CGRG Bibliography of Canadian Geomorphology
Search Results
Author : Cogley, J.G.
Date : 2003.
Title : Influence and redundancy in the global network of mass-balance measurements.
Publication : Technical Note. Department of Geography, Trent University, Peterborough.
Issue : 2003-2.
Page(s) : 14 p.
Abstract
Analysis of an updated global mass-balance dataset shows that we can be moderately confident that the average mass balance of small glaciers (those other than the ice sheets) has been growing more negative since the late 1960s, when it was near to zero. However the phrase “moderately confident” is not adequately quantified. I report here on progress in the use of cross-validation to address this problem. In the present context, cross-validation may be understood as 1) using the ensemble of N measured glaciers to model global average mass balance, with spatial biases corrected by a spatial interpolation algorithm; 2) repeating the calculation N times, leaving out one glacier each time; and 3) treating the distribution of the difference (step 2 minus step 1) as a guide to model performance. I interpret results in terms of the complementary concepts of “influence” and “redundancy”. Influential glaciers are those whose omission alters the global balance estimate substantially. Redundant glaciers are those whose omission makes little difference; they may therefore be seen (but only now that we have measured them) as “representative” glaciers. The great majority of measured glaciers are redundant. When any one of them is omitted, the global balance estimate (for the reference period 1961-1990) changes by less than 2 kg m
Bibliography of Canadian Geomorphology