Comparison of satellite-derived and in-situ observations of ice and snow surface temperatures over Greenland

Abstract

The most practical way to get spatially broad and continuous measurements of the surface temperature in the data-sparse cryosphere is by satellite remote sensing. The uncertainties in satellite-derived LSTs must be understood to develop internally-consistent decade-scale land surface temperature (LST) records needed for climate studies. In this work we assess satellite-derived "clear-sky" LST products from the Moderate Resolution Imaging Spectroradiometer (MODIS) and the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), and LSTs derived from the Enhanced Thematic Mapper Plus (ETM+) over snow and ice on Greenland. When possible, we compare satellite-derived LSTs with in-situ air temperature observations from Greenland Climate Network (GC-Net) automatic weather stations (AWS). We find that MODIS, ASTER and ETM+ provide reliable and consistent LSTs under clear-sky conditions and relatively-flat terrain over snow and ice targets over a range of temperatures from −40 to 0°C. The satellite-derived LSTs agree within a relative RMS uncertainty of~0.5°C. The good agreement among the LSTs derived from the various satellite instruments is especially notable since different spectral channels and different retrieval algorithms are used to calculate LST from the raw satellite data. The AWS record in-situ data at a "point" while the satellite instruments record data over an area varying in size from: 57 × 57 m (ETM+), 90 × 90 m (ASTER), or to 1 × 1 km (MODIS). Surface topography and other factors contribute to variability of LST within a pixel, thus the AWS measurements may not be representative of the LST of the pixel. Without more information on the local spatial patterns of LST, the AWS LST cannot be considered valid ground truth for the satellite measurements, with RMS uncertainty~2°C. Despite the relatively large AWS-derived uncertainty, we find LST data are characterized by high accuracy but have uncertain absolute precision.

Using MODIS land surface temperatures and the Crocus snow model to understand the warm bias of ERA-Interim reanalyses at the surface in Antarctica

“…By applying this method on 42 sites, Wan (2014) found MODIS T s errors within ±2 • C for all the sites but six bare soil sites (not including South Pole). For the South Pole site, MODIS LST error is only −0.5 • C. The accuracy of MODIS LST values depends primarily on the quality of the detection of clouds (Hall et al, 2008). When clear-sky conditions are detected, the generalized split-window land surface temperature algorithm of Wan and Dozier (1996) is used to retrieve LST values for each MODIS pixel along with emissivities in bands 31 (10.78 to 11.28 mum) and 32 (11.77 to 12.27 µm) (http:// modis.gsfc.nasa.gov/about/specifications.php).…”

Impact of MODIS sensor calibration updates on Greenland Ice Sheet surface reflectance and albedo trends

“…Our analysis investigates Terra and Aqua records from three product types, namely MOD09/MYD09, MOD10/MYD10, and MCD43 for the entire summer data record, though it does not directly compare MODIS C5 and C6 data with ground data. Several MODIS cryospheric calibration and validation investigations, some of which include in situ data, have been performed, such as Stroeve et al (2005Stroeve et al ( , 2013, Moody et al (2007), Hall et al (2008), Alexander et al (2014), Wright et al (2014), Polashenski et al (2015), and Zhan and Davies (2016). Recently, found Terra MOD10A1 albedo substantially improves in relative accuracy from C5 to C6, agreeing with the Greenland Climate Network and Programme for Monitoring of the Greenland Ice Sheet station data from midMay through August for the majority of GrIS south of 80 • N within 0.04 (unitless albedo from 0-1, see Box et al, 2017, Fig.…”

Feasibility of improving a priori regional climate model estimates of Greenland ice sheet surface mass loss through assimilation of measured ice surface temperatures

“…The reduction in uncertainty is evident by comparing the base of the posterior histogram with that from the prior estimates. The positive update by the DA system can be explained based on the fact that IST and air temperature are coupled and each one affects the other (Hall et al, 2008). Figure 3b illustrates that the median of the posterior estimate of γ l agrees well with the corresponding synthetic truth.…”