1043
A SIMPLE METHOD FOR INFORMATION EXTRACTION
OF FARM FIELD IN TARIM RIVER BASIN USING MODIS DATA
Shudong Wang 3 ', Suhong Liu 3 ’*, Xiaohua Wang b , Jianxia Guo 3 , Yanmin Shuai 3 , Ni Hu 3
State Key Laboratory of Remote Sensing Science; Beijing Key Laboratory for Remote Sensing of Environment and
Digital Cities; Center for Remote Sensing and GIS, School of Geography; Beijing Normal University, 19 Xinjiekou
Wai Street, Beijing, 100875, China-(wangsd @bnu.edu.cn
b School of Economic and management, Hebei University of Engineering, 199 Guangming South Street, Handan,
China,05 603 8-hdwxh@ 163 .com
KEY WORDS: Tarim River Basin, MODIS Vegetation Index, Time Series, Farm Field
ABSTRACT:
High spatial resolution remote sensing data are expensive and hard to cover the whole basin at the same time in Tarim river basin,
while, the low spatial resolution data, such as MODIS, which spatial resolution of first (red) and second (infrared) band are 250 m,
may cover the whole basin in same day. In this paper, based on previous models and methods, we compared the remote sensing
Vegetation index, NDVI, RVI and MSAVI (250m), which are thought most efficiently parameters and have been used widespread in
arid and semiarid area. Analyzing changing rule of soil, sparse natural and farm field of underlying surface, we put forward a new
method: calculating stand deviation using time series MODIS RVI data from April to October. Compared with extracting farm
information method of previous classification method and principle component analysis method using time series MODIS data, the
method is more efficient than others, and the accuracy exceed 80%, which verified by Landsat TM data and actual measurement by
GPS.
1. INTRODUCTION
Water is the foundation of the composition, development, and
stability of oasis ecosystems in arid areas [1]. It determines the
evolution of the ecological environment, including the two
contrary processes of oasis formation and development and of
desertification, and it is the key ecological factor in arid areas
[2]. It is widely recognized that human activities will affect
water resources and hydrological processes. Located in the arid
area of north-western China, Tarim River, about 1321 km long,
is the longest continental river in the world. It is mainly fed by
glacial/snow melt water, which accounts for 48.2% of the total
water volume in the basin [1]. The ecological environment in
the Tarim River Basin is extremely vulnerable. The
contradiction between ecological protection and economic
development is increasingly extrusive in the exploitation and
utilization of water resources, and the sustainable development
of the regional society and economy is seriously restricted [3-5].
Significant improvements have been made in regulating the
main stream of the Tarim River for flood control and water
supply mainly for agricultural purposes during the past several
decades. These improvements have made a great contribution to
the development of the economy in Tarim River Basin [2].
However, its negative influences such as deforestation,
desertification, and increased soil salinity have also been
considerable. Some eco-environmental problems including the
serious degeneration of the natural vegetation that relies on
groundwater for existence, the death of the herbs dominated by
Phragmites the degeneration of Tamarix chinensis shrubbery
and the Populus diversifolia forests in large areas, serious wind
erosion and land desertification, as well as the seriously
damaged ecosystems [6]. Both local and central governments
have paid significant attention to the altered hydrological
regime of Tarim River ecology. In an attempt to restore the
ecological system, ecological water releases from Boston Lake
through Daxihaizi Reservoir were carried out several times [7].
It was realized that these water releases were beneficial to the
potential restoration of the ecosystem.
But farm field water consumption is much larger than natural
vegetation in same area, so flood control and water supply
mainly for agricultural purposes during the past several decades,
the ecosystem still is facing deterioration [2, 7]. When water
supply for field isn’t enough, some field farm would be
abandoned, new ecological problems appear again [5], So
information extraction of farm field in Tarim River Basin is
important to decision making of both local and central
governments.
Compared with methods of other investigation, application of
remote sensing data is more efficient and timesaving. High
spatial resolution remote sensing data, such as Landsat TM, are
hard to cover the whole basin at the same time, and the low
spatial resolution data, such as MODIS, which spatial resolution
of first and second band are 250 m, may cover the whole basin
everyday. In this paper, based on previous models and methods,
we compared the remote sensing Vegetation index, NDVI, RVI
and MSAVI (250m), which are most efficient parameters and
are used widespread in arid and semiarid area, and selected
NDVI and RVI as basic data of information extraction.
2. SELECTION OF REMOTE SENSING VEGETATION
INDEX
Currently, the remote sensing Vegetation index, NDVI, RVI
and MSAVI (250m), are most efficient parameters and are
widespread used in arid and semiarid area.
* Corresponding author. Suhong. Liu