The accelerated burning of fossil fuels and the large-scale transformation of forests into agricultural land since the industrial revolution have led to rapid climate warming which, if not mitigated, is threatening human life on Earth. The field of climate science strives to better understand the complex climate system and its internal feedbacks and with that be able to provide reliable projections of future changes. These depend most and foremost on societal and political dynamics. In the Paris Agreement, the global community has committed to limit warming below 2â—¦ C, which requires global net-zero greenhouse gas emissions until the year 2050. However, global efforts to reduce emissions are still by far insufficient to reach the agreed target.
Although climate science long has identified immediate emission reduction as the most important mitigation strategy, a multitude of internal feedbacks and processes are still poorly understood. This creates the need for further investigations into potential impacts of past and future emissions, not only on the climate but also on the global biosphere, including potential irreversible tipping points or mitigation opportunities. One important sub-system of the global climate system and carbon cycle that is expected to see increased pressures from anthropogenic disturbance and climate warming are global wetlands. Wetlands are sources of methane, an important greenhouse gas, and thus hold the potential to contribute to future warming. A special type of wetlands, peatlands, also act as long-term carbon stores and thus could both help to remove anthropogenic carbon from the atmosphere on long timescales or lead to large additional release of carbon into the atmosphere.
This thesis presents model investigations of wetland dynamics, using the dynamic global vegetation model LPX-Bern, which is developed and maintained at the University of Bern. LPX-Bern is used to investigate changes in wetland area and greenhouse gas budgets from the past to the future, with a particular focus on peatlands.
Chapter 1 gives an introduction to the key concepts discussed in this thesis. First, the global carbon cycle and its different components are introduced. Then wetlands, the main subject of this thesis, are defined and discussed in detail. Wetlands are a key component of the global carbon cycle as they function both as large carbon stores and sources of methane. The history of wetlands and their reconstruction through proxy evidence is discussed in the context of past climate variability. The past and potential future effects of human activities on wetlands are examined. Finally, the world of wetland modeling is introduced, giving an overview of the historic model development and different model complexities.
Chapter 2 then introduces the LPX-Bern and presents model adjustments that were implemented during this thesis. The modules representing peatlands and wetlands are discussed in detail, including the formulation for the dynamic calculation of wetland and peatland area and the new treatment of dynamic peatland area in case of prescribed land-use change. The discussion of the methane module includes the emission calculation for different types of wetlands, the implementation of methane emissions from fires, and the description of a re-calibration of key emission factors in preparation for the different modeling studies presented in the following chapters.
In chapter 3, a modeling study published in Biogeosciences is presented which investigates the transient history of peatlands from the Last Glacial Maximum (LGM), about 21,000 years ago, to the present. Transient LPX-Bern simulations suggest that peatland area was highly dynamic in the past and changes in area were driven mostly by changes in precipitation and temperature. The study argues that to determine the net peat carbon balance, the full history of peatlands has to be considered, including peatlands that vanished over time. The simulated transient evolution of todayâ€™s northern peatlands is compared to data reconstructions and large model-data mismatches are found concerning the inception of peat in northern Asia. However, the simulated peatland distribution and carbon storage at present-day compare well to literature estimates. Additional time-slice simulations at the LGM show that uncertainties in the prescribed climate forcing propagate to large uncertainties in peatland variables.
Chapter 4 presents a follow-up study published in Biogeosciences which directly builds on results from the study presented in chapter 3. The transient simulation from the LGM to the present is taken as the basis for future projections of peatland dynamics. Different future climate and land-use scenarios are used to investigate potential future short-term and long-term changes in peatland area and carbon storage. The results suggest likely future losses of global peatland area and carbon, even under present-day climate, with large parts of todayâ€™s northern peatlands at risk. Losses in response to future climate and land-use change are expected to increase with increasing future emissions. Uncertainties connected to uncertain climate anomalies are quantified by using output from a climate model ensemble as forcing.
In chapter 5, model investigations into past wetland methane emissions are presented. Results from transient LPX-Bern simulations from the LGM to the present are compared to the methane ice-core record. Large model-data mismatches are found, most notably the absence of a simulated increase in emissions from the LGM to the pre-industrial period (PI). Driver attribution reveals a small temperature sensitivity and large sea-level driven tropical wetland loss as potential sources of the small LGM-PI methane emission increase. Preliminary investigations into model adjustments, addressing the temperature dependence of methane production and the dynamic wetland model, show potential to increase the LGM-PI methane emission rise, but alone are not sufficient to close the model-date gap. Furthermore, the discussed model changes could worsen model performance in other respects which would need to be addressed.
Chapter 6 presents a selection of two collaborative studies for which LPX-Bern model output was provided. First, simulations that contributed to the Global Methane Budget, a community publication that is part of the Global Carbon Project, are discussed in detail, with a focus on comparing wetland methane emissions between LPX-Bern simulations with prescribed and dynamically calculated wetland area. Emissions are found to be globally comparable, but with regional biases in wetland prediction translating into large regional differences in simulated emissions. In the second part, simulations are presented that contributed to a model-intercomparison project investigating projected future changes in peatland net carbon balance and methane emissions under different scenarios. The LPX-Bern simulations in this study, where the peatland area is prescribed, are compared to similar LPX-Bern simulations from chapter 4, where the peatland area is calculated dynamically. Future peatland area loss is found to mostly lead to larger predicted carbon loss but also smaller peatland methane emissions than if peatland area is held constant.
Finally, Chapter 7 gives an outlook over potential future model investigations and model development, progressing and building on the work presented in this thesis. An additional appendix describes the implementation of a new transient land-sea-ice mask for paleo simulations into the LPX-Bern, which however was not used in any simulations presented in the main text. The new implementation increases the update time-step and allows for variable grid cell land fractions.