1

Integrated remote sensing in determining areas of high geothermal potential in the exploration of the Amacan Geothermal Prospect, Philippines

Jeffrey T. Bermido, Kevin G. Guillermo, Oliver A. Briola, Leonardo L. Morales, Releo D. Contemplacion, and Joeffrey A. Caranto

Energy Development Corporation, One Corporate Center, Julia Vargas cor. Meralco Ave, Ortigas, Pasig City, Philippines

bermido.jt@energy.com.ph, guillermo.kg@energy.com.ph, briola.oa@energy.com.ph, morales.ll@energy.com.ph, contemplacion.rd@energy.com.ph, and caranto.ja@energy.com.ph

ABSTRACT

Leonard Kniassef is one of the active volcanoes in the Philippines, located in the eastern portion of Mindanao Island. It is interpreted to host the high temperature Amacan geothermal system, known for its Quaternary volcanism set on a favorable structural environment. In order to identify priority areas for detailed exploration, this study integrated three remote sensing techniques with minor statistics application namely: lineament analysis, hydrothermal alteration, and thermal anomaly mapping. Digital elevation models were processed for the lineament analysis, where slope statistics were correlated with the underlying tectonics. Thermal Infrared Sensor Band 10 of Landsat 8 was also processed using a series of raster calculations to highlight areas with high surface temperatures. The Operational Land Imager (OLI) bands of Landsat 8 were also processed using composite and band ratio operations to highlight alteration zones and discriminate the type of alteration present. Combining the three remote sensing results, five priority areas are identified to have high geothermal potential and activity. These areas are recommended for more detailed geoscientific assessments.

Keywords: remote sensing, geothermal exploration, lineament analysis, hydrothermal alteration, thermal mapping

1. INTRODUCTION

Remote sensing has been applied widely in the earth sciences, including geohazard assessments and natural resource management. This study extends the application of remote sensing to geothermal exploration as one of the first steps in an overall work program. As a time- and cost-efficient technology, remote sensing has high impact on how geological surveys are planned, especially on identifying priority areas for detailed assessment. It would also help address the data gap in areas that are difficult to reach due to challenging terrains and lack of access paths.

The study is focused on the Amacan Geothermal Prospect, centered on the Leonard Kniassef Volcano. Carbon dating of the youngest tephra deposits yielded an age of 1800 ka (Wood, 1980), which makes it one of the active volcanoes in the Philippines. The area is also generally active with the presence of the Philippine Trench offshore and the Philippine Fault inland. The sinistral Philippine Fault in Eastern Mindanao cuts through Holocene sandstones exposed in Mati, Davao Oriental signifying its active status (Yumul et al., 2008). The Philippine Fault bifurcates into two, namely the Eastern Mindanao Fault in the west and the Mati Fault in the east forming a spindle-shaped structure (PHIVOLCS, 2015). This mirrors the fault architecture of the Philippine Fault in Leyte Island (e.g. West Fault Line and Central Fault Line) where the Tongonan Geothermal Project and other geothermal prospects are all straddled. The Philippine Fault in Eastern Mindanao also forms a couple with the younging southward Philippine Trench (Quebral et al., 1996, Lallemand et al., 1998). The active status of this tectonic couple makes the area seismically active relative to the rest of the archipelago. This tectonism which translates to permeability is thereby favorable for mineralization and development of geothermal systems.

2

Figure 1: (Left) Regional geologic setting of the Amacan geothermal prospect in blue box. Modified from PHIVOLCS (2015). (Right) The prospect is hosted by the active Leonard volcano and bounded by two segments of the NW-SE trending Philippine Fault.

Given this impressive regional volcano-tectonic setting, the study aims to refine and further add to this existing information on the local scale through the application of remote sensing. This will be used as reference during the conduct of the geothermal exploration survey.

