A case study of the Khorat plateau and its vicinity in Thailand using high resolution satellite imagery

A case study of the Khorat plateau and its vicinity in Thailand using high resolution satellite imagery

Tokio Kachi, Yoshio Akiyama, Jiroh Komai
Earth Resources Satellite Data Analysis Center, Tokyo, Japan

S. Yaemniyom, K. Kuntawang, T. apakasetr
Economic Geology Division Department of Mineral Resources
Thailand

Abstract
Landsat TM and SPOT imagery of Nam Phong area and the surrounding areas were used to map surface geology. SPOT color composite images integrated with high resolution panchromatic data resulted in enhancing spectral differences in the vegetation at the surface. Vegetation patterns according to the subtle difference of topography clearly revealed several circular features.

Field spectra were acquired with hand held spectrometer fitted with seven spectral bands coinciding with the Japanese ERS-1 bands to with seven spectral difference in vegetation according determine the cause of the spectral difference in vegetation according to the difference of landuse observed in SPOT color composite images.

On the circular features extracted from the satellite imagery of the Khorat Plateau, the verification by the newly developed PLMT (Power Line Magnetotelluric) method was carried out in order to acquire the distribution of subsurface rock resistivities concordant to the working hypothesis that that the surface uneven terrain reflects the difference of subsurface lithology, resulting in being expressed as a circular structure.

Landsat TM image clearly revealed the north south trending phetehabun infra mountain basin as a garben type basin, where the inside Cenozoic sediments and the outside Paleozoic / Mesozoic rocks are bounded on both sides by two parallel lineaments, although no fault on the surface has been found to prove those lineaments to be large displacement faults. PLMT method and "Fingerprint" geochemical method were carried out at eh eastern boundary lineament and verified the existence f large displacement fault.

Introduction
In the Khoral Plateau area,(Fig.1) surface circular features on the Jurassic (Cretaceous terrain are particularly conspicuous (CCOP 1983). At the Nam Phong structure in the Western edge of the Khorat plateau, the surface subtle uneven terrains are adjusted to several types of landuse. Judging from this surface condition, it is possible to extract a circular feature caused by subsurface geological structure in spite of no outcrop at the surface (ERDAC, 1988). Fro the verification of this analytical result, field checking with portable spectra radiometer (Yamaguchi et al 1988) is necessary so that colors on the false color image of the Nam Phong structure can be compared to real field spectra of vegetation patterns.

Figure.1 LANDSAT TM false-color composite mosaic(bands 2,3,and 4)showing the khorat Plateau and the Phetchabun graben

At the south of Nam Phong structure located approximately 20km southwest of Khon Kaen City, a circular feature extracted from high resolution SPOT color composite image is also inferred to reflect subsurface differential rock resistivities. One of the magnetotelluric tools, PLMT method is thought to be convenient for the certification of such lithological differences in shallow under guard (Tsubota et al 1987).

Further more, at the Petchabun graben in the Central Thailand. North south trending intra mountain basin with 30 km in width flanks on Permian rock mountains on the both sides. Lineaments extracted from the Landsat TM image also have to be surveyed with the PLMT method as well as "Fingerprint" method, which is useful to verify the existence of deep-seated faults.

The purpose of this paper is to verify previously predicted geological hypothesis on circular features and large scale lineaments extracted from high-resolution satellite imagery.

Satellite data and processing
The data used in this case study were 9 Landsat MSS data, 7 Landsat TM data 6 SPOT multi spectral and 1 SPOT HRV Panchromatic data. All data except HRV Panchromatic one were enhanced by linear stretch processing method to produce full scene false color images. It subscene data in the interesting areas were also processed of principal component analysis techniques to elucidate the effectives of multisepctral data applied to the discrimination of surface objects.

Two HRD subscene composite images were also produced to obtain high resolution (10m) color images. As for HIS (Hue, Saturation, intensity) transformed multi spectral data (20m), intensity was replaced by panchromatic band (10m), and then it was inverse transformed to false color of HRV composite images.

Ground truth verification
Ground based surveys were conducted in 1987 and 1988to confirm the extracted circular features and lineaments. The survey was also carried out to collection more than 40 spectra of bedrock, soil and representative vegetation at 34 sites within the study area of the Khorat Plateau. All spectra were acquired with FPR 2000 spectroradiometer manufactured by Optical Science Co. Ltd. (Yamaguchi et. al 1988). This hand held radiometer has eight visible / very near infrared (VNIT) filters and eight short wave infrared (SWIR) filters. For this study we selected eight filters to measure spectral bans coinciding with Landsat TM badn 1 and the Japanese EERS-1 bands.

