Variation is a key concept in the study of statistics and its understanding is a crucial aspect of most statistically related tasks. This study aimed to extend and apply a hierarchy for describing students' understanding of variation that was developed in a sampling context to the context of a natural event in which variation occurs. Students aged 13 to 17 engaged in an inference task that necessitated the description of both rainfall and temperature data. The SOLO Taxonomy was used as a framework for analyzing student responses. Two cycles of Unistructural-Multistructural-Relational levels, one for qualitative descriptions and the other for quantitative descriptions, were identified in responses. Implications of the extended hierarchy for describing understanding of variation for research, teaching and assessment are outlined. (Contains 10 tables and 1 figure.)