2. MATERIALS AND METHODS

2.1 Slope Analysis

Shuttle Radar Topography Mission (SRTM) dataset from Earth Explorer was used in the geomorphological and lineament analysis. The DEMs are processed to produce hillshade, slope aspect, and slope gradient maps. Hillshade processing calculates the illumination value of a cell given a particular azimuth and altitude of a hypothetical light source. Slopes facing the illumination will be highlighted, while shadows will be casted on other slopes. For this study, illumination angle azimuths used are 045°, 135°, 225°, and 315°. The Z factor of 0.00000912 for the 0-10° latitude of Eastern Mindanao was also used in the analysis. The elevation angle of the illumination source is set at 20°.

Slope aspect reflects the direction into where the slopes are facing according to a color scheme displaying the 0-360° azimuth. Aspect maps are found to be suitable in defining abrupt changes in slope directions that may represent the presence of geologic structures. This may be indicated by the presence of valleys, ridges, and lineaments. Volcanic deposits can also be differentiated based on the variations in slope aspect. Using automated slope statistics from ArcGIS, the mean, maximum, and minimum value of slope orientation were also determined.

Slope gradient processing, usually generated for geohazard analysis, calculates the slope inclination or steepness of a surface in either degree or percentage values. According to Braganza (2014), while steep-sided slopes and flat terrains can be deduced from hillshade map, the values and degree of flatness or steepness of the topography is more properly illustrated in the slope gradient map. Broad flat areas are typically represented by white to green shaded regions while steep sided slopes are in the orange to red. Slope statistics was also applied to determine the mean, maximum, and minimum slope values.

3

2.2 Land Surface Temperature

The second dataset is the Landsat 8 package of Eastern Mindanao with acquisition dated August 2015. This is comprised of Operational Land Imager (OLI) bands covering Bands 1-7 and Thermal Infrared Sensor (TIRS) bands covering Bands 10-11. As recommended by the US Geological Survey (2016), Band 11 was not used due to its large uncertainty factor. A series of steps based on Avdan and Javanovksa (2016) were applied to convert the OLI Band 10 using raster calculations in ArcGIS.

Using the raster calculation tool in ArcGIS, the first step in the process (Equation 1) was the calculation of the Top of Atmospheric Spectral Radiance using the radiance rescaling factors included in the metadata file downloaded.

𝐿𝜆 = 𝑀𝐿𝑄𝑐𝑎𝑙 + 𝐴𝐿 (1)

Where Lλ is the Top of Atmosphere spectral radiance, ML stands for the band-specific multiplicative rescaling factors, Qcal is the quantized and calibrated standard product pixel values, AL is the band-specific additive rescaling factor from the metadata file.

The value of Lλ was then used to determine the brightness temperature (BT) in Equation 2 using the thermal constants provided in the metadata file. BT is defined as the temperature measured-at satellite and not the temperature on ground. Note that the resulting brightness temperature is in Kelvin and has to be converted into Celsius using conversion formulas.

B𝑇 = 𝐾2

𝐼𝑛 (𝐾1𝐿𝜆

+ 1) (2)

Where BT is the at-satellite brightness temperature (K), Lλ is Top of Atmosphere Spectral Radiance, and K1 and K2 are band-specific thermal conversion constants from the metadata file. On a separate calculation, Bands 4 and 5 were used to determine the Normalized Difference Vegetative Index (NDVI). The NDVI is used to quantify the green leaf vegetation around the globe.

𝑁𝐷𝑉𝐼 =𝑁𝐼𝑅 (𝐵𝑎𝑛𝑑 5)− 𝑅 (𝑏𝑎𝑛𝑑 4)

𝑁𝐼𝑅 (𝐵𝑎𝑛𝑑 5)+ 𝑅 (𝐵𝑎𝑛𝑑 4) (3)

Where NIR or near-infrared is represented by Band 5 and R or the red band is represented by Band 4. The value of NDVI was then used to calculate for the Proportion of Vegetation (PV).

