Reliability analysis for design of stake networks to measure glacier surface velocity

Abstract

AbstractMeasurement of glacier surface velocity provides some constraint on glacier flow models used to date ice cores recovered near the flow divide of remote high-altitude ice caps. The surface velocity is inferred from the change in position of a network of stakes estimated from the least-squares adjustment of geodetic observations – terrestrial and/or spaced-based – collected approximately 1 year apart. The lack of outliers in and the random distribution of the post-fit observation residuals are regarded as evidence that the observations contain no blunders. However, if the network lacks sufficient geometric redundancy, the estimated stake positions can shift to fit erroneous observations. To determine the maximum size of these potential undetected shifts, given the covariance of the observations and the approximate network geometry, expressions are developed to analyze a network for redundancy number and marginally detectable blunders (internal reliability), and the position shifts from marginally detectable blunders (external reliability). Two stake networks, one on the col of Huascar–n (9°–07'S, 77°–37'W; 6050 m a.s.l.) in the north-central Andes of Peru and one on the Guliya ice cap (35°–17'N, 81°–29'E; 6200 m a.s.l.) on the Qinghai–Tibetan Plateau in China, are examined for precision and internal and external reliability.