Optimization approach to Q-factor estimation from VSP data

Abstract

A zero-offset VSP survey is carried out with spaced apart receivers located in a vertical wellbore. Spectra of the signals at the receivers following wavefield separation are estimated. An absorption coefficient is estimated using differences in spectra between all pairs of receivers.

Description

BACKGROUND OF THE DISCLOSURE

1. Field of the Disclosure

This disclosure relates to a method of measuring the attenuation of seismic waves in earth formations. An attenuation coefficient may provide information about seismic lithology and fluids in earth formations. In addition, the attenuation coefficient is used in deconvolution of seismic data, thereby providing improved imaging of the subsurface. Quantitative analysis of amplitudes is complicated by Q during amplitude variation with offset (AVO) analysis of seismic data where attenuation effects are superimposed on AVO signatures.

2. Description of the Related Art

In surface seismic exploration, energy imparted into the earth by a seismic source reflects from subsurface geophysical features and is recorded by a multiplicity of receivers. This process is repeated numerous times, using source and receiver configurations which may either form a line (2-D acquisition) or cover an area (3-D acquisition). The data which results is processed to produce an image of the reflector using a procedure known as migration.

Sediments in the earth are attenuative, i.e., they absorb seismic energy. One result of the attenuation is that the bandwidth of a propagating seismic signal deceases as the wave propagates into the subsurface. As a result of this reduced bandwidth, the resolution of seismic reflectors decreases with depth. Knowledge of the attenuation coefficient (typically expressed by a constant α in nepers/wavelength of the seismic wave) can be used to deconvolve the seismic data and improve the resolution. In addition, Q is correlated with lithology and is highly dependent on the gas saturation of sediments. Knowledge of Q may thus provide a useful indication of lithology and fluid content of earth formations.

Knowledge of Q is very desirable, yet it is rarely measured. If a well has been drilled, core/laboratory and vertical seismic profiling (VSP) methods can be used. Each method has advantages as well as limitations. Many laboratory-based and field measurements of Q and its dependence on parameters such as lithology and gas saturation have been made on core samples.

The attenuation coefficient α is conventionally estimated using measurements from a Vertical Seismic Profile (VSP), though it may also be estimated from surface seismic data. Vertical (zero-offset) VSPs or check-shot surveys are nearly ideal configurations for estimation of Q. However, even in VSP data, a conventional approach normally provides low vertical resolution and quite often low accuracy. The reason is that in the conventional approach, only a small portion of input data is used to estimate Q. That is why the question of reliable Q-estimates remains. In theory, interval Q may be estimated for all two consecutive receiver depths, but in practice this is impossible. Two consecutive depth spectra may be too similar, and the difference is often so small that the Q estimates have significant errors.

The present disclosure is directed to an improved method of estimating attenuation from VSP data.

SUMMARY OF THE DISCLOSURE

One embodiment of the disclosure includes a method of evaluating an earth formation. The method comprises: acquiring seismic data using a plurality of seismic detectors, each at one of a plurality of spaced apart locations in a wellbore, responsive to activation of a seismic source near a top of the wellbore; estimating a spectrum at each of the plurality of seismic detectors; and estimating an absorption coefficient of the earth formation for at least one pair of the plurality of seismic detectors minimizing an objective function based on an exponential relation between the spectra of the at least one pair of seismic detectors.

Another embodiment of the disclosure includes a system configured to evaluate an earth formation. The system comprises: a plurality of seismic detectors, each positioned at a plurality of spaced apart locations in a wellbore, configured to provide a signal responsive to activation of a seismic source near a top of the wellbore; and at least one processor configured to: estimate a spectrum of the acquired signal at each of the plurality of seismic detectors; and estimate an absorption coefficient of the earth formation for at least one pair of the plurality of seismic detectors minimizing an objective function based on an exponential relation between the spectra of the at least one pair of seismic detectors.

Another embodiment of the disclosure includes a non-transitory computer-readable medium product having stored thereon instructions that when read by at least one processor cause the at least one processor to execute a method. The method comprises: using seismic data acquired by a plurality of seismic detectors, each at one of a plurality of spaced apart locations in a wellbore responsive to activation of a seismic source near a top of the wellbore, for estimating a spectrum of a downgoing wavefield of the acquired seismic data at each of the plurality of seismic detectors; and estimating an absorption coefficient of the earth formation for at least one pair of the plurality of seismic detectors minimizing an objective function based on an exponential relation between the spectra of the at least one pair of seismic detectors.

