Lowry et al.
(2000) data/output files:
This directory contains some (but not all) of the data files and estimated
topography fields described in Lowry, Ribe and Smith (2000), Dynamic
elevation of the Cordillera, western United States, J. Geophys. Res.,
All files are ascii in a three-column (longitude, latitude, field) format.
If you have questions or would like to see something else included here,
please write and let me know.
WUStopo.llz contains the raw topography (derived from an old USGS data set
described in Simpson et al. (1986), J. Geophys. Res., 91, 8348-8372).
These data were used to estimate effective elastic thickness Te
and as raw elevation for the estimate of "dynamic topography".
WUSgrav.llz contains Bouguer gravity data (derived from an old USGS
data set described in Simpson et al. (1986), J. Geophys. Res., 91,
8348-8372). These data were used to estimate effective elastic thickness
WUSq_s.llz contains surface heat flow data (optimally interpolated from
Blackwell's compilation of measurements, circa 1998). These data were used
to estimate the conductive thermal contribution to elevation and the variable
thickness of mechanical lithosphere.
WUSTe.llz contains estimates of effective elastic thickness Te,
using the methodology described in Lowry and Smith, J. Geophys. Res.,
100, 17,947-17,963, 1995. These Te estimates were used to calculate
surface load contributions to elevation, for isostatic filtering of estimated
internal contributions to topography from crustal mass and conductive thermal
variations, and in the estimation of variable thickness of mechanical lithosphere.
(1) The estimation windows used here are small (200 to 400
km). Subsequent research on synthetic data has shown that such small estimation
windows can yield large variance in Te estimates.
estimates used an old approach to load deconvolution that neglected the higher
(1000 kg/m^3) density of surficial fluid in ocean regions, which can bias
Te toward lower values in oceanic lithosphere.
(3) The old style of
load deconvolution also used an a priori fixed surface-to-internal load ratio of
-1 at long spatial wavelengths prone to singularity. This approach can bias
toward higher or lower values of Te depending on the relationship of the
assumed load ratio to the true average.
(4) Elevation and Bouguer gravity
data grids were extrapolated up to 50 km past the raw data constraints in some
parts of the Pacific ocean, which can also introduce errors near the
surf_h.llz contains estimates of surface load contributions to total
elevation, derived from isostatic deconvolution of the surface and internal
loading. This was subtracted from the raw elevation as the first step toward
estimating "dynamic topography".
(1) Load estimates are
very sensitive to assumed Te (see caveats above).
(2) At long
wavelengths prone to singularity (for which load ratios were assumed fixed in
original Te estimation), all elevation was assumed to result from
internal loading. Subsequent studies have shown that surface loading can have
substantial amplitude at long wavelengths (e.g. Lowry and
Zhong (2003) J. Geophys. Res., 108(E9), 10.1029/2003JE002111,
crust_h.llz contains estimates of crustal mass contributions to total
elevation (incorporating variations in both thickness and density estimated from
seismic refraction data). This was also subtracted from raw elevation as a step
toward estimating "dynamic topography".
of seismic velocity to density has very large uncertainty, resulting in elevation
uncertainties of order 600 to 1000 m (one-sigma!).
(2) The seismic
refraction data used to constrain crustal mass are very sparsely sample, have
highly variable quality of original data and used various different
modeling/inversion approaches to arrive at velocity structure. (This portion of
the analysis really should be updated with more recent data). The crustal mass
estimate is consequently the largest source of error in the estimate of "dynamic
therm_h.llz contains estimates of (mantle lithospheric) conductive
thermal contributions to total elevation (estimated from a locally
one-dimensional approximation of conductive geothermal variation given surface
heat flow and crustal thickness). This was also subtracted from raw elevation
as a step toward estimating "dynamic topography".
Mantle temperatures depend on poorly-known crustal thermal conductivity and
(2) Modeling assumes the geotherm is in steady-state.
Consequently large spurious negative buoyancy is modeled in Cascadia and
California (where mining of heat by ancient and/or ongoing subduction yields
artificially low estimates of deep temperature). Negative buoyancy may also be
overestimated in areas heavily influenced by Pleistocene glaciation, as surface
temperatures may not have had time to fully re-equilibrate.
was referenced to a potential temperature but never grafted to an adiabat, which
may introduce errors due to variable thickness of the thermal boundary layer.
(This was corrected in later versions of the codes but the calculation was not
dynamic.llz contains the estimate of dynamic topography after
subtracting surface load, crustal mass and mantle conductive thermal mass
contributions to raw elevation.
(1) Uncertainties in
the fields subtracted from raw elevation are large (800+ m at one-sigma, see
above) and high-frequency (less than 1000 km wavelength) variations are highly
suspect owing to modeling assumptions and data sampling problems.
is not true "dynamic elevation" but rather a deep mantle contribution to
elevation incorporating BOTH sublithospheric thermal contributions AND mantle
contributions relating to variable mineralogy (particularly variable
garnet/pyroxene concentrations resulting from melt removal) and variable water
content; the latter contribution may be as large or larger than the former.
| || This material is based upon work supported by NASA under SENH grant number
NAG5-7619. Any opinions, findings, and conclusions or recommendations expressed
in this material are those of the author and do not necessarily reflect the
views of NASA. |