Please cite this article in press as: Feng, J., et al., Comparison of lithological mapping results from airborne hyperspectral VNIR-SWIR, LWIR and combined data. Int. J. Appl. Earth Observ. Geoinf. (2017), http://dx.doi.org/10.1016/j.jag.2017.03.003 ARTICLE IN PRESS G Model JAG-1403; No. of Pages 14 International Journal of Applied Earth Observation and Geoinformation xxx (2017) xxx–xxx Contents lists available at ScienceDirect International Journal of Applied Earth Observation and Geoinformation journal homepage: www.elsevier.com/locate/jag Comparison of lithological mapping results from airborne hyperspectral VNIR-SWIR, LWIR and combined data Jilu Feng a , Derek Rogge b , Benoit Rivard a,∗ a 1-26 Department of Earth and Atmospheric Sciences, University of Alberta, Edmonton, T6G2E3, Canada b German Remote Sensing Data Center, DLR, Munchnerstr. 20, D-82234, Germany, Germany a r t i c l e i n f o Article history: Received 31 October 2016 Received in revised form 6 March 2017 Accepted 10 March 2017 Available online xxx Keywords: Short wave infrared Long wave infrared Wavelet analysis Airborne hyperspectral Geological mapping a b s t r a c t This study investigates using the Airborne Hyperspectral Imaging Systems (AISA) visible and short- wave infrared (SWIR) and Spatially Enhanced Broadband Array Spectrograph System (SEBASS) longwave infrared (LWIR) (2 and 4 m spatial resolution, respectively) imagery independently and in combination to produce detailed lithologic maps in a subarctic region (Cape Smith Belt, Nunavik, Canada) where regionally metamorphosed lower greenschist mafic, ultramafic and sedimentary rocks are exposed in the presence of lichen coatings. We make use of continuous wavelet analysis (CWA) to improve the radiometric quality of the imagery through the minimization of random noise and the enhancement of spectral features, the minimization of residual errors in the ISAC radiometric correction and target tem- perature estimation in the case of the LWIR data, the minimization of line to line residual calibration effects that lead to inconsistencies in data mosaics, and the reduction in variability of the spectral con- tinuum introduced by variable illumination and topography. The use of CWA also provides a platform to directly combine the wavelet scale spectral profiles of the SWIR and LWIR after applying a scalar correc- tion factor to the LWIR such that the dynamic range of two data sets have equal weight. This is possible using CWA as the datasets are normalized to a zero mean allowing spectra from different spectral regions to be adjoined. Lithologic maps are generated using an iterative spectral unmixing approach with image spectral end- members extracted from the SWIR and LWIR imagery based on locations defined from previous work of the study area and field mapping information. Unmixing results of the independent SWIR and LWIR data, and the combined data show clear benefits to using the CWA combined imagery. The analysis showed SWIR and LWIR imagery highlight similar regions and spatial distributions for the three ultramafic units (dunite, peridotite, pyroxenite). However, significant differences are observed for quartz-rich sediments, with the SWIR overestimating the distribution of these rocks whereas the LWIR provided more consis- tent results compared with existing maps. Both SWIR and LWIR imagery were impacted by the pervasive lichen coatings on the mafic rocks (basalts and gabbros), although the SWIR provided better results than the LWIR. Limitations observed for the independent data sets were removed using the combined spec- tral data resulting in all geologically meaningful units mapped correctly in comparison with existing geological maps. © 2017 Elsevier B.V. All rights reserved. 1. Introduction Regional lithological mapping in Northern regions is costly and challenging owing to poor accessibility and a short working sea- son. Remote sensing is particularly suited to assist in mapping of expansive regions, such as northern Canada. Satellite and airborne ∗ Corresponding author. E-mail addresses: jfeng@ualberta.ca (J. Feng), derek.rogge@dlr.de (D. Rogge), benoit.rivard@ualberta.ca (B. Rivard). imaging systems are becoming more readily available and can be used in support of regional mapping and mineral exploration. This study makes use of a unique data set that comprises visible (VIS), short wave infrared (SWIR) and long wave infrared (LWIR) airborne hyperspectral imagery to assess improved mapping capa- bilities over a geologically interesting and economically important sub arctic region in Nunavik, Canada. Mapping the distribution and compositional diversity of the lithologies in the region, specifically the ultramafic rocks (e.g. dunite, peridotite, olivine pyroxenite, and pyroxenite) which are host to Nickel-Copper-Platinum Group Ele- http://dx.doi.org/10.1016/j.jag.2017.03.003 0303-2434/© 2017 Elsevier B.V. All rights reserved.