BRIEF DESCRIPTION OF THE DRAWINGS

The present disclosure is best understood by reference to the attached figures in which like numerals refer to like elements, and in which:

FIG. 1 illustrates the geometry of data acquisition of a VSP according to the present disclosure;

FIG. 2 shows spectral ratio curves for 40 ms of propagation time for different methods of Q estimation in the presence of noise;

FIG. 3 shows a layered model illustrating the method of the present disclosure;

FIG. 4 shows estimation through two layers (three receivers); and

FIG. 5 shows a flow chart of some of the steps in the method of the present disclosure.

DETAILED DESCRIPTION OF THE DISCLOSURE

For the present disclosure, the acquisition geometry of a VSP is illustrated in FIG. 1. Shown therein is the surface of the earth 123 with a rig 121 thereon. This may be a drilling rig or it may be a mast rig which conveys a wireline into a borehole 101. The borehole 101 penetrates the subsurface 103. Positioned in the borehole 101 are seismic sensors (detectors) denoted by 111a, 111b, 111c, 111d. Each of the sensors may include one or more of (i) a hydrophone, (ii) a single-component geophone and (iii) a multi-component geophone. Data for a single offset VSP is typically acquired using a single seismic source such as 125 at the surface (or within a body of water at the surface). Exemplary raypaths (112a, 112b, 112c, 112d) from the source location 123 to the plurality of receiver locations 111a, 111b, 111c, 111d are shown. In addition to receiving the direct arrivals from the source, each receiver would also receive reflected signals from boundaries below the reflector. To simplify the illustration, these boundaries and reflected raypaths are not shown. However, as discussed later, these reflected signals may be removed using prior art methods for wavefield separation.

In a typical VSP, data resulting from operation of a source at a single position such as 125 are recorded in each of the receivers 111a, 111b, 111c, 111d in the borehole. Analysis of the downgoing signals can provide information about the seismic velocities in the subsurface and the absorption in the subsurface.

Q-estimation from zero-offset VSP data is based on a linear absorption dependence on frequency (constant Q). Because of this linear dependence with frequency, the attenuation process may be expressed as:

$\begin{array}{cc}{S}_2\left(f\right)=\mathrm{Const}\times S_1\left(f\right)e^{-\pi f\frac{\Delta t_{1,2}}{Q_{1,2}}},& \left(1\right)\end{array}$ where S₁(f) and S₂(f) are amplitude spectra of a downgoing wave at the levels z₁ and z₂; Δt_1,2 is one-way time between levels z₁ and z₂, f is frequency in Hertz. Eqn. (1) is used in the spectral ratio method to estimate Q from zero-offset VSP data. Normally Q-factor is estimated for the well interval from shallow to deep receivers. The downgoing wavefield produced by a source at the surface 123 may be used to calculate amplitude spectra and used for estimating Q. To make the estimation more stable, several shallow receiver amplitude spectra may be averaged, as well as several deep receiver amplitude spectra.

Eqn. (1) gives:

$\begin{array}{cc}\log \frac{S_N\left(f\right)}{S_1\left(f\right)}=C-\mathrm{\alpha \Delta }{t}_{1,N}f=C-\mathrm{bf},& \left(2\right)\end{array}$ where C is a constant (attenuation factor that does not depend on frequency), and α is the effective absorption coefficient:

$\begin{array}{cc}\alpha =\frac{\pi }{Q_{1,N}}& \left(3\right)\end{array}$ S₁(f) and S_N(f) are averaged amplitude spectra for the shallow and deep receivers. Coefficient α is the slope of the line fitted to function P_1,N(f):

$\begin{array}{cc}{P}_{1,N}\left(f\right)=\log \frac{S_N\left(f\right)}{S_1\left(f\right)}& \left(4\right)\end{array}$ In other words, a straight line may be estimated using the least-squared method by minimizing quadratic function F(a,α):

$\begin{array}{cc}F\left(b,\alpha \right)=\sum _{f=F_1}^{F_2}{\left[P_{1,N}\left(f\right)-b-\mathrm{\alpha \Delta }{t}_{1,N}f\right]}^2& \left(5\right)\end{array}$ where F₁ and F₂ are minimum and maximum frequencies used for the Q estimate.