Filter number	1	2	3	4	5	6	7	8
Median wave length (nm)	485	560	660	810	21650	2060	2190	2340
Band width (nm)	70	80	60	100	110	110	120	130

For the verification of geological interpretation from satellite image, it was also recommended to apply Power Line Magneto telluric (PLMT) method (Tsubutoa et al. 1987) and "Fingerprint" method (Viellenave et al 1986) because there was no outcrop on the surface. The PLMT method consists of a hand held instrument to measure the surface magnetic field generated by commercial power lines and immediately calculates the frequency dependent (50 Hz and harmonies apparent resistivity.

Figure.2 Index map of northeastern Thailand showing the locations of study areas

If near surface layer was homogenous, penetration depth (D) could be describes as follows :

D = 503 x (d/f)^1/2

d: apparent resistivity (Wm)
f: frequently (Hz)

Since actual underground is composed of multi layers, underground resistivity distribution has to be obtained through reiterating simulation under the multi layer model.

The fingerprint method is convenient to verify the lineaments interpreted as large faults. The activated charcoal absorbent is designed to trap microgas seepage coming up to the surface through faults. The microgas analysis with the curie point gas pyrolizer and quadruple mass spectrometer results in separating low background data operations from high background data locations and detects the anomalies of heavy gas components which indicate possible fault locations.

Results

The Nam Phong Structure
In the following paragraph some of the words are illegible in the source

Figure.3 Physiographic map of the Nam Phong area showing the locations of spectral measurement

Figure.4SPOT HRV Composite image of the Nam Phong structure

A ____ feature at the Mesozoic terrin can be easily extracted from the SPOT HRV composite image. This typical circular feature is composed of four colors. Blue color areas form linear patterns in which some rills and small streams can be identified. The areas composed of white, light red, and dark red _____ also form annular pattern, which are separated from blue color areas.

As a result of spectral survey with FPR-2000, typical spectral patterns of 8 bands representative vegetation type, soil, sandstone and river were obtained. Band 4 of FPR 2000 mainly represents the volume of vegetation the same as TM band 4 and HRV band 3.

Fig. 4 shows that the relative strength of reflection for band 4 increases in descending order. Accordingly, the comparison of relative strength for each object is as follows:

1 (Eucalypts, teak, mango, banana, coconut, tobacco, cassava, bamboo) 2 (______) 3 (sugar cane, cone) 4 (grass, rice field) 5 (the vegetation area soil). Fig.2 and 3 correspond to dark red and light red color areas of the HRV false color image. 4 is shown as blue and white color areas and 5 as white color areas.

On the false color image whose red component is allocated to HRV band 3. White color area is inferred to have more vegetation than blue color area because white means very high reflection of HRV band 1 and 3, while blue color means less reflection of HRV band 3. However, blue color areas (No. 21008) have actually more vegetation and higher reflection of HRV and 3 than white color area (N0. 21001). This discrepancy is caused by linear stretch processing in stretching the strength of reflection for each band and some grass/shrub area and non-vegetation area are both represented as the same white color on the dales color image as a result.

This problem can be resolved by FR-2000's pseudo color ratio composite image of band ratios ¼ or ¾ Spectral survey results over the Nam Phong and other areas show the advantage f FRPS-2000's multi band processing capability.

Figure.5 Spectral Pattern of the Nam Phong area

Figure.6Block diagram of the nothern part of the Nam Phong structure
The South of Nam Phong Structure:
This tone anomaly is inferred to reflect possible geological structure because it consists of red, white and blue patterns and resembles the Nam Phong Structure, in addition, it is located between the Nam Phong and the Chonnabot anticlinal structures. However this tone anomaly is not shown on the published geological map, (DMR, 1983).

Figure.7 SPOT HRV composite image of the South of Nam Phong structure

Because no outcrop has been found within this anomaly, the PLMT method was carried out to verify the geological hypothesis that the surface the even terrain reflects the difference of subsurface lithology resulting in being expressed as a circular structure. The apparent resistivity profile of PLMT simulation showed the distribution of low resistivity layer (appr. 1000Wm). Consequently, the _________ of two layer was down to appr. 200m beneath the surface, and the suggested the existence of possible anticlunal structure.

