Application of novel fabric and texture analysis techniques to GISP2.

Recent development of a fully automated digital instrument for fabric analysis has led to a dramatic increase in the quality and quantity of c-axis data from ice cores (Hansen and Wilen, 2002 J. Glaciol.). The new data allow straightforward application of statistical techniques to evaluate nearest-neighbor correlations between grains for probing active grain formation processes such as polygonization and/or recrystallization (Alley and others, 1995 J. Glaciol.). After demonstrating some of the software tools we have developed to facilitate this type of analysis, we will present and discuss results for horizontal and vertical sections from a range of depths at GISP2. These results show unambiguous correlations in c-axis direction for nearest-neighbor grains, but there are fairly large variations in the strength of the correlation even among multiple sections from a single depth. We will also present experimental results of correlations between grain shape and c-axis orientation. The results will be discussed in relation to expected correlations from simple models of fabric evolution. There are surprising discrepancies that might be attributed to polygonization. If this conjecture proves correct, this type of analysis may provide an alternative technique (to that described above) to probe active grain formation.