Ice-sheet mass balance can be inferred from comparison between total net accumulation and total ice discharge, and this data set contains results from such a comparison for higher-elevation parts of the Greenland ice sheet. During 1993-1997, velocities were measured at a line of stakes about 30 km apart completely around the ice sheet approximately along the 2000 m elevation contour except in the south west, where it was higher because of high mountains, nunataks, and crevasses [Thomas et al., 2000]. More information on these velocities can be found on our web site. Ice thickness was also measured along the traverse by a low-frequency ice-sounding radar [Gogineni et al., 1998]. Snow accumulation was from existing estimates [Ohmura and Reeh, 1991], supplemented by recent German and PARCA measurements and with estimates of total precipitation at weather stations for coastal accumulation rates. Each estimate refers to the time period over which measurements were made, ranging from years to centuries, but the overall set of results is heavily influenced by results from recent PARCA cores, most of which refer to the past two or three decades. Total snow accumulation within catchment areas corresponding to selected groups of stakes, expressed as a volume flux of ice, was compared to ice discharge between the stake locations. This was calculated as the product of ice thickness and velocity integrated between the stakes, after correcting for the ratio (R) between surface and column-averaged velocity, based on model simulations of the velocity/depth profile [Huybrechts, 1996]. Errors for small catchment areas are quite large because of uncertainties in local accumulation rates, the size of the catchment area, and R [Thomas et al., 1998]. For areas larger than about 100,000 sq km, the random nature of these errors reduces their overall effect to less than seven percent of the snow-accumulation rate, and considerably less for the entire region under study [Thomas et al., 2000]. Snow-accumulation rates range from less than 0.1 m of water equivalent per year in the northeast to more than one meter per year in parts of the southeast, with an average value for the entire ice sheet of about 0.3 m/yr [Bales et al., this issue]. The data set presented here is in tabular form, with a listing for individual gates of: station locations forming the gates; velocities, ice thicknesses and R estimates at the gates; total upstream accumulation; and resulting estimates of average ice-thickening rates (dH/dt) for the catchment areas appropriate to the gates. Because errors in dH/dt are large for single gates, additional Tables are provided giving:

1. Values for larger ice-sheet catchment areas (approximately 30,000 sq km) formed by including several adjacent gates.
2. Values for 12 larger gates (up to >100,000 sq km) that represent regions of the ice sheet with distinctive patterns of dH/dt. These estimates are compared to independent estimates from altimetry measurements in Thomas et al.

Errors in accumulation rates are the largest source of uncertainty in these mass-balance estimates, but they are progressively reduced as more information is acquired. The data presented here were compiled before all PARCA accumulation measurements had been analyzed, so they include different estimates of dH/dt based on accumulation estimates ranging from those in Ohmura and Reeh [1991] to our own attempts to improve these using then available PARCA data, and details of these can be found on our web site. Consequently, the dH/dt estimates provided here will be progressively improved as our estimates of snow accumulation also improve. Ongoing work with these data include:

• Investigation of the regions that are significantly out of balance, in an attempt to identify causes for the imbalance.
• Progressive updating of the dH/dt estimates using improved accumulation estimates, flow lines based on better maps of ice-sheet topography, and better estimates of R. This will be continued into the GLAS mission, using accumulation estimates from atmospheric analyses, for comparison with GLAS measurements of surface-elevation change.