Modeling forcings and responses in the global carbon cycle-climate system: Past, present and future

Roth, Raphael (2013). Modeling forcings and responses in the global carbon cycle-climate system: Past, present and future (Unpublished). (Dissertation, Universität Bern, Philosophisch–naturwissenschaftliche Fakultät, Physikalisches Institut, Abteilung für Klima– und Umweltphysik)

Text roth13phd.pdf - Other Restricted to registered users only Available under License BORIS Standard License. Download (32MB) | Request a copy

This thesis investigates the interactions within the coupled carbon cycle-climate system. It addresses questions regarding both its past and future evolution using a computationally efficient carbon cycle-climate model. Therefore, no specific single topic or mechanism is studied, but a suite of applications is presented in the following.
The cycling of carbon in the Earth system is driven by a complex interplay between physical, biological and chemical processes in the major carbon reservoirs: atmosphere, land, ocean, sediments and lithosphere. Of primary interest is the net exchange of carbon dioxide ( CO2) from/to the atmosphere, as this is the process that ultimately influences Earth’s climate, as CO2 absorbs electromagnetic radiation. Carbon fluxes on decadal timescales are controlled by the interactions between atmosphere, surface ocean and living biomass on land. On longer timescales, atmospheric CO2 concentrations are driven by the oceanic carbon cycle, which is influenced both by changes in ocean circulation, marine biology, seawater chemistry and sediment interactions. On geological timescales, atmospheric CO2 stabilization is primarily a result of balancing igneous rock weathering, volcanism and seafloor burial rates. The tool invoked in this theses to study these interactions on different timescales is a cost-efficient Earth system model of intermediate complexity: the Bern3D-LPJ model. This model does not only simulate the cycling of carbon, but also tracks isotopes 13C and 14C (radiocarbon), thus helping to understand the complex nature of the global carbon cycle.
The introductory chapter of this thesis gives an overview on the global carbon cycle, with a focus on carbon-climate interactions and isotopes. It emphasizes paleoclimatic investigations and benchmarking to strengthen our understanding of current and future human-driven climate change. Moreover, the most important concepts and methods are introduced as a basis for the subsequent chapters.
In Chapter 2, the Bern3D-LPJ carbon cycle-climate model is shortly described. In addition, a summary of changes incorporated in the recently updated ocean component is given. A special focus is set on the adjustment of biogeochemical parameters in order to improve the modeled distribution of dissolved carbon, alkalinity and nutrients. In a final step, the model’s preindustrial solution, i.e. ocean circulation, tracer distributions and global fluxes is presented and discussed in the light of the latest datasets available.
Chapter 3 consists of a study published in Earth and Planetary Science Letters in which a recent hypothesis is tested, namely whether the deglacial retreat of Northern Hemisphere ice sheets triggered enhanced deglacial volcanic activity in these regions. As volcanoes release carbon stored in Earth’s mantle as they erupt, this process may considerably contributed to the observed glacial-interglacial CO2 variation of ∼100 parts per million (ppm).
In order to test the implications of the proposed mechanism, we apply a range of 40 kyr volcanic emission scenarios of carbon, 13C and 14C to the Bern3D model. We find a CO2 increase of 46 ppm for the central scenario, peaking at around 6 kyr before present (BP). By comparing the modeled response in the atmosphere-ocean-sediment system to a range of ice-core and marine proxy records, we show that the proposed mechanism cannot be rejected: the uncertainties in the emission history as well as in other processes involved in deglacial and early Holocene CO2 dynamics prevent a firm conclusion. One of the drawbacks is that carbon isotopes serve as a weak constraint here, because i) magma δ13C does not differ significantly from the signature of the contemporary atmosphere and ii) the ∆14C decrease induced by a weakening of the geomagnetic shield and outgassing of old oceanic carbon may screen the ∆14C decline induced by old volcanic CO2. On the other hand, the pronounced CO2 increase in the early Holocene resulting from the central and high scenarios is in conflict with the observed CO2 decline during this period. We therefore conclude that the volcanic feedback to deglaciation is small.
