3 The study area covers mapsheet NTS 87H/05 and is within the western Minto Inlier on Victoria Island (Figure 1). The Minto Inlier, first recognized and mapped by Thorsteinsson and Tozer (1962), is a northeast-trending belt of early Neoproterozoic sedimentary rocks of the Shaler Supergroup, which are intruded by gabbro sills and dykes capped by coeval flood basalt. The succession was gently folded forming an open syncline and a smaller anticline before deposition of an unconformably overlying shallow marine succession of Cambro-Ordovician age. The Shaler Supergroup is a ca. 4-km-thick succession of shallow marine carbonate rocks, shallow marine and fluvial clastic rocks and peritidal evaporites that were deposited in a very broad and shallow intracontinental sea, the Amundsen Basin (Young 1981). Other remnants of this basin are preserved in other inliers along the northern mainland coast and Banks Island and these can be traced to similar exposures in the northern Cordillera (Rainbird et al. 1996; Long et al. 2008). The maximum age of the Shaler Supergroup is ~1000 Ma based on detrital zircon geochronology (Rainbird et al. 1992; 1997) and its minimum age is 720 Ma, the age of the gabbro sills that intrude it (Heaman et al. 1992). The Shaler Supergroup includes, in ascending stratigraphic order, the Rae Group, Reynolds Point Group, Minto Inlet Formation, Wynniatt Formation, Kilian Formation and Kuujjua Formation (Figure 1; Rainbird et al. 1994). The Minto Inlet, Wynniatt and Kilian formations are exposed in the study area and are described, in detail, in Table 1.

4 The Minto Inlier and overlying Cambro-Ordovician rocks are cut by a prominent NE-striking fault zone, present in the northern half of the study area.

5 At about 720 Ma, mantle-derived mafic magmas of the Franklin event were injected into a broad area of the Canadian Shield from northern Baffin Island to the western Arctic islands (Jefferson et al. 1994) manifested as an extensive NW-trending dyke swarm (Heaman et al. 1992). In the Minto Inlier, the Franklin magmas were intruded mainly as 5–50m-thick sills between sedimentary rocks of the Shaler Supergroup. This geologic setting is reported to be favourable for Ni, Cu and PGE (Fahrig et al. 1971; Heaman et al. 1992). Jefferson et al. (1994) compared the Neoproterozoic Franklin event of Arctic Canada with the Permo-Triassic Noril’sk-Talnakh Ni–Cu–PGE deposits of Russia. They rated a high potential for Ni–Cu–PGE mineralization based on similarities in the scale and duration of the magmatic events, host rock types, volcanic stratigraphy and petrology, tectonic setting, associated major faults, sources of S, Se, and As, magmatic differentiation, geophysical anomalies and other features. In particular, the northeast-striking fault zone, described above could be an analogue to the Noril’sk-Kharaelakh Fault which is thought to be critical to localization of the massive sulphide deposits at Noril’sk (Jefferson et al. 1994).

6 The Minto Inlier has also been reported to have high potential for other types of Cu mineralization, such as Kuperschiefer and volcanic Red Bed (Rainbird et al. 1992). The potential for carbonate-hosted zinc-lead (MVT) mineralization, similar to that in the Mackenzie Mountains (Dewing et al. 2006) exists in dolomitized carbonates of the Wynniatt Formation.

7 The very high spatial resolution and stereo capability of GeoEye-1 imagery were utilized to create a predictive geology map of the study area. The GeoEye-1 pushbroom satellite, formerly known as OrbView-5, was launched on September 2008 from Vandenburg Air Force Base in California. The satellite orbits the Earth 15 times per day in a near-polar, sun-synchronous orbit at an altitude of 681 km with a nominal equator crossing at 10:30 AM local. GeoEye-1 simultaneously collects panchromatic imagery at 0.41 m pixel resolution and multispectral (VNIR) imagery at 1.65 m resolution (GeoEye 2009). The nominal swath width of the satellite is 15.2 km at nadir and the revisit time is less than three days. Table 2 presents the main specifications of the GeoEye-1 imagery (GeoEye 2009).

8 Remotely sensed images can be used for stereoscopic viewing provided the two images forming a stereo pair are geometrically compatible (Gupta 2003). A remote sensing satellite with stereo-imaging capability is able to capture images in stereo mode using either across-track or along-track configurations. Along-track stereo pairs (e.g. ASTER imagery) are images of the same location taken from two different perspectives during a single pass of the satellite. Across-track images (e.g. SPOT imagery) are typically taken from a location by the same sensor on multiple orbits. GeoEye-1, being capable of rotating or swivelling forward, backward or side-to-side can capture images in any direction (up to 60° off nadir, along- and across-track). The GeoEye stereo pairs are collected on the same orbital pass (along-track). This results in superior image quality because of a short time span between the two images resulting in the same atmospheric lighting conditions and scene content, as well as better metric accuracy. Each stereo pair contains an image collected at a low elevation angle (above 60°) as well as an image collected at a higher elevation angle (above 72°) with 30°−45° convergence (0.54−0.83 base-to-height ratio (Eurimage 2010). A Rational Polynomial Coefficient (RPC) camera model maps the respective ground coordinates (latitude, longitude, height) to image coordinates (line, sample).