𝑃𝑉 = (𝑁𝐷𝑉𝐼−𝑁𝐷𝑉𝐼𝑚𝑖𝑛

𝑁𝐷𝑉𝐼𝑚𝑎𝑥−𝑁𝐷𝑉𝐼𝑚𝑎𝑥)

2 (4)

Where NDVImin and NDVImax are the minimum and maximum values of the NDVI, respectively.

The value of PV was then used to calculate for the Land Surface Emissivity (LSE). The LSE is defined as the proportionality factor that scales blackbody radiance or the efficiency of transmitting thermal energy across the surface into the atmosphere (Avdan & Jovanovska, 2016). A simplified version of the equation is presented in Equation 5.

𝐿𝑆𝐸 = 0.0004𝑃𝑉 + 0.986 (5)

For the final calculation of the Land Surface Temperature (LST), the values of the Land Surface Emissivity, brightness temperature, and other scientific constants were used in the equation.

𝐿𝑆𝑇 = 𝐵𝑇

1+𝑤[(𝐵𝑇

𝑝) (ln(𝐿𝑆𝐸)] ]

(6)

Where LST is the Land Surface Temperature, BT is the brightness temperature, and LSE is the land surface emissivity. W is the wavelength of emitted radiance at 11.5µm in the TIRS bands. The value of p was calculated using Equation 7 which is equal to 1.438 x 10-2 mK.

p=h c/s (7)

4

Where h is the Planck’s constant equal to 6.626 x 10-34 J.s., c is the Boltzmann constant equal to 1.38 x 10-25 J/K, s is the speed of light in vacuum (3 x 10-8 m/s).

2.3 Hydrothermal Alteration Mapping

The OLI bands were processed using ArcGIS Model Builder for the hydrothermal alteration mapping. Color composite processing was performed to show the spatial distribution of hydrothermal alteration relative to the background. Three additive colors were used to display multispectral bands (Mia, 2012) corresponding to the RGB values of a false composite image. The ratio applied in this study was 5:7:6. Band ratio operation, on the other hand, was employed to identify the type of alteration present. Abrams ratio (6/7:4/3:5/6) was applied in this study to distinguish areas with iron oxide and clay alteration.

The results from the three main techniques were integrated to show priority areas with geothermal activity and potential, in terms of the presence of structures and overlaps of hydrothermal alteration with high surface temperatures.

3. RESULTS AND DISCUSSIONS

3.1 Geomorphology and General Topographic Description

The digital elevation model in Figure 2.A was analyzed using several raster conversions in ArcGIS. Hillshade processing (Figure 2.B) was able to display the terrain variations in the area. The slope gradient map (Figure 2.C) showed the gradation of surface inclinations while slope aspect (Figure 2.D) showed the directions into where the slopes are facing. Figure 3 combines the hillshade and slope gradient maps to better reflect the terrain variations in the area.

The prospect encapsulates an oblong shaped terrane centered on Leonard (Figures 2 and 3). Previous geomorphologic interpretations by Paguican (2012) classified that this terrane is a volcanic massif given its low height/basal width ratio of 0.05, irregular plan shape, with intermediate to high ellipticity. In terms of dimensions, the massif has a basal area of 366.4 km2. The highest peak in Amacan is Mt. Dasuran at the south which rises to about 1889m. The lowest elevations with about 100m are in the sediment laden plains at the southeast in the town of Maragusan and at the north in Mawab and Nabunturan (Figure 3).

In general, the volcanic massif is characterized by moderately steep slopes with around 60% in the 18°-35° interval. An exception of this is the wide area east of Leonard which is differentiated by a smoother surface of slopes at 11°-18°. The more rugged terrain possibly reflects the extent of older volcanic deposits or the flank of older edifices. This is also supported by the radial drainage pattern surrounding the center, which is typical in volcanic systems. Contrarily, the smoother terrain east of Leonard represents the possibly younger volcanic deposits. These have relatively less fault incisions and not as deep as the older deposits.

The slope aspect map also shows the differences between the old and young deposits within the massif. The older deposits are characterized by slopes facing the northeast. The younger deposits on the other hand have no dominant slope directions. Paguican and Bursik (2016) explained that areas having slopes facing in many directions may indicate rough material deposits, such as emplaced lava flows or avalanche and slide toes.

