Transcript How accurately can we measure density within the Earth?

Density within the Earth?
Mike Warner
Earth Science & Engineering
Imperial College London
Outline
• Earth structure
• Methods
• Accuracy
• Recommendations
Earth Structure
• crust vs mantle
• lithosphere vs asthenosphere
• ocean vs continent
chemical composition
less accurate
~ 10 – 50 km
depth
more accurate
thermal
crustboundary
layer
convecting
adiabatic
mantle
interior
highly heterogeneous,
chemically differentiated
(rather) homogeneous
& chemically uniform
temperature
less accurate
thermal
boundary
layer
lithosphere
heterogeneous,
strong, complicated
depth
~100 km
more accurate
convecting
adiabatic
interior
“asthenosphere”
highly viscous,
convecting at high
Raleigh number
well-mixed,
homogeneous,
simple
Oceans vs Continents
Oceans – more accurate
•
•
•
•
young
simple
thin crust
mantle density related to cooling
Continents – less accurate
•
•
•
•
long history
complicated
thick crust
no simple density relationships
Methods
• gravity
• seismology
• others
Gravity
• low resolution
• non-uniqueness
• density heterogeneity
• much improved if boundaries are
known from another technique
Seismology
• high spatial resolution
• good absolute accuracy
• determines boundaries
• does not measure density directly
• elastic properties correlated with
density
• more money –> better results
Other methods
Density varies with
• porosity
• composition
• temperature
• pressure
Measure or estimate each of these
Accuracy
• absolute densities
– conversion from seismics
– determination of composition
– determination of temperature
• boundaries
– base of sediments
– base of crust
– base of lithosphere
Profile length
• European
~2500 km
• Trans-continental
~9000 km
European
lithosphere
2500 km
European - continental
crust
European - oceanic
crust
Errors in density
location
length
“a priori”
“best”
Continental
2500 km
4.7%
2.9%
Oceanic
2500 km
2.6%
1.7%
Continental
9000 km
2.0%
1.7%
Oceanic
9000 km
1.8%
1.5%
Errors are ~2 sigma
(errors not really Gaussian)
Recommendations
If at all possible:
• source and receiver on oceanic crust
• maximum profile length beneath ocean
• use maximum profile length
• custom seismic experiments
Recommendations
Avoid:
• Alps
• central Europe
• thick crust (e.g. Fenoscandia)
• Europe to Japan
Recommendations
Best profiles:
• Western Europe to Eastern US
• Atlantic Islands (Canaries, Maderia,
Azores) to Portugal, western Spain, NW
France, southern Ireland, western
England
Errors in density
location
length
“a priori”
“best”
Continental
2500 km
4.7%
2.9%
Oceanic
2500 km
2.6%
1.7%
Continental
9000 km
2.0%
1.7%
Oceanic
9000 km
1.8%
1.5%
Errors are ~2 sigma
(errors not really Gaussian)