V23C-2146
Sources of Non-Monotonicity for the Earth's Thermal State and Heat Transfer
Monotonic models of Earth's thermal evolution do not account for temporal variations of heat loss out of the
mantle due to changes in the geometrical pattern of convection, for instance in relation to the cycle of
aggregation and dispersal of continents. These variations are significant, of the order of 10~TW on a
timescale of a few hundreds million years, and not negligible by comparison to the expected monotonic
variation of Earth's heat loss due to secular cooling (around 5~TW on a timescale of one billion years).
An event that may influence the heat loss of a terrestrial planet in a more dramatic way is the possible
cessation of plate tectonics. This possibility is studied here using numerical models of thermal convection
including a temperature-dependent viscosity and plastic yielding, which leads to either a stagnant lid regime,
an episodiplate-like regime or a stable one, depending on the chosen parameters
for plastic yielding, on the vigour of convection, on the thermal state of the mantle and on the geometry of the
flow. The model also includes the possibility of partial melting within the mantle, which was widespread when
the Earth was hotter.
We discuss the consequences of time variations of the convection pattern and of the possible cessation of
plate tectoncis on the thermal state of the mantle, and study the temporal variations in melting events and
their consequences on mantle heat loss.
V23C-2147
Dynamic Mechanisms Behind Periodic and Episodic Phenomena in the Earth's Interior
The evolution of many planets and the Earth in particular is clearly influenced by convective processes in the
interior. Convection in the strongly nonlinear regime, as to be expected in planetary interiors, typically
exhibits time dependent behavior, ranging from simple periodic to chaotic signatures. What are relevant
mechanisms producing periodic and/or episodic events and what are typical scales. We employ numerical
models allowing for realistic rheologies, i.e. temperature, pressure and stress-depenent visocities in order to
identify key-mechanisms leading to time varying behavior. Clearly plumes, developing from the 670 km/ and
or the Core mantle boundary develop periodically in a realisttic strong temperature dependence of the
viscosity is assumed. With an appropriate rheology surface plates form naturally in our model, being an
integral part of the convective system. We observe changes in the style of plate tectonics, showing
transitions from periods of plate tectonics to stagnant lid convection and vice versa. This can potentially
explain global changes in the surface dynamics of planets. We further demonstrate that separately
convecting layers can originate from the interplay of compositional and thermal contributions. The creation of
layers and their final destruction also induces periodic and episodic fluctuation in the planetary evolution
http://earth.uni-muenster.de
V23C-2148
Temperature Beneath Continents as a Function of Continent Size and Convective Wavelength
Geodynamic modeling studies have demonstrated that mantle global warming can occur in response to the aggregation of supercontinents, possibly leading to large scale melting and associated continental breakup. Such feedback calls for a recipe describing how continents help to regulate the thermal evolution of the mantle. Here we use 3D spherical mantle convection models with continents to quantify variations in subcontinental temperature as a function of both continent size and convective wavelength. Through comparison to a simple analytical boundary layer model, we show that larger continents beget warming of the underlying mantle, with heating compounded by the formation of broader convection cells associated with the biggest continents. Our results hold well for purely internally heated and partially core heated models with Rayleigh numbers of 105 to 107 containing continents with sizes ranging from that of Antarctica to Pangea. Results from a time dependent model with multiple, mobile continents of various sizes suggests that the tendency for temperatures to rise with continent size persists on average over time scales of billions of years. This model could provide a standard against which to measure theories of continental driven thermal events, such as proposed flood basalts of the Archean.
V23C-2149
Searching for catastrophic mantle-melting events in an imperfect geologic record
Peaks in the age distribution of the continental crust (CC) and in the distribution of Os and He isotopic values in mantle-derived samples suggest that the CC grew in pulses during large mantle-melting events. However, a significant question remains. How can one confidently identify peaks in such time series data when the geologic record is incomplete and sampling is highly biased? Histograms of age versus frequency and other representations of time series data such as probability density functions (PDFs) are highly susceptible to systematic biasing by both non-uniform geologic preservation and by non-random sampling. While sophisticated methods (feature analysis) exist to identify statistically significant peaks in such data, these methods require that sampling of the data is random, which the geologic record is not. Though non-uniform preservation generally cannot be corrected for, over-sampling can be mitigated using geologic processes to sample an area, such as zircon studies that use river drainage basins to collect zircons from continents. Plots of age versus value (such as an isotopic ratio) are fairly robust to over-sampling. Though if an area contains samples with a range of values, more sampling will reveal the more extreme values, which can create artificial peaks. Given the inherent systematic biases, it is difficult to assign statistical significance to peaks in any single geologic time series. A more robust approach is to use the correlation of peaks between independent data sets. The only assumption that is required is that the systematic biases in each data set are independent of each other. In this case, the likelihood of peaks in different data sets corresponding in age can be quantitatively evaluated. Applying this approach to the CC, Os and He record for the solid Earth and the S- MIF and organic C isotope record for the atmosphere, our results indicate significant global events around 2.7, 1.9 and 1.2 Ga (in descending order of probability). An event at 3.3-3.5 Ga is weakly suggested.