Remote Sensing for Environment of Angkor Site

1.Background of REAS Project
The project named “Remote Sensing for Environment of Angkor Site” (hereinafter called “REAS”) was launched in July, 2013, which is supported by HIST, APSARA and the UNESCO Phnom Penh Office. The background of this project is as follows:
The Memorandum of Understanding (MOU) between APSARA of the Kingdom of Cambodia and HIST in the People’s Republic of China was signed on 14 June 2013, which aims to strengthen and build APSARA capacity for the application of space technologies for the conservation and management of the Angkor World Heritage site and the sustainable development of the Region of Siem Reap of Cambodia.
Most of the existing projects in Angkor Site mainly focus on restoration and protection of the monuments; To dynamically monitor the surrounding environment and its impact on Angkor Site is of great significance to the conservation of Angkor and its sustainable development; Space technologies provide a systematic approach for studying the complex natural environment and human-environment interactions in Angkor (400 km2) and its surrounding areas (6,000 km2);
The Royal Government of Cambodia and APSARA have expressed strong interest in using remote sensing technologies for dynamic monitoring and applied research of Angkor Site. H.E. Dr. Sok An, Deputy Prime Minister also expressed his full support for the proposed collaborative project when he met with the HIST/RADI delegation in Phnom Penh in June, 2013.
HIST and APSARA carried out the field studies at Kulen Mountain in June and October, 2013 respectively.
2.REAS Project
The project aims to monitor Angkor site and its surrounding environment by space technology, and to better understand the relationship between them. This project is composed of the following contents: environmental management of the Angkor Landscape including dynamic monitoring of Kulen Mountains in the north of Siem Reap and regulation of human settlements surrounding Angkor site; water resources management and ancient water system reconstruction using satellite optical and SAR images; SAR applications to monitor land subsidence as a threat to Angkor monuments; spatial database construction of Angkor Site; capacity building.

Figure 1 Framework of the REAS project

Figure 2 Study area (red circle, about 6000 km2) of the REAS project

3.Research Progress of REAS Project
(1)Land Use and Land Cover Changes of Angkor Site during the Past 30 Years
Data used: Nine satellite images acquired between 1985 and 2003 were used in the study for analyzing the land use and land cover changes of Angkor site (see Table 1).
Table 1. Satellite images used in the REAS project for LULC mapping

Image Classification: Maximum Likelihood Classification was used for LULC identification, and seven LULC types were classified, i.e. forest, water, crop, grassland, urban area, barren, and wetland.

Figure 3. LULC classification map of the 2009 TM image

Figure 4. Areas for each LULC types from 1985 to 2013

(2)Urban area change of Siem Reap city
Based on the statistical results of the above time-series LULC classification maps, the conclusions for urban change issue are: high speed of urbanization since 2000; no urban area in 1985 (the MSS image); the increased urban areas were mainly converted from crop, forest and barren areas.

Figure 5 Urban area changes of Siem Reap city from 1989 to 2013

Figure 6. Urban changes from 2004 to 2012 (Quickbird, 0.61m)

Forest change of Kulen Mountains
The conclusions of forest change issue are: forest area reduced 593.2 km2 (18% of the whole forest) in the past 30 years; the forest loss was mainly due to conversion to crop fields and grassland areas.

Figure 7 Forest area changes of Kulen Mountains from 1985 to 2013

Figure 8. Forest changes from 1985 to 1996 (TM images, 30m)

(4)PS-InSAR for ground subsidence monitoring in Angkor site
Data used: the TerraSAR-X mission, characterized with high quality, higher spatial and temporal resolution, is indeed promising for an operational monitoring of the surface dynamics in the Angkor site. In the REAS project, 42 SAR images in the observation period of February, 2011 to July, 2013 were employed for time series analysis.
Methodology: one category of MT-InSAR, so-called Persistent Scatter Interferometry (PSI) was applied for the land subsidence monitoring over the whole scene. Compared with the current PSI method, two improvements have been achieved: 1) differential phases of neighboring target points were introduced to avoid the phase unwrapping procedure; 2) the velocity and height components were integrated by a least square solution; owing to the developed routine searching method based on the minimum spanning tree, low-reliable arc-connections could be avoided.
Results and interpretation: Using the described PSI method, the ground subsidence was derived, as illustrated in Figure 9. It is clear that the displacement on the Angkor site was evident. The annual velocity rates were primarily in the range of -50 to +12 mm/yr in the observation period of 2011~2013. Further studies will focus on the relationship between monument collapses and the ground subsidence from geology, hydrology as well as human-being activity aspects.

Figure 9. Time series analysis of TerraSAR-X images; (upper) the interferogram formation of PSI, (bottom) the derived annual surface deformation rate in the observed period of 2011-2013.