From forest to farmland and meadow to metropolis: Integrated modeling of Holocene land cover change

Background/Question/Methods

Did humans affect global climate over the preindustrial Holocene? While this question is still debated, the co-evolution of human-environment interactions over the last 11,700 years had an undisputed role in influencing the development and present state of terrestrial ecosystems, many of which are highly valued today as ecological, cultural, and scenic resources. Yet we still have a very incomplete picture of human-environment interactions over the Holocene, both spatially and temporally. In order to address this problem, we combined a global dynamic vegetation model with a new model of preindustrial demographic, technological, and economic development. We drive this model with paleoclimate from GCM scenarios to simulate land cover and land use change, fire, soil erosion, and emissions of CO₂ and CH₄ from 11 ka BP to AD 1850. We evaluate our simulations in part with a new set of continental-scale reconstructions of land cover based on records from the Global Pollen Database.

Results/Conclusions

Model results show that climate and tectonic change controlled global land cover in the early Holocene, e.g., shifts in forest biomes in northern continents show an expansion of temperate tree types far to the north of their present day limits, but that by 3 ka, humans in Europe, east Asia, and Mesoamerica had a larger influence than natural processes on the landscape. At 3 ka, anthropogenic deforestation was widespread with most areas of temperate Europe and southwest Asia, east-central China, northern India, and Mesoamerica occupied by a matrix of natural vegetation, cropland and pastures. Burned area and emissions of CO₂, CH₄ from wildfires declined slowly over the entire Holocene, as landscape fragmentation and changing agricultural practices led to decreases in burned area. In contrast, soil erosion increased with increasing human pressure over the last 11 ka, except in areas where topsoils became exhausted, e.g., in the Andes and the eastern and southern Mediterranean. While we simulate fluctuations in human impact on the landscape, including periods of land abandonment, approaching the Industrial Revolution nearly all of Europe and east Asia is dominated by anthropogenic activities. In contrast, the collapse of the aboriginal populations of the Americas following 15^th century European contact leads to a period of ecosystem recovery. Our results highlight the importance of the long histories of both climate change and human land use on the development of continental-scale landscapes. We emphasize the utility of combining paleo-archives with remote sensing-based datasets for land cover reconstruction and model evaluation.