9 The GeoEye dataset comprises four stereo pairs acquired from June 27 to July 27, 2009, with each stereo pair collected at the same time. An orthorectified mosaic covering the NTS87H/05 was also used for 2D interpretations. A DEM with a resolution of 15 m, derived from the stereo GeoEye imagery, was also produced. The DEM was used to orthorectify the GeoEye imagery but was not used in the interpretation process. The GeoEye multispectral bands were resampled to 1 m resolution using a bi-linear interpolation algorithm in both image types.

10 A LANDSAT-7 ETM ⁺ image, acquired on July 14, 2002, together with a SPOT-5 image acquired on August 22, 2009, were used in concert with the GeoEye-1 images to provide higher spectral resolution compared to GeoEye-1 imagery (Table 2). The SPOT image had a small gap for the northwest part of the study area. The satellites are both well known and well-used in various geoscience applications. More details on these sensors can be found in Lillesand and Keiffer (2000), Drury (2001) and Harris (2008).

15 The resulting predictive map (Fig. 5b) presents much more detailed information compared to the existing 1:500,000 scale geological map of the area (Fig. 5a; Thorsteinsson and Tozer 1962; Hulbert et al. 2005). The main lithological units exposed in the study area are the Minto Inlet, Wynniatt, and Kilian formations of the Shaler Supergroup, along with gabbro sills and dykes of Franklin magmatic event and Paleozoic succession of CambroOrdovician age. With the aid of recent reconnaissance-level field observations and previous mapping (Rainbird, 1998), one of the main map units, the Wynniatt Formation, was divided into four sub-units (members) and these are recognized in the GeoEye images throughout most of the map area.

16 The LANDSAT and SPOT SWIR bands proved to be superior for distinguishing gabbro sills and dykes from sedimentary rocks of the Shaler Supergroup. The SPOT image because of the one SWIR band and 10 m spatial resolution, proved to be superior for capturing small outcrops of gabbro, which were not detected on the LANDSAT and GeoEye images. Thus the LANDSAT and SPOT data were employed as complimentary sources of lithologic information for producing the predictive map. In the areas covered by the glacial till, other overburden or vegetation, the contacts between the units were inferred by extrapolation and expert knowledge. Table 3 presents the spectral and spatial characteristics of each lithological unit (derived from the existing geology maps; Figs.1 and 5a and recent field studies) for the LANDSAT, SPOT and GeoEye images respectively.

17 The high spatial and moderate spectral resolution of the GeoEye imagery enabled the identification of a black shale unit (black shale member) characterized by low reflectance (Figure 6) and to distinguish (by subtle spectral and textural differences) parallel-stratified to massive stromatolitic dolostones (stromatolitic carbonate member) from underlying and overlying dolostones containing fine-grained interlayers (lower and upper carbonate members, respectively). Figures 6 and 7 are bands 4,3,2 RGB colour composite GeoEye images of small portions of the study area (boxes A and B in Fig. 3a) which provide invaluable structural and lithological details of the Wynniatt Formation and the gabbro sills. As well, an important distinction could be made between Proterozoic sedimentary strata and unconformably overlying interlayered sandstone and carbonate rocks of Cambro-Ordovician age (Fig. 8; Box C in Fig. 3a), which is not indicated in the existing geology map (Fig. 5a) (see Fig. 8d) The high spatial resolution of GeoEye images also facilitated the identification and mapping of the thin (up to 1 m thick) sandstone interlayers in the carbonate rocks of Cambro-Ordovician succession. Figure 8-a, b and c compare the spatial resolution of GeoEye, LANDSAT and SPOT images. The very high spatial resolution of the GeoEye image, despite its lower spectral resolution compared to LANDSAT and SPOT images, provides more detailed textural information, which is very useful for identifying different rock types and geologic structures (see Fig. 8d). While generating the remote predictive map, the Cambro-Ordovician succession was not sub-divided even though in some areas different sub-units could be clearly distinguished on the GeoEye Imagery (Figs. 8a and 9b). This was because, initially, the Paleozoic rocks were not the focus of our study, but after work in the field, we realized that they occupied a much more significant part of the map area. Furthermore, in many areas, Cambrian carbonates were visually very similar to Proterozoic carbonates and could only be separated based on their fossil content.

18 Figure 9a is a 4,3,2 GeoEye RGB colour composite image that provides a more regional view of the Cambro-Ordovician rocks. The carbonate unit (labelled tan dolostone member) is shown in cream and has a clear mappable contact with the lower clastic member containing quartz-arenite interlayers. Figure 9b is also a 4,3,2 GeoEye RGB colour composite image covering the northeast corner of the map area showing some details of the Cambro-Ordovician strata.

19 A field mapping program was carried out in 2010 and 2011 over the study area under the Geo-mapping for Energy and Minerals (GEM) program, Natural Resources Canada. The bedrock geology map (Fig. 10) was produced largely from field observations using SPOT imagery as a visual mapping guide. The field (Fig. 10) and predictive maps (Fig. 5b) look similar in terms of the regional distribution of lithological units but the field map provides important updates to the predictive map particularly in areas where a thin veneer of Quaternary surficial deposits masks defining spectral attributes.

