Pathways to future cropland : Assessing uncertainties in socio-economic processes by applying a global land-use model

Global agricultural production almost tripled within the last five decades. The production increase wasbased on expanding cropland and pastures, as well as the intensification of agriculture, including increased use ofhigh yielding crop varieties, machinery, irrigation, artificial fertilisers, and pesticides. Both, agriculturalintensification and the expansion of agricultural land-use lead to environmental degradation, pose threats tohuman health, and contribute to climate change. Transitioning towards sustainable agricultural land use, therefore,is one of the major challenges facing humanity in the 21st century. This challenge is aggravated by the need tofeed the growing and increasingly affluent population, the effects of climate change on agriculture and theincreasing demand for land to mitigate climate change, through for example bioenergy production. This thesisassesses how uncertainties in the development of socio-economic drivers and processes, such as populationgrowth, dietary shifts, technological change, and bioenergy production, affect the outcome of future land-use andland cover change (LULCC). Future development of socio-economic drivers and climate forcing are described bythe latest scenarios developed for environmental and climate-change research, i.e. the Shared Socio-economicPathways (SSPs) and the Representative Concentration Pathways (RCPs). The impacts of the changing driverson the land system are assessed with the global Parsimonious Land Use Model (PLUM). PLUM was shown toreproduce observed global agricultural land use change at the global to country scale for 1991-2010. Future globalcropland changes were found to be very sensitive to the assumed yield growth rate. In a subsequent study,estimates of future yield were therefore derived with a global dynamic vegetation model, and included impacts ofclimate change. Without assumed land-based mitigation strategies, simulated future cropland ranged from 970 to2280 Mha by 2100, compared to current cropland area of 1500 Mha. This range is consistent with those found inthe recently published literature. Accounting for the uncertainties related to the interpretation of socio-economicprocesses and drivers described in the scenarios expanded the simulated range for global cropland to 890-2380Mha (± one standard deviation) by 2100 and led to strongly overlapping cropland ranges for three out of fivescenarios. Uncertainties related to scenario interpretation are thus of similar importance as uncertainties acrossdifferent models for estimating the possible outcome of future LULCC. When land-based climate change mitigationstrategies are considered, additional cropland requirements of 603-1115 Mha by 2100 were simulated for theproduction of bioenergy. However, considerable uncertainties related to the strength of mitigation efforts and cropyields accompany this estimate. Continuous expansion of cropland into grasslands and forest, as in scenarios withstrong population growth and low technological change or scenarios with large bioenergy production, wassimulated to transform the terrestrial biosphere from a carbon sink into a carbon source. Moreover, remainingwithin the estimated planetary boundary for global cropland (15% of ice-free land) is not possible when aiming toensure food security while simultaneously producing bioenergy that significantly contributes to strong climatechangemitigation efforts by 2050. In a local to regional case study future food security was shown to be at riskunder the assumed future socio-economic developments, demonstrated here for countries in the Sahel region ofAfrica. Implementing sustainable agricultural management practices as well as global trade will be important toensure food security in the future. Overall, uncertainties in population development, technological change,resource intensity and land degradation were shown to contribute to a wide range of future agricultural LULCC