| Summary: | A commercially-available remotely controlled boat (http://www.viperfishing.co.uk/Site/Viper Storm 2.html) was customized and equipped with a GPS/sonar (HDS-5 Lowrance), an above-surface irradiance sensor for incoming solar radiation, a below-surface downward looking radiance sensor, a spectrometer (Ocean Optics) and a micro computer (used to synchronize the data from the different instruments). The boat is made of glossy acrylic capped ABS and propelled by dual jet-pump engines. Lake depth was measured by means of a 50/200 kHz transducer. The depth uncertainty was estimated to be of the order of 20–30 cm by testing the set up in a swimming pool before the shipping and deployment to Greenland. However, the uncertainty on depth might be higher than the one obtained in the pool because of the roughness of the bottom of the supraglacial lake, which might be responsible for multiple echos. The boat was deployed 2-5 July 2010, on the west margin of the GrIS (Lake Olivia, 69 3603500 N, 49 2904000 W), with deployment time usually occurring when the sun was at zenith and measurements lasting for a few hours. As the boat average speed was 1 m/sec and the GPS data were recorded every second, we estimate a spatial resolution of 1m for the in-situ data. The spectral (450–1050 nm, with a 3 nm resolution) and depth data collected between two subsequent GPS acquisitions were averaged and assigned to the first of the two GPS locations. The total number of samples used in our analysis is 6000. Though the boat here used is much smaller than manned boats and casts a relatively small shadow, we acknowledge that this might still be a source of uncertainty. During the maneuvering of the boat, we paid attention in avoiding that the boat would not cast a shadow across the instrumental field-of-view while driving it. Another mitigating factor includes the use of the arm on the side of the boat in order to position the spectral sensor at a certain distance from the hull. The irradiance and radiance sensors were cross-calibrated in the laboratory following Mueller (2003). The diffuse irradiance leaving the outlet of an integrating sphere, illuminated by a calibrated xenon lamp, was measured by the irradiance sensor directed at the outlet. Radiance measurements were then taken of a Spectralon© target (99% reflectance) put in place of the irradiance sensor. These measurements were used to derive reflectance from irradiance and radiance intensity measurements and cross-calibrate the two sensors. See this citation for data visualization and interpretation: Tedesco, M. and N. Steiner. "In-situ multispectral and bathymetric measurements over a supraglacial lake in western Greenland using a remotely controlled watercraft." The Cryosphere, 5, 445–452, 2011 www.the-cryosphere.net/5/445/2011/doi:10.5194/tc-5-445-2011 These spectral data (2226 points from 2 & 3 July 2010) were also convolved with the relative spectral response functions of a range of multispectral sensors (Landsat 7 ETM+, Landsat 8 OLI, MODIS, ASTER, WorldView-2) to emulate satellite reflectance retrievals. In addition, sensor-specific dynamic range (brightest and darkest possible measurements) and radiometric resolution (8-bit or 12-bit) were used to create the most realistic emulation possible. Stated another way, minimum and maximum reflectances were imposed, and reflectance values were rounded with the appropriate precision. Data citation: Tedesco, M., N. Steiner, and A. Pope (2015): In situ spectral reflectance and depth of a supraglacial lake in Greenland. UCAR/NCAR - CISL - ACADIS. Dataset. http://dx.doi.org/10.5065/D6FQ9TN2
|
|---|