satellite images during the period April 2000 to December 2005 are analyzed to quantify the spatial distribution and
temporal changes of Ejin Oasis. The results will help improving monitoring techniques to evaluate land degradation and
to estimate the newest tendency of landscape green cover dynamics in the Ejin Oasis.
The desertification in Northwestern China and Mongolia shows the result of conflicts between economic
development and natural conservation. Many researches have proven the desert areas are growing in these regions. The
variations of bi-weekly NDVI satellite images are used as one of the parameters to evaluate the vegetation dynamics
over large scale studies. In this study, remotely sensed satellite images are conducted to provide multi-temporal
vegetated and non-vegetated areas in order to assess the status of desertification in East Asia. Spatial data derived from
these satellite images are applied to evaluate vegetation dynamics at regional scale to find out the hot spot areas
vulnerable to desertification. The results show that the desert areas are mainly distributed over southern Mongolia,
central and western Inner-Mongolia, western China (the Taklimakan desert). The desert areas were expanded from 2000
to 2002, were shrunk in 2003, and were expanded from 2003 to 2005 again. The hot spot areas of desertification are
mainly distributed over southeastern Mongolia and eastern Inner-Mongolia. The results will help administrators to refine
the planning processes in defining the boundaries of protected areas and will facilitate to take decision of the priority
areas for conservation of desertification.
East Asian dust sources are mostly located in remote areas where the geographical characteristics are not well understood. This imposes difficulties in the modeling of atmospheric mineral dust. Satellite remote sensing is a viable way of deriving temporal dynamics and spatial distributions of dust emission over large areas. Although the present-day satellites provide only column-integrated aerosol properties, it is possible to retrieve from them the dust emission strength, which is also a column denomination. In this study, dust emission over East Asia is estimated using aerosol optical depth retrieved from MODIS sensors with the aide of a regional meteorological and dust model. Differences of daily AOD and meteorological data are applied to the mass conservation equation to estimate net emission of dust over specified desert areas. The derived emission is then compared with the old data set used in the model for dust simulation in East Asia. The area of study covers northern and western China as well as Mongolia. Asian dust events identified during the period January 2003 to December 2004 are selected for emission retrieval and for modeling comparisons. The results show that the patterns of emission coefficient retrieved from Aqua-MODIS AOD data are consistent with spatial characteristics of land cover over Northern China and Mongolia.
For the assessment of climatic impact of aerosols, the knowledge of both the temporal and spatial distributions of aerosol is essential. Laser radar, more popularly known as Lidar, has becoming one of the most powerful techniques for active detection of aerosols in the atmosphere. Lidar can provide vertically resolved of extinction and backscatter coefficients, and thereby the height of the planetary boundary layer or the nighttime residual layer. As the long-term changes in the structure and dynamics of the lower and middle troposphere is now becoming a priority, a pulsed Nd:YAG Lidar system is applied for measuring the vertical distribution of aerosol properties in the metropolitan Taipei. Two years (2004-2005) of aerosol optical depth (AOD) measured by Lidar, Cimel Sunphotometer and MODerate resolution Imaging Spectroradiometer (MODIS) were compared. The AOD shows strong seasonal variation with maximum values (AODLidar > 1, AODCimel > 1 and AODMODIS > 0.39) occurred in April. AODMODIS shows significant underestimation. AODLidar has good correlation with AODCimel, but the Lidar measurement is biased toward lower values as presented by the 0.725 slope in the linear regression. This bias is mostly caused by the Lidar blind distance at the lowest part of the atmosphere. The R-squared of AODCimel and surface PM2.5 concentration is about 0.44. This reflects the fact that the atmospheric boundary layer is often not well-mixed, so aerosols there cannot represent the total AOD value. Particles in the free troposphere also need to be concerned. Further comparison of our Lidar data with the CALIPSO measurements is intended.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.