Land use mapping error introduces strongly-localised, scale-dependent uncertainty into land use and ecosystem services modelling

We analysed the impact of land use mapping error on uncertainty in projections of land use and ecosystem services in Australia's agricultural land from 2013 to 2050. We simulated land use mapping error at four levels then modelled potential land use transitions and indicators of ecosystem services supply using the Land Use Trade-Offs (LUTO) model. Increasing error resulted in larger areas of uncertainty in land use transition. This influence was highly localised, with concentrations in mixed cropping areas, rather than homogeneous extensive grazing areas. Error effects also varied with spatial scale. While barely discernible when aggregated at the national level, error effects increased markedly with granularity of assessment. Sensitivity to land use mapping error also differed between ecosystem services. The results were driven largely by geographic heterogeneity and nuance in the environmental and economic drivers of land use peculiar to the study area. The finding of strongly-localised, scale-dependent impacts that differed between ecosystem services is likely to hold for other areas. Understanding the effects of land use mapping error propagation is necessary both for guiding efforts to reduce mapping error, and for supporting policy and governance at multiple scales.