Stable carbon isotopic (δ¹³C) analyses of soil organic matter (SOM) have been used in the past to characterise C3–C4 vegetation changes. However, the temporal and spatial resolution of these isotopic data are not well established. Here, we present data from δ¹³C analyses of whole and size-separated SOM, which are discussed in conjunction with organic (total organic carbon (TOC) content) and inorganic (%clay) soil data. These data are put into context with the current vegetation state (assessed from tree size-class distribution) and the 50-year vegetation history (assessed from aerial photographs). By linking below- and above-ground datasets, we show that δ¹³C analyses of SOM can accurately record vegetation-change histories over short- (10 and 50 years) and longer-term (hundreds of years) time scales. Our data also show that spatial variability was relatively small for the clay TOC content but was much larger for δ¹³C data, indicating that the number of soil cores required for statistical significance is highly dependent on the kind of measurements intended. Finally, interpretation of δ¹³C data from SOM to assess the history of C3–C4 vegetation change is complicated by the inherent ¹³C-enrichment of SOM, owing to decomposition processes, which occurs regardless of vegetation change. We suggest a method for distinguishing ¹³C-enrichment of SOM that is due to soil-inherent (decomposition-related) processes from ¹³C-enrichment that is due to increased inputs of C4 organic matter.