Figure.8 Geological imterpretation of SPOT IIRV composite image of the South of Nam Phong structure and PLMT apparent resistivity distribution
The North of Petchabun :
This graben feature has two parallel lineaments, and in their __________ the outside Mesozoic/Paleozoic sedimentary rocks bound on the inside Cenozoic sediments.

The DMR's geological map also depicts these lineaments as normal faults, although the outcrops of these faults have not been found yet, that is the graben has not been proved by geological survey yet. The PLMT method was carried out at 38 points near the Highway 12 of the eastern boundary of graden and then, the eastern part of the major lineament at the eastern boundary showed the apparent resistivity of more than 1000Wm. The resisitivity in the vicinity of the lineament was transitional between 100 to 400Wm at 150Hz responses. The simulation results of the PLMT data showed that the assumed depth of high resistivity basement was more than 400m at Loc. 28, thereby indicating that Cenozoic sediments are so thick near the more lineament that the boundary between the eastern Permian and the western Cenozoic rocks should be interpreted as a lag displacement fault.

The fingerprint method was carried out at 3 points along the highway 12. The major lineament was presumed as a boundary fault at near Loc-18, and the minor lineament was presumed as a small fault within Tertiary sediments at near Loc.2. Analytical result is presented in the diagrm of HMWGR (high molecule weight gas ratio) VS MTIC 9modified total ion count) in which data are classified into (1) gas anomaly samples, (2) background high HMWGR samples, (3) background-A samples, and (4) background-B samples. Since Loc-16 and 18 are classified into (1) these data also indicate the existence of the major lineaments. Since Loc-2 is classified into (1) in spite of 100 meters sampling interval, this will seem to indicate a fault corresponding to the minor lineament.

Figure.9LANDSAT TM Principal Component Image of the North of Phetchabun area

Figure.10 Index map of the Norht of Phetchabun area showing the locations of PLMT and Fingerprint surveys

Figure.11 Geological imterpretation of SPOT IIRV composite image of the South of Nam Phong structure and PLMT apparent resistivity distribution

Figure.12Gas feature diagram for sites on Highway 12 shown in Figure 10

Recommendation
Guideline of future research is suggested by the result of this case study as follows.

At the outset, ground truth checking with land held geophysical and geochemical survey tools such as PLMT and "fingerprint" should be essential to the geological mapping of heavily vegetated areas so that geological interpretation from high resolution satellite imagery will become more reliable.
In case of the discrimination of vegetation and landuse, the only use of false color image should not be recommended from the reason that relative spectral characteristics among multi bands might be changed by stretch processing of each band. Rationing image among the JERS-1 multi bands will be rather recommended for the detailed classification in terms of the conservation of spectral pattern of each band.

Conclusion
This case study revealed the possibility to distinguish circular features and lineaments from currently available satellite data. Through this case study the valuable experiences for verification method have been gained and we, then, can use this result as a model example for future verification research on the Japanese ERS-1 multi band data, in conjunction with the ever improving satellite data provided, promises more convenient access for the geological survey even in vegetation covered areas.

References

CCOP-IOC (1981) Studies in East Asian Tectonics and Resources (SEATAR) CCOP Technical paper, pp.25-36 D.M.R (1983) Geological Map of Thailand, scale 1/500,000, Department of Mineral Resources.
ERDAC (1988) Report on Research and Development of Remote Sensing Technology for Non-Renewable Resources, Earth Resources Satellite Data Analysis Center (in Japanese)
Tsubota, K., Noritake, K., Sakuma, H., Ohya, T., Saitoh, A. (1987) Development pf a New Mt System Using the Commercial Power line as a Source. Proc. The Society of Exploration geophysicists of Japan Autumn Meeting (in Japanese)
Viellenave, J.H. an Bloom, D.N (986) the Petrex Fingerprint Technique as Applied to Dry Gas Exploration Bounde Creek Field, Colusa and Glenn Counties, California. Proc. AAPG Pacific Coast Section Meeting.
Yamaguchi, Y., Sato., I,. Ohkura, T. (1988) Development of a Field Radiometer as a round truth equipment for the Japanese ERS-1. proc. International Society for Programmertry and Remote Sensing 16th Congress, Vol.27, Part B7, p. 718-727.