Outside the Leonard massif, the eastern terrane is characterized by a N-S trending morphologically homogenous body with almost 80% of the slopes falling under the 34-72° interval. Slopes on its eastern side face the north to the east while slopes on its western side face the south to the west. Correlation with existing maps indicated that this is part of the intrusive complex in the area, identified as diorite (Quebral et al., 1996).

The western block shares almost similar characteristics with the eastern terrane. This is characterized by steep slopes with more than 50% of which are within the 34-72° slope interval. It presents an irregular topography with alternating valleys and ridges. Slope aspect maps shows

5

that it is dominated by slopes facing the NNW-SSE and E-W directions. Correlation with existing regional maps revealed that the area is underlain by Late Oligocene to Middle Miocene sediments and limestones (Quebral et al., 1996).

Figure 2: DEM processing for geomorphology and structural analysis A) Unprocessed B) Hillshade C) Slope Gradient D) Slope Aspect

6

Figure 2: Superimposition of slope gradient and hillshade. Broken lines refer to the trace of the massif (Paguican, 2012)

3.2 Slope Statistics and Correlation with Geology

Slope gradient statistics show that the minimum value is 0° while the maximum value is 72°. The minimum value corresponds to the flat regions in the area, specifically in the valley plains of Maragusan and Mawab-Nabunuturan. Minimum slope values are also observed along major drainage systems and lake. The maximum slope value of 72° corresponds to the slopes of the mountains composing the western and eastern terranes bounding the prospect. The mean value of the slopes is 19° or gently dipping. This may indicate moderate to high degree of erosion across the prospect, or the presence of young volcanic ash deposits that drape and fill the topography. For the slope aspect, the minimum value calculated is 0° which corresponds to the north orientation while the maximum value is 360° which corresponds to the south orientation. N-S slope aspect values corresponds to E-W trending slopes. The mean value is 96° which means that the dominant slope orientation which is perpendicular to this direction would be NNE-SSW or almost N-S.

Slopes or inclined surfaces form in response to various geological processes such as erosion and faulting. Considering the highly active tectonics in the region, faulting is deemed as the most dominant influence for the development of slopes. The almost N-S trending slopes are mainly attributed to the presence of the similarly oriented splays of the Philippine Fault. These are represented by the Eastern Mindanao Fault which bounds the prospect at the west and Mati Fault at the east. The westward dipping subduction of the Philippine Sea Plate would also produce N-S trending structures that would be compressive in nature based on tectonic models. The subduction may also be responsible for the E-W trending slopes that may represent E-W trending structures. Tectonic models show that structures of this orientation would open up and become tensional in nature.

The figure belows presents the structural map, with structures delineated from combined inspection of slope aspect, slope gradient, and hillshade maps.

7

Figure 3: Remote sensing based structural Map of the Amacan geothermal prospect.

3.3 Thermal Mapping

Land surface temperature (LST) is defined as the skin temperature of the land surface or the temperature felt when the ground is touched (Avdan and Jovanovska, 2016). The LST map (Figure 5.E) was generated using a series of raster calculations as presented by Figures 5.A to 5.D. As shown in the final map, thermal highs are displayed in warmer reddish coloration while relatively colder areas are in blue. The maximum temperature calculated is about 31°C, which is similar with other land surface temperature maps in other geothermal areas (Meneses, 2015).

Figure 5.E also shows the documented locations of thermal manifestations in the area such as hot and warm springs according to the Philippine Department of Energy. A linear high temperature anomaly is observed near the Manat thermal springs. Similarly, the locations of the Leonard-Mainit-APEX, Manat-Bucal, and Amacan thermal areas also correlate well with the thermal anomalies. Other thermal areas with nearby high temperature anomalies are in Maragusan and Maraut. Contrarily, relatively colder areas are observed in the northwest, northeast, and central-south portion of the prospect. This may be due to thick vegetation or the lack of geothermal activity at the surface. The TIRS image was also captured in the morning given the sun elevation of 67°.

