Calving is a complex process subject to several cooperating atmospheric, oceanographic and glaciological forcings that vary in space and time, and whose relative effects are challenging to separate. Statistical ‘Systems Analysis’ is commonly used in engineering and economics to extricate complex ‘force–response’ relationships. Here we apply Systems Analysis to the Amery rift system, East Antarctica. We develop a scalable ‘System Model’ driven by a coarsely-sampled dataset characteristic of glaciological observations in remote locations, and validate it using rift lengths observed in 2000–06 and 2012. In this initial demonstration, we forecast a detachment date of ~2019 ± 5 years for the large tabular iceberg colloquially known as the ‘Loose Tooth’, for which relative humidity surprisingly emerges as the best statistical predictor. RACMO2 climate modelling reveals that relative humidity correlates best with surface albedo and snowmelt, both of which are intimately linked to firn compaction and ice shelf temperature and flow. We postulate that relative humidity can therefore serve as a proxy for internal stress, a known key control of ‘Loose Tooth’ calving. Although no physical causality is implied in Systems Analysis, postulates such as this can aid in setting priorities in studies of complex glaciological processes.