Monitoring and Analysis of Deforestation Process using Satellite Imagery and GIS (Case Study in Myanmar)
Ko Ko Lwin, Ryosuke, Ryosuke Shibasaki
Center for Spatial Information Science, University of Tokyo
E-mail: Iwin@skl.iis.u-tokyo.ac.jp
Abstract: One of the major limitations in detecting and tracing spatio-temporal changes of land covers using satellite imagery is the spatial and temporal resolution of the sensors. The high spatial resolution obtains low temporal high temporal resolution like Landsat TM, which cannot detect rapid change, On the other hand, low spatial resolution obtains high temporal resolution like NOAA AVHRR, which cannot detect detailed spatial change. In this study we used both Landsat TM and AVHRR images, where deforestation were monitored by AVHRR Time Series Images. In order to understand causes of deforestation process, we also analyzed relation between deforestation and landscape attribution using GIS and generated a deforestation risk map based on the analyzed results.
Key Words: Land Cover Change Detection and Monitoring, Deforestation, Difference resolution satellite images. Annual spatial Expansion of Deforestation, GIS role in analysis of deforestation process.
1. Introduction
Deforestation, desertification, pollution and resources depletion are now serious problems in the world, which threaten the survival of half of the living things in the world. By understanding of these environmental change processes, we can conserve renewable or non-renewable resources that can be used for future generation without seriously damage.
On the other hand, one of the major limitations in study of change detection is satellite resolution problem due to different observational frequencies. Classical change detection studies were performed by single data image processing, which cannot allow to detect subtle change or annual change.
In this study, we demonstrated that combination that combination of two-resolution satellite image (Landsat TM and AVHRR) allows to detect and monitor changes with finer spatial and temporal resolution. In order to understand causes of deforestation process, we also analyzed relation between deforestation process and landscape attributes deforestation process and landscape attributes using GIS.
2. Description of the Study
2.1 Study Area
The study area is lower part of Myanmar, located between Thailand and India.
Figure 1. Location of Study Area
The total study area is 2,798,936 hectares (170 km X 165 Km) and elevation is ranging from between 0 to 2000m. Western part of the mountain is ranging from 0 to 200m and eastern part of the mountain is ranging from 0 to 2000m. The central area is flat region.
2.2 Study Objectives
The following are objectives of this study.
- To develop a method of land cover change detection and monitoring using different resolution satellite images.
- To analyze the relation of between deforestation process and Landscape attributes.
- To generate a deforestation risk map based on current deforestation process.
2.3 Study Material and Media
The following table shows by study materials and media that are used in this study.
Table 1. List of Study Material and Media
| Satellite Images |
| Landsat |
Date Acquired |
Spatial Resolution |
Source |
| Landsat TM |
9Jan 1999 |
30 x 30m (Raw Image) |
EROS Data Center |
| Landsat TM |
7 Mar 1995 |
50m x50m(Geo-coded) |
JAFTA
Japan Forest Technical
Association |
| NOAA AVHRR |
2 Jan 1989
31 Dec 1990
5 Jan 1992
29 Dec 1993
13 Jan 1994
1992 data set
1993 data set
1995 data set |
1Km x 1Km
(Level 1 Raw)
1Km x 1Km
|
National Institute for
Environmental Study
(Japan)
10 days Composite NDVI
data
(USGS EROS data center) |
| GIS Data |
| Township Vector Boundary (ARC Info Coverage Format)Road Network Vector file (ARC Info Coverage Format)DEM (Digital Elevation Model) 1 Km |
| Map and Statistical Data |
Forest Cover Map of Lower Myanmar Regions, Paper Map, Scale 1:250,00 (JAFTA)
World Atlas (Digital Format)
Myanmar Data on CD ROM
(CSO, Ministry of National Planning and Economic Development, Myanmar) |
| Hard Ware and Software |
PC and UNIX
ERDAS Image Version 8.3.1 Windows NT o/s |
2.4 Study Flow
There are two steps in this study, first we generate the land cover change map between 1989 and 1995. The second step is monitor the spatial expansion of deforestation and analyzed the deforestation process.
Figure 2. Over All Study Flow