8

Figure 4: Raster calculations converting Band 10 to Land Surface Temperature. The prospect is highlighted by the blue boxes. (A) Top of Atmosphere Radiance (B) Brightness Temperature (C) Normalized Difference Vegetative Index (NDVI) (D) Land Surface Emissivity (E) Land Surface Temperature Map with documented thermal areas in red circles (DOE, n.d.). Legend shows temperature in degrees Celsius.

3.4 Hydrothermal Alteration Mapping

Previous studies explained the fact that certain minerals associated with hydrothermal processes, such as iron-bearing minerals (e.g., goethite, hematite, jarosite, and limonite) and hydroxyl bearing minerals (e,g. kaolinite and K-micas) show diagnostic spectral features that allow their remote identification (Hunt, 1980; Mia, 2012). Iron oxide is said to be a common constituent of alteration zones associated with hydrothermal sulfide deposits (Poormirzaee and Oskouei, 2009). The major iron oxide species, formed from the weathering of sulfides absorb energy at different frequencies providing a means of discrimination using hyperspectal scanners (Taranik et al., 1991). Hydroxyl bearing minerals on the other hand form the most widespread product of alteration.

Using OLI bands of Landsat 8, hydrothermal alteration maps were prepared as shown by Figure 6. Composite ratio operation 5:7:6 remotely identified areas with hydrothermal alteration in the study area. The spatial distribution of altered grounds is observed along major drainages where the reflectance is coming from either the altered exposed bedrock or fluvial deposits with altered material. Hydrothermal alteration is also evident in mining areas within the Masara Caldera and near thermal areas. Relatively wider distribution of hydrothermal alteration is observed along Manat-Bucal, Maraut, Mainit-Apex, Maragusan, Amacan, and around Lake Leonard.

On the other hand, Abrams ratio (6/7:4/3:5/6) was able to differentiate the type of alteration mineral present (Figure 6). Iron-oxide minerals such as limonite, hematite, and jarosite were discriminated as yellow to yellow-green while hydroxyl bearing minerals such as clays appeared as red. The purple coloration which dominates the resulting image corresponds to the background

9

that is possibly unaltered or vegetated with no exposures. As with band ratio operation, yellow to yellowish green areas are concentrated along major drainages, known thermal areas, and iron-oxide altered outcrops. The largest iron oxide altered area is in Mainit-APEX. Increasing the contrast of the image also enabled better accentuation of the clay altered regions in red. The largest clay altered area is observed in Amacan. Patches of clay alteration is also observed in Maraut and Lake Leonard.

Figure 5: Left: Composite ratio operation 5:7:6 showing areas that are hydrothermally altered. Right: Abrams ratio operation (6/7:4/3:5/6) distinguishing iron oxide (FeO) and clay (Cl) minerals

4. INTEGRATED REMOTE SENSING AND RECOMMENDED PRIORITY AREAS FOR EXPLORATION

Figure 7 superimposes the three generated images to present an integrated remote sensing map. The prospect is narrowed down into five priority areas: Area A – Lake Leonard-Mainit-APEX, B – Amacan, C- Manat-Bucal, D – Maraut, and E – Maragusan. These areas appear to indicate geothermal activity given the overlaps of thermal anomalies and alteration zones, as well as the presence of major structures. The map is dominantly purple colored due to the Abrams ratio (Figure 6) basemap. The temperature legend in the map shows that lighter regions are thermal highs while dark regions are thermal lows or no anomalies.