Chapter 4 focuses on 14C and its application as a proxy for solar activity. By running simulations with the Bern3D-LPJ model from 20 kyr BP to 1950 AD with prescribed atmospheric ∆14C from the IntCal09 and SHCal04 records, the Holocene radiocarbon budget is solved for atmospheric production (Q). We do not only consider past changes in ∆14C , but also take into account changes in CO2 and its radiative forcing, continental ice-sheet extent, sea level, dust flux, shallow-water carbonate deposition and anthropogenic landuse change. The results of these experiments are summarized in a study published in Climate of the Past. With these experiments, we partially resolve the discrepancy between carbon-cycle and particle-simulation based estimates of the contemporary production rate; our estimate of Q is in line with the results of the most recent production-model estimates. We do not only reconstruct Q for the first time with a 3-D model over the past 10 kyr, but Q is also applied as a proxy for the solar modulation potential and total solar irradiance (TSI) using published methodologies. We show that linking past TSI to instrumental records is problematic, both because i) early instrumental records are prone to considerable uncertainty and ii) humans disturb the natural 14C cycle not only by atomic bomb testing, but also by the combustion of fossil fuels and landuse emissions, both introducing uncertainty in the radiocarbon budget and thus our Q reconstructions. Our best-guess normalization of the solar activity record suggests a high, but not exceptionally highcontemporary solar activity compared to the past 10 kyr. We also show that a bias is introducedby globally applying the Northern Hemisphere record (IntCal09) as done in past studies. Beside the solar activity reconstruction, Chapter 4 also includes results on pre-Holocene 14C dynamics. For the past 50 kyr, 10Be and paleointensity-based Q reconstructions are prescribed to the Bern3D model in order to simulate atmospheric ∆14C . We find that most of the reconstructions of Q are in conflict with high glacial ∆14C levels as preserved in natural archives; past carbon cycle changes are probably too small to explain the remaining offset.
In Chapter 5, we present the results of a multi-model effort to determine the response of the carbon cycle-climate system to a pulse-release of CO2. In this study, which we published in Atmospheric Chemistry and Physics, fifteen models ran the same CO2 pulse-emission experiments under contemporary CO2 conditions (389 ppm). On a time horizon of decades to centuries, atmospheric CO2 and therefore its radiative impact is controlled by how much of the (excess) CO2 is taken up the by ocean and land carbon reservoirs. The evolution of atmospheric CO2 can then be represented by impulse response functions ( IRFCO2 ). IRFCO2 serves as a basis for the computation of the global warming potential (GWP), a metric to compare the radiative effect of emissions of different greenhouse gases. We provide an updated estimate (and uncertainty range) of the absolute GWP of CO2 at e.g. year 100 of 92.5 (68–117) ×10−15 yr W m−2 per kg CO2. Further, we set up numerous sensitivity experiments with the Bern3D-LPJ model in order to quantitatively estimate the influence of the pulse size, the chosen background scenario and carbon-climate feedbacks. In addition to analyzing the response in atmospheric CO2 levels, the model responses in air-sea and air-biosphere fluxes, atmospheric temperature, sea level rise and ocean heat content are quantified and fitted by analytical functions. As an extension to the published study, a simple representation of the continental weathering feedback is implemented in the Bern3D and its influence on IRFCO2 quantified in terms of a pulse-release of CO2 under preindustrial conditions.
In Chapter 6, the feedback of the terrestrial biosphere under global warming scenarios is explored. In a study published in Nature Climate Change, we i) reconcile the observed 20th century increase in atmospheric N2O and CH4 levels and ii) quantify the terrestrial response to CO2 increase and climate change in terms of feedback parameters. The feedbacks of N2O and CH4 are not well understood so far, as most current Earth system models do not simulate the cycling of these compounds. By running numerous experiments with the Bern3D-LPJ model applying the extended Representative Concentration Pathway (RCP) emission scenarios, we separate feedbacks arising from snow-albedo changes as well as changes in CO2, N2O and CH4 fluxes. Our results show that under RCP8.5 (a business-as-usual scenario), the land biosphere likely turns into a net source of CO2 by 2100 AD. We find that by the year 2300 AD, N2O and CH4 feedbacks add another 0.4–0.5 ◦C on top of the 0.8–1.0 ◦C caused by the albedo and carbon cycle feedbacks.
Chapter 7 outlines how the Bern3D-LPJ model can be applied in the future, for example by modeling the response of multiple greenhouse gases (CO2, N2O, CH4) during abrupt climate events or by applying inverse modeling methods to paleoclimatic questions. In the appendices, supplementary material is given, mainly related to the Bern3D-LPJ model but also containing additional material to the multi-model IRFCO2 study.

Interest & Impact

Downloads

0 since deposited on 07 Mar 2024
0 in the past 12 months

Citations

5 Citations in Web of Science ®
6 Citations in Scopus

Search

in Google Scholar™

Services

Actions (login required)

Edit item

Actions (login required)

Edit item