Mapping the Mankayan Mineral District Using Satellite Thermal & Synthetic Aperture Radar (SAR)
The Mankayan mineral district on northern Luzon, Philippines, is a remarkably prolific collection of mineral deposits of different styles. Several significant ones occur in a tight 25 km2 region of interest including several Cu-Au porphyry and high sulfidation epithermal deposits, an intermediate sulfidation epithermal Au-Ag vein deposit and an epithermal Au-Ag vein deposit as well as several exploration targets. A comprehensive review paper by Chang et. al. (2011) provides an overview of geology and traditional exploration techniques applied.
Porphyries represent large accumulations of economically import metals and are often associated with extensive alteration signatures. Which makes them attractive targets for remote sensing exploration which can cover large areas quickly and at a much lower cost than geophysical surveys, geochemical sampling and fieldwork. Remote sensing is routinely used by explorationists to take a first look at a region.
At Mankayan, several of the deposits are associated with quartz-alunite alteration halos. Such halos have been used to vector into mineralized zones with the mapping of chlorite and epidote in Indonesia presented by Neal et. al. (2018) just one recent example that has generated much interest among explorationists.
Neal et. al. propose short-wave infrared (SWIR) reflectance spectroscopy to detect and characterise hydrothermal alteration. SWIR spectroscopy can identify not only mineral species but also changes in the major element composition of minerals. Although it is worth noting that several minerals of economic importance lack diagnostic SWIR features. Quartz is a good example.
Nevertheless, SWIR reflectance spectroscopy has been widely applied to map sericitic (phyllic), argillic and advanced argillic alteration domains because it is particularly effective in discriminating bright clay minerals. Many Chilean porphyries were discovered by mapping “clay band” (2.08 - 2.35 µm at 30 m spatial resolution) anomalies in Landsat 4 in the early 1980s.
Propylitic alteration has been less used because propylitic rocks are usually dark and produce relatively poorly-defined spectra with noise often an issue on account of low signal levels. At Mankayan, the exploration problem is compounded by the presence of significant vegetation cover which raises issues for traditional visible, near and shortwave infrared remote sensing techniques which sense only the top millimeter of the earth’s surface.
Longwave infrared cameras achieve some penetration thanks to the emissivity property of minerals. In particular, quartz, alunite and pyrophyllite may be mapped beneath vegetation thanks to their distinctive LWIR signatures while the wavelength shift of emittance maxima of K- and Na-alunite may be used as a geochemical vector to mineralization.
We processed an Aster image collected over Mankayan on 1 November 2001. The area covered by a single Aster scene is 60 Km (rows) and 60 km (columns) and extensive vegetation is present (red tones in Fig. 1 below). The region of interest is the bright square.
Figure 1. Aster visible near infrared false colour composite over Makayan
While the spatial resolution of Aster visible bands is 15 m it is only 90 m in the thermal and consists of 5 bands imaged at 8.2910, 8.6340, 9.0750, 10.6570 and 11.3180 microns. The first step in estimating mineral abundances is to separate reflectance from emission effects in the imagery as these data were collected during daytime. 16 spectral endmembers were derived for the image as it is assumed that each 90x90 m parcel of ground is a nonnegative linear combination of 16 pure endmembers. 16 is an arbitrary number, chosen to account for likely geological spectral variation in a scene.
Each pixel is then expressed as a sum of 16 spectral abundances, most of which will be zero as they are estimated in such a way as to produce a sparse representation of the 5 dimensional data in 16 dimensional space. Each endmember hopefully corresponds to a geological meaningful unit and interpretation consists of the process of identifying these endmembers.
Two endmembers have very good correlation (0.99 and 0.93) with K-alunite and Na-alunite spectra collected in a laboratory and resampled to the Aster spectral bandpasses, as shown in Fig. 2.
Figure 2. Image and laboratory alunite spectra
The corresponding mineral abundances are plotted below and agree well with the geology map.
Figure 3. Na-alunite abundance distribution derived by unmixing Aster thermal imagery
Figure 4. K-alunite abundance distribution derived by unmixing Aster thermal imagery
A further complicating factor for satellite remote sensing exploration is cloud cover. Clouds and cloud shadows are especially a problem in equatorial regions. Satellite synthetic aperture [SAR] L-band radar sees through vegetation thanks to the 23.5 cm wavelength of the interrogating wave.
Multipolarized SAR data may be used to estimate the dielectric constants of the scattering target (Pendock, 2018) which allows pyrite and other metallic minerals to be mapped beneath vegetation. In particular it is a useful tool for mapping mineralized quartz veins obscured by vegetation.
Figure 5. Electrical conductivity estimated from ALOS SAR imagery
If we focus in on rocks with the largest conductivities, we find targets spatially coincident with many of the mapped deposits.
Figure 6. Conductivity anomalies
Quartz and alunite have dielectric constants between 6 and 7 while pyrite is over 33 so we are likely mapping the pyrite in the various units as conductivity anomalies.
Interpreted in conjunction with mineral abundances derived from thermal imagery, SAR is a useful tool for exploration, especially since its spatial resolution of less than 10 m allows for sharper definition of targets.
 Chang, Z., Hedenquist, J. W., White, N. C., Cooke, D. R., Roach, M., Deyell, C. L. and Cuison, A. L. (2011). Exploration Tools for Linked Porphyry and Epithermal Deposits: Example from the Mankayan Intrusion-Centered Cu-Au District, Luzon, Philippines. Economic Geology, 106(8), 1365-1398. doi:10.2113/econgeo.106.8.1365
 Neal, J., Wilkinson J.J., Mason, P.J and Chang, Z. (2018). Spectral characteristics of propylitic alteration minerals as a vectoring tool for porphyry copper deposits. J. Geochemical Exploration, 184 (part A), 179-198. https://doi.org/10.1016/j.gexplo.2017.10.019
 Pendock, N (2018). Satellite Synthetic Aperture Radar for Mine and Tailings Dam Monitoring. Proc. Copper Cobalt Africa, Livingstone, Zambia, SAIMM.