Last week’s problems a) Mass excess = 1/2πG × Area under curve 1/2πG = × in kgs 2 m -3 Area under curve = -1.8 ×10-6 x 100 m 2 s -2 So Mass excess (actually deficit) = -4.3 ×10 5 kg per m b) The anomaly decreases to zero in m. This suggests that z is < 5m. c) The area of the car park = excess mass per unit length/density contrast, hence the area = 4.3 ×10 5 /1800 = 239 m 2 Assuming a rectangular shape, the height is 239/100 = 2.39m. Problem 4
Problem 5 x 1/2 = 2.3 cm = 575 m For a spherical body z max = 1.3 x 1/2 = 750 m
Today’s lecture Seismic reflection Overview Data collection - Sources and receivers Theory Seismic reflection processing Two new developments Problems 6 and 7
Marine streamer carries 100s of hydrophones. Detect changes in pressure Birds on streamer keep it level Land receivers detect movement Reflections occur at boundaries where velocity and/or density change
Horizontal distance ~10 km ~1500 traces TWTT (seconds) Earth’s surface ~6 km depth Oil exploration colour = reflection amplitude 0 km depth 1 trace
3D image subsurface Slice at t = 400 ms Can also track horizons and generate fly-bys
Seismic stratigraphy Unravel depositional changes with time
Normally only interested in p-wave (smallest amplitude/fastest) The higher the frequency the better – reflections appear sharper
Velocity versus density
Seismic boat Airgun array guns
Tows 6 streamers at once, each 6 km long Each streamer ~480 channels (group of hydrophones)
5 am in Derbyshire Dynamite buried ~ 5 m below surface Packed with filler Encourage energy to travel down – not up Ideal source – high frequency
Truck vibrates up and down to produce a continuous seismic signal
Horizontal (S-wave) source
A hammer source
Geophones
Ocean-bottom seismometersBroad band seismometers
For small angle of incidence Gas beds produce high-amplitude (bright) negative reflections
Rays represent the direction of travel of the seismic wave, and are perpendicular to the 3D wavefront The energy in the wavefront decreases in with increasing distance from the source (geometrical spreading) and is inversely proportional to 4πr 2. The amplitude of a seismic wave is proportional to the square root of energy, hence amplitude is proportional to r -1 Absorption and scattering also decrease the energy in the seismic wave as it passes through the Earth This decrease in energy with increasing travel time (so- called attenuation) can be written as: Three effects combined
White = primary Red = source ghost Black = multiple
Attenuation - decrease in amplitude and frequency content Multiples and ghosts - wavelet extended
Seismic reflection processing Raw data contains noise, refractions, S-wave reflections, surface waves, multiples The aim of processing is to try to remove all of these and leave in only the primary reflections
Filtered trace Raw data from a broad-band seismometer
Statics If sources and receivers are not horizontal then reflections from interfaces will not line up correctly in the time section. After applying statics, horizontal reflectors will appear horizontal
Reflections are hyperbolic Stacked trace gets plotted at x position that lies at midpoint of traces being stacked
Test velocity (2-5km/s) Power of stacked trace at each velocity (Amplitude 2 ) Pick velocity that gives highest amplitude (white dots) Multiples and S-wave reflections have lower velocities than primaries Refractions are not hyperbolic Semblance plot
Deconvolution Run a autocorrelation on the seismic traces to determine the effective wavelet (The wavelet at t = large) Design an inverse filter (the deconvolution) to convert the wavelet back
As the distance between two reflectors decreases, their reflections start to overlap, and eventually do not appear as two separate reflections. The minimum vertical resolution (the closest two reflectors can be and still be identified as two separate reflectors) is assumed to be ~ ¼ seismic wavelength A reflection does not arrive from one single reflection point – but from circular area. Thus small gaps in a reflector will not be obvious in the seismic data The minimum horizontal resolution (the smallest observable gap) is assumed to be ~ the width of a Fresnel zone
Relatively new developments 4D time lapse Survey 1985 and 1999 Track changes in reservoir
State-of-the-art Permanent arrays on sea bed Real-time monitoring of reservoir e.g. Hydro-fracturing