14 | Halliburton Landmark
New Frontiers in Plate Boundaries
by : Jean-Christophe Wrobel-Daveau , Bruce Eglington and Graeme Nicoll
Crustal dynamics showing earthquakes , plate boundaries , and volcanoes . Image from https :// svs . gsfc . nasa . gov / 155
with credit to NASA / Goddard Space Flight Ce Research Project ( GCRP ), National Oceanic and Atmosphere Administration ( NOAA ), United States Geological Survey , National Science Foundation ( NSF ), Defen ( DMA ), New York Film and Animation Company , Silicon Graphics , Inc . ( SGI ), Hughes STX Corporation .
Given a plate tectonic model is meant to be a singular representation of what the Earth looked like back through geological time , one might ask , why are so many different models available , and which one should I use ? In order to answer these questions , it is useful to understand the genealogy of the different plate tectonic models , what data underpin them , what methods and techniques are used to construct and update them , and what uncertainties are involved .
We will go on to consider how the scope and complexity of these models went hand-in-hand with growing computational power over the last 60 years . In order to compare and contrast the variety of plate tectonic models that exist today , we need to set out the main difference between plate tectonics and continental drift models , and consider the importance of plate tectonic boundaries .
PLATE TECTONIC MODEL GENEALOGY
Exploration Handbook | 27
circulation models consist of a three-dimensional representation of the atmosphere coupled to the land
surface. As atmospheric models are unable to model ocean processes they need to be provided with
summer and cold winter. This also applies to wind and ocean
currents, which tend to follow a distinct seasonal pattern. Of
course, in some years, summers may be cooler than normal
and winters warmer, but climate is the 30-year average of
weather and, as such, is more predictable.
There are also ocean model counterparts to atmospheric models. The most complex models couple
both atmospheric and ocean models, so that the whole system is considered. These represent the
“flagship” models used in paleoclimate studies, and many are also able to simulate additional features,
such as vegetation coverage and type. Even the most sophisticated models do not incorporate
all aspects of the Earth’s systems. These other features, such as tides, need to be modeled
Like all models, paleoclimate models are a simplification of
reality. Model resolution may result in an over-simplification of
certain aspects of the Earth system. For instance, the width of
ocean upwelling zones, important for source rock formation,
may be broader than they were in reality, or rain shadows
may be missed due to an averaging of the topography. Some
processes may not be able to be resolved by a model, such
as the rate of precipitation from clouds. In these instances,
a value is assigned within the model, which may be an over-
simplification as in reality this value may change spatially.
There are also uncertainties in the boundary conditions, such
as the paleogeography, topography, and bathymetry, and
the concentration of CO 2 in the atmosphere. This is why it is
important to run several simulations to explore the range of
possible scenarios, rather than relying on a single simulation.
The choice of model depends on how the simulations will be used and is normally a pragmatic decision
that reflects the best trade-off between resolution and sophistication, and computational cost and the
time taken to run the simulation. The most sophisticated climate models have geographic cells sizes
of several kilometers and contain over a hundred vertical levels. Some models have a cell size of many
hundreds of kilometers and contain a small number of vertical layers. The most sophisticated models
may only be able to simulate 0.5 years in 24 hours, even using supercomputing facilities, while other
models can simulate several thousand years in the same time.
Why is this important? Climate is defined as the 30-year average of the weather. Therefore, at least
30 years of simulation are needed to model climate. However, it is not as simple as this. For a climate
simulation to be reliable, it is imperative that the climate has reached equilibrium. For instance, if CO 2
levels are changed in a model, the climate is perturbed and becomes unbalanced. The climate must
warm or cool to reach a new equilibrium consistent with the new CO 2 level.
The true test of the ability of a climate model to predict past
conditions is to compare simulations with data. Unfortunately,
it is impossible to measure many of the simulated parameters,
such as sea temperature, directly in the geological past
and we must, therefore, rely on a variety of paleoclimate
proxies. These proxies range from sophisticated geochemical
paleotemperature estimates to the simple occurrence of
climatically controlled facies, such as coral reefs. In general,
there is very good correspondence between climate
simulations and climatically controlled facies (Figure 3), but
this is not always the case for more complex geochemical
proxies. These proxies may possess uncertainties from
a range of sources, including diagenesis, calibration, and
seasonal bias (e.g. Davies et al., 2019), so they must be used
When paleoclimate simulations are generated, it is not only atmospheric CO 2 levels that are changed,
but also many other boundary conditions, such as the land elevation and the bathymetry of the ocean.
It is, therefore, necessary to simulate thousands of years for robust results. To ensure the most
sophisticated models are providing robust simulations, they would need to be left running for several
years! This is not practical for most applications, least of all hydrocarbon exploration. This is further
compounded, as it is good practice to run several simulations with different boundary conditions, due to
uncertainties in the boundary conditions (such CO 2 levels).
CAN WE REALLY MODEL CLIMATE?
Is it possible to model past climates accurately? After all, meteorologists are often unable to predict
the weather a few days in advance! It is important not to confuse weather with climate. Weather is the
phenomenon that happens every day. It is extremely dynamic (often chaotic) and is, therefore, often
hard to predict. However, it is possible to predict, with a high degree of certainty, that it will be warm in
Large benthic foraminifera
Mixed photozoan and
18 ºC Cold month mean
14 ºC Cold month mean
l l i b
2 0 2
2 0 2
l l i
r t o
Figure 3> A) Distribution of modern-day tropical carbonates versus cold month mean sea surface temperature. Most corals are bound by the 18°C cold month temperatures; whereas, larger benthic foraminifera can tolerate conditions as low as 14°C. B) Comparison
of an early Eocene simulation against carbonate distribution. Notice the high level of correspondence.
“Unfortunately, it is
impossible to measure
many of the simulated
parameters, such as sea
in the geological past
and we must, therefore,
rely on a variety of
These proxies range from
to the simple occurrence of
climatically controlled facies,
such as coral reefs.”
26 | Halliburton Landmark
The first attempts at reconstructing the paleoposition of continental land masses were hand-drawn as far back as the 17 th
century by Dutch map makers , and later in the 20 th
century by Alfred Wegner
and Boris Choubert ( Kornprobst et al ., 2018 ). However , these were really just singular snapshots of often poorly constrained geological times , and are more akin to paleogeographic maps ( Figure 1 ). Indeed , they lacked the understanding of geodynamic processes , such as the absolute and relative motion of plates on a sphere , the driving mechanisms behind plate motion , and the existence of plate limits .