20 One of the main differences between the two maps is the subdivision of Cambro-Ordovician rocks into subunits on the field map. This precludes a quantitative comparison of the two maps using cross-tabulation and as such our comparison is based on a visual assessment of the maps on a unit-to-unit basis as discussed in the next sections on a formation basis.

21 The lithological contact between the evaporite rocks of Minto Inlet Formation and the lower carbonate member (Wymiatt Formation) is covered by vegetation in most areas (Fig. 11) and locally by cloud in the GeoEye image. There are scattered exposures that are distinguished only in LANDSAT and SPOT images. Figure 11a is a 7,4,2 LANDSAT RGB colour composite image of the western exposure of these units (Box A in Fig. 10) that shows the Minto Inlet in bluish tone in scattered exposures. There was no SPOT coverage for this part of map area. Figure 11b is a 4,3,2 GeoEye RGB colour composite image of the same area. Despite the higher spatial resolution there is no subtle spectral difference between the rock units. The scattered exposures were used to infer the lithological contact for the covered areas for the rest of map area, which accounts for the small disagreement between the predictive and field maps. Otherwise the two maps are in good agreement regarding this lithological unit.

22

23 The level of agreement between the field and predictive maps is high with respect to the Wynniatt Formation carbonate rocks. However, on the predictive map, the upper carbonate member has been incorrectly mapped as a stromatolitic and lower carbonate member, places. This is because the spectral response of the carbonate members is very similar. The very high spatial resolution of the GeoEye imagery show subtle textural differences between the massive stromatolitic dolostone and the fine-grained dolostone layers (Fig. 7), which led us to map the stromatolitic dolostone as the stromatolitic member of Wynniatt Formation. Field work revealed that these strata actually belong to the upper carbonate member, which also contains stromatolitic biostromes. The black shale member is also less extensive on the field map in the central-northern part of the map area compared to the predictive map. Again, most of the difference between the predictive and field maps occurs in areas covered by glacial/surficial material.

24 The evaporite and carbonate rocks of the Kilian Formation are mappable on the LANDSAT and SPOT images because of the higher spectral resolution of these images compared to that of the GeoEye imagery. Figure 12a is a 4,3,2 SPOT RGB colour composite image (Box B in Fig.10) showing these rocks in purple and bluish colour. Figure 12b is a 4,3,2 GeoEye RGB colour composite image for the same area. The Kilian Formation is covered with the glacial till especially in the western parts of the map area for which the lithological contact was inferred by extrapolation. Comparison of the predictive and field maps shows some discrepancy especially in areas covered by surficial materials.

25 One of the main contributions of the field work was subdivision of the Cambro-Ordovician rocks. These units in the map area include, in ascending stratigraphic order, the lower clastic member, tan dolostone member, and stripy member of Cambrian age, and Victoria Island Formation of Ordovician age. (K. Dewing and T. Hadlari, personal communications, 2011). The lower clastic and tan-dolostone members are well exposed in the north-central part of the map area where they unconformably overlie Proterozoic rock units (Figs. 8 and 9a). These rock units are covered by surface material in surrounding areas although some outcrops appear in the northern part and northeast corner of the map area (Fig. 9b). Thus the concealed contact was inferred using scattered outcrops in many areas. These units were not mapped on the predictive map despite clear differences in spectral response because the original focus of the mapping was on the Proterozoic rocks. The SWIR bands (LANDSAT and SPOT) were particularly useful in distinguishing the tan-dolostone unit from Wynniatt Formation carbonates. The other exposure of Cambro-Ordovician rocks occurs in the western part of the map area where they are in fault contact with the upper carbonate unit of the Wynniatt Formation. These areas are covered mostly by vegetation and, unfortunately, some cloud (in the GeoEye imagery) and were mapped as undifferentiated Wynniatt Formation as they were surrounded by Wynniatt Formation carbonate.

26 Gabbro sills and dykes provide the highest level of agreement between the predictive and field maps. They are easily mapped on the remotely sensed data (LANDSAT, SPOT especially) as their spectral characteristics in the SWIR bands are very different (moderately reflective) from those of other rock types. There are only a few areas in which the sills are unrecognized either on the predictive map or on the field map. Areas, in which the sills are unrecognized on the predictive map, always coincide with the areas covered with either vegetation or Quaternary deposits. In a few cases some small sills were not identified on the field map. These sills were, in fact, added to the field map after re-examining the remotely sensed images. The SPOT imagery, because of higher spatial resolution, was particularly useful in capturing the small sills which could not be captured on the LANDSAT image. Also, outcrops south of Minto Inlet, which are mapped as Victoria Island Formation carbonate on the field map (Fig. 10), were originally mapped as gabbro sills based on the LANDSAT and GeoEye imagery and latter were corrected to the stromatolitic member of Wynniatt Formation using SPOT image (Fig. 5b).