Area A as shown by Figure 8.A is centered on Leonard Caldera, postulated to be the youngest volcanic center in the area as it cross cuts the Masara Caldera. The Leonard Caldera is located in between the NW-SE trending Amacan and Manat Faults, where majority of the thermal manifestations are all confined. Area A hosts the NE-SW trending Lumanggang Fault in the west and parallel Maraut Fault in the east. High thermal anomaly and hydrothermal alteration composed of clay and iron oxides also coincide around the lake, most notably in its southern portion. West of the lake is a high iron oxide anomaly which measures about a kilometer in length and follows the trend of Lumanggang Fault. Minor overlaps between the high thermal and alteration anomalies are also noted within the southern half of the Masara Caldera which

10

encloses a gold-copper district. Peculiarly, no distinct anomalies are mapped in the reported Mainit thermal area.

Figure 6: Five priority areas identified for detailed exploration in yellow circles

About three kilometers south of Lake Leonard is Area B (Figure 8.B) which corresponds to the Amacan area. A large thermal anomaly of about two kilometers is observed which also coincides with dominant clay alteration. This anomaly area is transected by arcuate features that represent the eastern extent of the Masara Caldera. Permeability is also facilitated by the major Amacan Fault which cuts through the western side of the area. Area B is also located on the flanks of Ugos Dome which is a candidate heat source (DOE, n.d.).

Area C (Figure 8.C) is mainly delineated following the trace of the major NW-SE trending Manat Fault. High permeability from this structure is perceived considering the overlaps of the thermal anomalies and alteration zones following its trace. The remarkably linear thermal anomaly which measures for about two kilometers also hosts several documented hot springs in Mainit and Bucal areas. The hot springs and anomalies are also observed to occur at the intersections of the Manat Fault and the secondary NE-SW trending minor structures.

11

Area D (Figure 8.D) is located in Maraut at the southwestern portion of the Amacan prospect, where similar high temperature springs are also documented. Unlike Areas B and C, the anomalies do not follow the trend of the major Maraut Fault or the Maraut Collapse. Thermal anomalies occur as sporadic bodies around the area with alteration by clay and iron oxides. Area E (Figure 8.E) is also a recommended area for exploration in Maragusan town. The area hosts two reported warm springs, with nearby high clustering of iron oxide alteration enclosed in a high thermal anomaly body.

On the other hand, Figure 8.F shows an area in Mawab at the northwestern portion of the prospect showing generally no thermal anomalies, hydrothermal alteration, documented thermal manifestations, and major structures. This represents the background in the prospect showing no clear geothermal activity at the surface.

Figure 7: Close up maps of the five priority areas showing the overlap of thermal anomalies, hydrothermal alteration using Abrams ratio, and presence of structures. Green dots are the location of the thermal manifestations. Light purple regions are thermal highs while dark purple are relatively thermal lows. (A) Leonard-Mainit-APEX for Area A (B) Amacan for Area B (C) Manat-Bucal for Area C (D) Maraut River for Area D (E) Maragusan for Area E and (F) Mawab area showing no anomaly. FeO indicates iron oxide anomalies while Cl shows clay.

12

5. SUMMARY AND CONCLUSIONS

The study integrated known remote sensing techniques to determine priority areas for geothermal exploration in the Amacan Geothermal Prospect in Eastern Mindanao. Geomorphologic analysis from slope aspect, slope gradient, and hillshade identified two major deposits within the massif. The western and eastern terranes bounding the massif also show different geomorphologic signatures relative to the volcanic complex.

Lineament analysis revealed that the major structures are oriented NW-SE, NE-SW, N-S, and E-W. These structures are resultant from the activity of the Philippine Fault in the area associated with the ongoing subduction along the Philippine Trench. The N-S and E-W structures also conform to the minimum, mean, and maximum slope values calculated.

Thermal mapping showed areas of high temperatures relative to the background, especially near thermal areas and open grounds with mining activity or lack of vegetation. The highest temperature measured is 31°, similar with other geothermal areas in the Philippines. Hydrothermal alteration mapping was also able to discriminate iron oxide and clay alteration. Based on the integrated remote sensing results, the large prospect is narrowed down into five priority areas for detailed geoscientific assessments: Leonard-Mainit-Apex, Amacan, Manat-Bucal, Maraut, and Maragusan.

