The roles of statistical inference and historical sources in understanding landscape change: the case of feral buffalo in the freshwater floodplains of Kakadu National Park

The statistical analysis of Bowman et al. (Journal of Biogeography, 2008, 35, 1976– 1988) revealed the weak relationship between the rate of woody cover encroachment onto the freshwater floodplains in the central section of Kakadu National Park (KNP) over a 40-year period and estimates of proximate water buffalo (Bubalus bubalis) density. The analysis relied on detailed mapping of buffalo tracks, the best historical record of spatial variation of buffalo density in KNP. In their reply, Petty & Werner (Journal of Biogeography, 2009, doi: 10.1111/j.1365- 2699.2009.02185.x) prefer to privilege an amalgam of historical sources to claim that buffalo removal is the primary driver of the woody expansion on floodplains. The contrasting weight placed on data analysis and differences of interpretation underscore a tension between statistically based historical ecology approaches and environmental history narratives, a tension that forms part of the broader cultural clash between the Sciences and Humanities.