The potential of multi-sensor satellite data for applications in environmental monitoring with special emphasis on land cover mapping, desertification monitoring and fire detection

Huang, Shengli (2005): The potential of multi-sensor satellite data for applications in environmental monitoring with special emphasis on land cover mapping, desertification monitoring and fire detection. Dissertation, LMU München: Fakultät für Biologie

Abstract

Unprecedented pressure on the physical, chemical and biological systems of the Earth results in environment problems locally and globally, therefore the detection and understanding of environmental change based on long-term environmental data is very urgent. In developing countries/regions, because the natural resources are depleted for development while environmental awareness is poor, environment is changing faster. The insufficient environmental investment and sometimes infeasible ground access make the environment information acquisition and update inflexible through standard methods. With the main advantages of global observation, repetitive coverage, multispectral sensing and low-cost implementation, satellite remote sensing technology is a promising tool for monitoring environment, especially in the less developed countries. Multi-sensor satellite images may provide increased interpretation capabilities and more reliable results since data with different characteristics are combined and can achieve improved accuracies, better temporal coverage, and better inference about the environment than could be achieved by the use of a single sensor alone. The objective of this thesis is to demonstrate the capability and technique of the multi-sensor satellite data to monitor the environment in developing countries. Land cover assessment of Salonga national park in the democratic republic of Congo of Africa, desertification monitoring in North China and tropical/boreal wildland fire detection in Indonesia/Siberia were selected as three cases in this study for demonstrating the potential of multi-sensor application to environment monitoring. Chapter 2 demonstrates the combination of Landsat satellite images, Global Position System (GPS) signals, aerial videos and digital photos for assessing the land cover of Salonga national park in Congo. The purpose was to rapidly assess the current status of Salonga national park, especially its vegetation, and investigated the possible human impacts by shifting cultivation, logging and mining. Results show that the forests in the Salonga national park are in very good condition. Most of the area is covered by undisturbed, pristine evergreen lowland and swamp forests. No logging or mining activity could be detected. Chapter 3 demonstrates the one full year time series SPOT VEGETATION with coarse resolution of 1 km and the ASTER images with higher resolution of 15 meters as well as Landsat images for land cover mapping optimised for desertification monitoring in North-China. One point six million km2 were identified as risk areas of desertification. Results show within a satellite based multi-scale monitoring system SPOT VEGETATION imagery can be very useful to detect large scale dynamic environmental changes and desertification processes which then can be analysed in more detail by high resolution imagery and field surveys. Chapter 4 demonstrates the detection of tropical forest fire and boreal forest fire. Firstly, the ENVISAT ASAR backscatter dynamics and ENVISAT full resolution MERIS characteristics of fire scars were investigated in Siberian boreal forest, and results show these two sensors are very useful for fire monitoring and impact assessment. Secondly, the general capability and potential of ENVISAT multi-sensor of MERIS, AATSR, ASAR as well as NOAA-AVHRR and MODIS for tropical forest fire event monitoring and impact assessment in tropical Indonesia were investigated, and results show the capability of ENVISAT to acquire data from different sensors simultaneously or within a short period of time greatly enhances the possibilities to monitor fire occurrence and assess fire impact. Finally, the multi-sensor technology was applied to the disastrous boreal forest fire event of 2003 around East and West Lake Baikal in Siberia, and results show that 202,000 km2 burnt in 2003 within the study area of 1,300,000 km2, which is more than the total burnt area between 1996-2002. 71.4% of the burnt areas were forests, and 11.6% were wetlands or bogs. In total 32.2% of the forest cover has been burnt at least once from 1996 to 2003, 14% of the area has been affected at least twice by fire. These demonstrations show that in spite of the two disadvantages of indirect satellite measurements and the difficulty of detecting the cause of environment change, multi-sensor satellite technology is very useful in environment monitoring. However more studies on multi-sensor data fusion methods are needed for integrating the different satellite data from various sources. The lack of personnel skilled in remote sensing is a severe deficiency in developing countries, so the technology transfer from the developed countries is needed.

• Dokumententyp: Dissertationen (Dissertation, LMU München)
• Keywords: Satellite, remote sensing, multi-sensor, environment monitoring, desertification, fire, land cover
• Themengebiete: 500 Naturwissenschaften und Mathematik ; 500 Naturwissenschaften und Mathematik > 570 Biowissenschaften, Biologie
• Fakultäten: Fakultät für Biologie
• Sprache der Hochschulschrift: Englisch
• Datum der mündlichen Prüfung: 8. April 2005
• 1. Bericht­erstatter:in: Siegert, Florian