Overall, remote sensing proved its efficiency as a tool in determining areas with high geothermal potential and activity. This study however stresses the importance of ground validation to continuously improve and calibrate the remote sensing process. Given that the area also hosts mining districts, the anomalies may not just reflect the current hydrothermal system but from the fossil systems as well. Statistics also played a role in providing the most dominant slope orientations, which possibly represents the most dominant structures as well.

REFERENCES

Aurelio, M.A. (2000), Tectonics of the Philippines revisited. Journal of the Geological Society of the Philippines, 55, 119−183

Avdan U. and Jovanovska G. (2016) Algorithm for Automated Mapping of Land Surface Temperature Using LANDSAT 8 Satellite Data; Journal of Sensors, Volume 2016

Braganza, J. (2014). Volcano-tectonic evolution of the Bacon-Manito Geothermal Project, Sorsogon, Philippines (MSc Thesis) The University of Auckland

Buhari, U. (2015). Landsat 8: Estimating Land Surface Temperature Using ArcGIS. Retrieved September 22, 2017, from https://www.youtube.com/watch?v=uDQo2a5e7dM

Buckley, A. (2010). Understanding curvature rasters / ArcGIS Blog., 2017, Retrieved from http://blogs.esri.com/esri/arcgis/2010/10/27/understanding-curvature-rasters

Department of Energy. (n.d.). Amacan Geothermal Prospect. Retrieved July 1, 2019, from https://www.doe.gov.ph/amacan-geothermal-prospect

Hunt G. R. (1979) Near Infrared (1.3–2.4 μm) spectra of alteration minerals – potential for use in remote sensing; Geophysics 44 1974–1986.

Jensen J. R. (1996) Introductory digital image processing: A remote sensing perspective; 2nd edition, Prentice Hall Series in Geographic Information Science; 318p.

Lallemand, S., Popoff, M., Cadet, J.-P., Deffontaines, B., Bader, A.G., Pubellier, M. and Rangin, C., (1998) The junction between the central and southern Philippine Trench. Journal of Geophysical Research, 103, 933−950.

Meneses, S.M., (2015). Thermal Remote Sensing at Leyte Geothermal Production Field using Mono-window Algoriths. Proceedings World Geothermal Congress. Melbourne, Asutralia.

13

Mia, B., & Fujimitsu, Y. (2012). Mapping hydrothermal altered mineral deposits using Landsat 7 ETM+ image in and around Kuju volcano, Kyushu, Japan. Journal Earth System Science , 1049-1057.

Paguican E.M.R. (2012). The structure, morphology, and surface texture of debris avalanche deposits: field and remote sensing mapping and analogue modelling. Earth Sciences. Université Blaise Pascal - Clermont-Ferrand II.

Paguican E.M.R and Bursik M.I. (2016) Tectonic Geomorphology and Volcano-Tectonic Interaction in the Eastern Boundary of the Southern Cascades (Hat Creek Graben Region), California, USA. Front .EarthSci.4:76.

PHIVOLCS (2015). Distribution of active faults and trenches in the Philippines

Quebral, R., Pubellier, M. and Rangin, C., (1996). The onset of movement on the Philippine fault in eastern Mindanao: A transition from a collision to strike slip environment. Tectonics, 15, 713−726.

Taranik D. L., Kruse F. A., Goetz A. F. H. and Atkinson W.W. (1991) Remote sensing of ferric iron minerals as guides for gold exploration; Proceedings Eighth Thematic Conference on Geologic Remote Sensing, Denver, Colorado, pp. 197–228.

US Geological Survey (2016). Landsat 8 Data Users Handbook

Wood, P. (1980). A Report on a visit to the Philippines, February 1980. Unpublished report. NZ DSIR, 1980.

Yumul G.P., Dimalanta C.B., Maglambayan, V.B., and Marquez, E.J. (2008). Tectonic setting of a composite terrane: A review of the Philippine arc system. Geosciences Journal Vol. 12, No. 1, p. 7-17