Biophysical and Human Controls of Land Productivity under Global Change : Development and Demonstration of Parsimonious Modelling Techniques

Sammanfattning: Net primary production (NPP) serves as an indicator for plant-based resources such as food, timber and biofuel for human appropriation. It is defined by the annual production of plant matter and is mainly controlled by climate and human activities. Climate change in combination with human activities is altering NPP. As the controls of NPP are expected to further change in the future, it is vital to investigate alterations in NPP and their magnitudes. The impacts of climate change and human activities on NPP can be explored in integrated assessment (IA) frameworks, where sectoral models are coupled and interact rapidly. For such frameworks, parsimonious models are desired because they enable rapid estimates and facilitate easy model coupling for explorations of multiple global change scenarios (i.e. large volumes of data). This thesis aims to advance parsimonious modelling techniques for quantifying current and future NPP on land. This is accomplished by developing and testing rapid models that facilitate easy model coupling to explore the impacts of multiple global change scenarios on NPP. The model development is based on the meta-modelling concept, which can be applied to simplify the dynamic vegetation model LPJ-GUESS in a parsimonious model. For this, multiple climate change and [CO2] perturbations are applied to LPJ-GUESS to simulate NPP. The NPP data are then used to define biophysically motivated relationships between NPP and the driving climate variables along with [CO2]. The relationships are then combined in a synergistic function – the meta-model. Thereafter, the meta-models are assessed for their performance in estimating NPP by comparing them to LPJ-GUESS NPP simulations, to independent field observations and to NPP experiments under enriched [CO2] on biome level. The results provide confidence in the modelled NPP estimates for the most productive biomes, which are important for global quantifications of NPP. The meta-models capture NPP enhancement under enhanced [CO2] adequately in the majority of the studied biomes. Finally, the NPP meta-models are coupled with other sectoral models in two IA modelling-frameworks in order to explore the impacts of global change on ecosystem indicators. The first framework enables an IA of climate change impacts and vulnerabilities for a range of sectors on the European level. This thesis conducts a sensitivity analysis on the effects of climatic and socio-economic change drivers on model outputs related to key sectors. This provides better quantification and increased understanding of the complex relationships between input and output variables in IA modelling-frameworks. The second framework addresses the NPP supply-demand balance in the Sahel region by coupling two sectoral models in order to analyze the timings and geographies of NPP shortfalls in the 21st century Sahel under global change. The results show consistent regional NPP shortfalls in the Sahel for the majority of global change scenarios.Overall, the parsimonious modelling techniques developed in this thesis contribute with rapid NPP estimates on the biome and global scale. BME NPP estimates agree reasonably well with NPP observations in the majority of biomes (especially in the most productive biomes). This thesis demonstrates that NPP meta-models facilitate easy model coupling for exploring the impacts of global change on human-environmental systems in IA modelling-frameworks.