Validation of AVHRR- and MODIS-derived albedos of snow and ice surfaces by means of helicopter measurements

Abstract

We describe the validation of surface albedos of snow and glacier ice as derived from Advanced Very High Resolution Radiometer (AVHRR) and MOderate Resolution Imaging Spectrometer (MODIS) satellite data. For this purpose we measured surface albedos from a helicopter over Vatnajokull, Iceland, and the Kangerlussuaq transect (western part of the Greenland ice sheet) in Thematic Mapper (TM) bands 2 and 4 and AVHRR bands 1 and 2, and converted these values to ‘measured albedos’ in three MODIS bands. Relative to other validation methods, our helicopter measurements have the advantages of larger spatial coverage and of (almost) direct measurements in satellite-sensor spectral bands. We found the smallest differences between the satellite-derived and helicopter albedos for the Kangerlussuaq transect: for AVHRR data a mean difference of 0.01 in both bands (with the satellite in near-nadir position) and for two MODIS images a mean difference of 0. 00-0.02 for bands 2 and 4, and 0.03 for band 1. For two AVHRR images of Vatnajokull, we found mean differences of up to 0.06. Differences are primarily due to errors in the satellite-derived albedos, which, in turn, are mainly caused by errors in the calibration coefficients of the satellite sensors and insufficient knowledge of the angular distribution of the radiation reflected by snow and ice. Satellite data obtained from view zenith angles larger than ~50-55° appeared to be unsuitable.

Climatic mass balance of the ice cap Vestfonna, Svalbard: A spatially distributed assessment using ERA-Interim and MODIS data

“…Comparison of remotely sensed surface albedo with the measurement record of AWS2 yield a mean difference of +0.02 over the period May 2008 to July 2009 and a root mean square error (RMSE) of 0.12. This result is in agreement with uncertainties obtained by Greuell and Oerlemans [2005] on the Greenland ice sheet and is regarded as an adequate reproduction of in situ conditions, given that the AWS only samples a small point within the much larger MODIS footprint [ Stroeve et al , 2006a]. A characteristic annual evolution of the albedo bias depending on solar zenith angle as described by Wang and Zender [2010] is not clearly identifiable due to the limited length of the study period.…”

Changes in surface albedo after fire in boreal forest ecosystems of interior Alaska assessed using MODIS satellite observations

“…No single satellite‐derived estimate of surface albedo functions perfectly under all conditions and it is likely that high accuracy can only be obtained at a regional scale by combining multiple satellite products with high quality surface and aircraft observations. There is some empirical evidence, for example, that MODIS estimates of pure snow albedo are low by ∼2–7% [ Greuell and Oerlemans , 2005; Stroeve et al , 2005; Salomon et al , 2006]. This type of bias would reduce the magnitude of the fire‐induced changes in surface albedo during winter and spring (and consequently our estimates of shortwave surface forcing).…”

“…Summer albedo also increases above prefire levels because early successional plant functional types, including grasses and deciduous trees and shrubs, have leaves and branches with higher albedo than those of evergreen needleleaf trees [ Betts and Ball , 1997; Roberts et al , 2004; Amiro et al , 2006; McMillan and Goulden , 2008]. Over an annual cycle, the dynamic range of albedo within boreal forest ecosystems is bounded by black carbon coatings (char) on vegetation and soil surfaces that have an albedo of 0.06 [ Chambers and Chapin , 2002; Chambers et al 2005] and snow that has an albedo between approximately 0.7 and 0.9 [ Greuell and Oerlemans , 2005; Stroeve et al , 2005].…”

Impact of spatial, spectral, and radiometric properties of multispectral imagers on glacier surface classification

“…highly reflective in much of the visible and near-infrared) and ice is darker, therefore as the melt season progresses the glacier as a whole gets darker overall -specifically in proportion to the relative contributions of different glacier facies. In this way, it is possible to monitor glacier albedo as a tool for monitoring glacier mass balance (Dumont et al, 2012;Greuell & Oerlemans, 2005;Greuell et al, 2007).…”