Spatial distribution of debris thickness and melting from remote-sensing and meteorological data, at debris-covered Baltoro glacier, Karakoram, Pakistan C. MIHALCEA, 1 C. MAYER, 2 G. DIOLAIUTI, 1 C. D’AGATA, 1 C. SMIRAGLIA, 1 A. LAMBRECHT, 3 E. VUILLERMOZ, 4 G. TARTARI 4 1 Department of Earth Sciences ‘Ardito Desio’, University of Milan, Via Mangiagalli 34, I-20133 Milan, Italy E-mail: claudia.mihalcea@unimi.it 2 Commission for Glaciology, Bavarian Academy of Sciences, Alfons-Goppel-Strasse 11, D-80539 Munich, Germany 3 Institute for Meteorology and Geophysics, University of Innsbruck, Innrain 52, A-6020 Innsbruck, Austria 4 CNR–IRSA, Water Research Institute/National Research Council, Localita ` Occhiate, I-20047 Brugherio, Milan, Italy ABSTRACT. A distributed surface energy-balance study was performed to determine sub-debris ablation across a large part of Baltoro glacier, a wide debris-covered glacier in the Karakoram range, Pakistan. The study area is  124 km 2 . The study aimed primarily at analyzing the influence of debris thickness on the melt distribution. The spatial distribution of the physical and thermal characteristics of the debris was calculated from remote-sensing (ASTER image) and field data. Meteorological data from an automatic weather station at Urdukas (4022 m a.s.l.), located adjacent to Baltoro glacier on a lateral moraine, were used to calculate the spatial distribution of energy available for melting during the period 1–15 July 2004. The model performance was evaluated by comparisons with field measurements for the same period. The model is reliable in predicting ablation over wide debris-covered areas. It underestimates melt rates over highly crevassed areas and water ponds with a high variability of the debris thickness distribution in the vicinity, and over areas with very low debris thickness (<0.03m). We also examined the spatial distribution of the energy-balance components (global radiation and surface temperature) over the study area. The results allow us to quantify, for the study period, a meltwater production of 0.058 km 3 . 1. INTRODUCTION Calculating the distribution of ice melt over large areas of debris-covered glaciers represents a challenge in assessing water resources from glacier melt in remote areas such as Karakoram, Pakistan, and the Himalaya. Remote-sensing data are now available and can be used to analyze the spatial distribution of surface energy fluxes, along with traditional field studies that can be used to calibrate the new techniques. While quite a few energy- and mass-balance studies have been performed on debris-free glaciers, studies on debris- covered ice are not numerous, especially for distributed energy- and mass-balance studies. Recently, Nicholson and Benn (2006) presented a modified surface energy-balance model to calculate melt beneath a debris layer from daily mean meteorological data on two European debris-covered glaciers (Ghiacciaio del Belvedere, Italy, and Larsbreen, Norway). Han and others (2006) proposed a simple model to estimate ice ablation under a thick debris layer by using surface temperature and debris thermal properties on Koxkar glacier, Tien Shan, China. During the last decade, a few papers have focused on debris-covered glaciers in the Himalaya and Karakoram (e.g. Hewitt and others, 1989; Mattson and others, 1989, 1993; Young and Hewitt, 1993; Nakawo and Rana, 1999; Kayastha and others, 2000; Nakawo and others, 2000; Takeuchi and others, 2000). Some studies have utilized remote-sensing data to analyze the spatial distribution of surface temperature to calculate the energy available for melting (Nakawo and others, 1993; Rana and others, 1997; Nakawo and Rana, 1999). Un- fortunately these studies only calculate melt over small areas and short time-spans. One of the crucial input parameters for a successful application of melt models over large areas is a reliable debris-cover distribution. This can be obtained from remote-sensing data but still represents a challenge (Mihalcea and others, in press). Several studies have dealt with the debris-covered area of Baltoro glacier and its tributaries in the past (Desio, 1954; Desio and others, 1961; Pecci and Smiraglia, 2000; Diolaiuti and others, 2003). More recently Mayer and others (2006) and Mihalcea and others (2006) analyzed the glaciological and meteorological characteristics of Baltoro glacier (glacier velocities, ablation rates and meteorological data from field expeditions), providing the observational and experimental basis for the present study. Here we present the results of the application of a distributed energy-balance model calculation of sub-debris melt on a large Karakoram debris-covered glacier. The model predicts the magnitude of buried-ice melt using a debris thickness distribution derived from remote-sensing data and field meteorological data from a local automatic weather station (AWS), Urdukas (4022 m a.s.l.), at Baltoro glacier. The study time frame and area were determined by the 2004 Italian scientific–alpine expedition on Baltoro glacier ‘K2 2004 – 50 Years Later’. The study period is quite short (1–15 July 2004), but different data sources were available during this time span, permitting intercomparisons (i.e. satellite data and field investigations, including an AWS on the glacier surface) and model validation. In addition, our approach for calculating debris thickness and melt distribution had not been applied before on Karakoram glaciers, avoiding duplication of existing literature. Annals of Glaciology 48 2008 49