The study of the surface mass balance of valley glaciers has a long history but one that is dogged by uncertainty and errors, and uncertainty about those errors. These problems are acknowledged by the glaciological community and have been examined and formalised in several publications. The latest of these stems from a workshop organised by the World Glacier Monitoring Service, the results of which are presented in the first paper of this thesis. The paper examines two common methods and some of their associated errors, with the aim of unifying them and providing more robust data sets. New methodologies, such as ground penetrating radar (GPR), are used in the second paper herein to provide richer and alternative data sources for approaching measurement problems related to snow depth and, to some degree, snow density. The third paper is concerned with both technical survey issues and glaciological definitions when surveys of glacier surfaces are performed for use in mass balance calculation. Many of these issues are common to remote sensing methods and ground based surveys but the paper attempts to make this commonality more explicit. Whilst the first three papers concern themselves with the act of measurement and calculation the fourth paper considers errors brought about by logistical constraints restricting the timing of surveys. Such errors are technically avoidable to a large degree but inevitable in practice. In the case presented here the error is one of unobserved accumulation, falling late in the season, after the last survey. By modelling expected ablation from minimal data, such as temperature, and comparing this with measured mass balance an estimate of unobserved accumulation is made.

Also contained in this work is an assessment of glacier front surveys, specifically those performed by the Tarfala Research Station. Such surveys are assumed to act as a monitor of glacier response to climate change and are assumed to be a large scale proxy of mass balance but the resolution of the response as well as the resolution of survey methods indicates that the frequency of such surveys should be reduced and that remote sensing methods may be more effective.

Common to all glaciological field surveys is the relative sparsity of data rendering error analysis and many statistical methods ineffective but new technologies such as Lidar, Global Navigation Satellite System, GPR and remote sensing indicate a way forward and the potential for future work to deliver detailed and reliable data.