Satellite-retrieval and modeling of glacier mass balance

Abstract

In this research project we use satellite measurements to infer the mean specific mass balance (Bm) of glaciers. Vatnajökull, the largest ice cap in Europe, is being used as a test-case because this ice cap has often been studied. Only one aspect of Vatnajökull has not been investigated so far, and that is the relation between its mass balance and climatological conditions. We therefore also construct a mass balance model, the results of which can be compared to the satellite images. On Vatnajökull an extensive meteorological experiment took place in the summer of 1996, and the dataset collected during this experiment can be used for validation of the model. The mass balance model is calibrated with in situ measurements. We find that the incoming longwave radiation is best modeled as a function of meteorological variables in the free atmosphere just above the relatively thin katabatic layer. Also, the ratio of changes in the 2 m temperature to changes in the free atmospheric temperature (the climate sensitivity) is smaller than 1. Horizontal precipitation gradients over Vatnajökull are large, which results in a strongly varying sensitivity to external temperature changes over the ice cap. Local climatic conditions thus highly determine the mass balance and its sensitivity. From the mass balance model we construct a Seasonal Sensitivity Characteristic (SSC) of Vatnajökull, which consists of the sensitivity of Bm to monthly perturbations in temperature and precipitation. Temperature sensitivities are high in summer and nearly zero in winter, while precipitation sensitivities are high in winter and low in summer. With the SSC we reconstruct the mass balance of Vatnajökull since 1825. The results for two Icelandic glaciers correlate very well with mass balance records that are extracted from length records with a linear inverse model. For the south of Vatnajökull we find that after 1900, the length record is well explained by temperature variations alone, while another Icelandic glacier (S¢lheimajökull) was also influenced by precipitation variations. Equilibrium Line Elevation at the end of the melting season (ELAe) is a proxy for Bm, but the Equilibrium Line (EL) is mostly not visible on AVHRR images when it is located above its position of the previous year(s). EL detection is further hindered by clouds and a gradual transition between ice and firn or snow. Consequently, detection of the ELAe on albedo images is not particularly useful for estimating Bm. Instead, we propose to study the mean albedo of the entire ice cap throughout the melting season, so that all available information about the surface albedo is taken into account. The mean net potential global radiation, which can be estimated from the mean surface albedo alone, is found to relate linearly to Bm. Like AVHRR images, Synthetic Aperture Radar (SAR) images only display the surface firn line as a distinct boundary and not the snow line (i.e. melting snow and firn have the same backscatter and albedo signatures). On the other hand, winter SAR images of Vatnajökull display sub-surface firn-ice transitions, but these do not correspond to the late summer surface firn line. Unlike albedo images, SAR images do not display inter-annual variations of the signal within the accumulation area that are clearly related to Bm and are therefore less usefull for mass balance retrieval. For the north-western part of Vatnajökull the indirect observations (satellite and modelled) correspond with the direct (in situ) observations. Only for one year the different estimates strongly diverge, probably due to a lack of satellite images and an erroneously modeled distribution of precipitation. We compute the best estimate of the mass balance of north-western Vatnajökull as the weighted mean of the individual estimates. For the part of Vatnajökull where the mass balance has not been measured, we can use the indirect methods to estimate relative variations in Bm.