This spectral-ratio approach has two major drawbacks:

• 1. Function log(x) is strongly non-linear, and for small argument (small values of x) changes dramatically even for small changes in x.
• 2. The method uses only few amplitude spectra, and can lead to a poor vertical resolution and non-realistic Q values, including negative values.Eqn. (2) may be rewritten as:

$\begin{array}{cc}\alpha =\frac{C-P_k}{f_k\Delta t}\mathrm{where}{P}_k=\log \frac{S_2\left(f_k\right)}{S_1\left(f_k\right)}& \left(6\right)\end{array}$

The present disclosure uses a new method for Q determination: an optimization approach based on exponent (not ratio) estimate of Q-factor for all receiver pairs. First, the calculation approach to estimate Q for two receivers with amplitude spectra S₁(f) and S₂(f) may be modified. Instead of taking a logarithm of the ratio of the two receivers, an objective function may be used that calculates the average squared difference between these spectra. Eqn. (1) may be rewritten as:S₂(f)=C×S₁(f)e^−αf,  (7)where C is a constant and α is defined by formula (3). Then, to find C and α, minimization of an objective function G:

$\begin{array}{cc}G\left(\alpha ,C\right)=\sum _{f=F_1}^{F_2}{\left[S_2\left(f\right)-C\times S_1\left(f\right)e^{-\alpha f}\right]}^2& \left(8\right)\end{array}$ may be performed.

Determination of the constant C may be performed by solving the linear equation

$\begin{array}{cc}\frac{\partial G}{\partial C}=0,& \left(9\right)\end{array}$ which leads to

$\begin{array}{cc}C=\frac{\sum _{f=F_1}^{F_2}\left[S_2\left(f\right)S_1\left(f\right)e^{-\alpha f}\right]}{\sum _{f=F_1}^{F_2}\left[S_1^2\left(f\right)e^{-2\alpha f}\right]}& \left(10\right)\end{array}$ Substitution of eqn. (10) into eqn. (9), may yield:

$\begin{array}{cc}G\left(\alpha ,C\right)=\sum _{k=K_1}^{K_2}{\left[S_2\left(f_k\right)-S_1\left(f_k\right)e^{-\alpha f_k}\frac{\sum _{j=J_1}^{J_2}\left[S_2\left(f_j\right)S_1\left(f_j\right)e^{-\alpha f_j}\right]}{\sum _{f=J_1}^{J_2}\left[S_1^2\left(f_j\right)e^{-2\alpha f_j}\right]}\right]}^2& \left(11\right)\end{array}$ The absorption coefficient α may be estimated by scanning the objective function over a range of α and using the value that minimizes the eqn. (9).

The discussion above is illustrated by a model. For example, a homogeneous model with Q=100 may be used. The travel time between two receivers may be 30 ms. Amplitude spectra for the first receiver may be calculated for a real VSP downgoing trace. After using eqn. (1), random noise of 0.5% of average spectrum amplitude may be added to the spectrum at the second trace. This is shown by the curve 201 in FIG. 2. For high frequencies (between 80 and 100 Hz) the random noise may be about 30% of the amplitude spectrum, which distorts logarithm of the ratio for these frequencies. Curve 203 shows linear approximation of no-noise spectra ratio (corresponds to Q=100), curve 205 stands for linear approximation of noisy spectra ratio and leads to Q=48. Curve 205 corresponds to exponent estimate and gives Q=77, which is closer to model Q=100. Curve 207 corresponds to a spectral ratio estimate and gives Q=48.

To illustrate the influence of a propagation time on Q estimation in Table 1, the same added noise level (0.5% of average spectrum amplitude) was used, but with different times between points 1 and 2. Time intervals for Δt=10, 20, 40, 60, 80 and 100 ms were used. Table 1 shows the results on the modeling. It can be seen that for this noise/signal ratio (0.5% of average spectrum amplitude), for the time delay less than 40 ms, both methods lead to a large error in Q. For Δt=50 ms, exponent estimates provide accurate Q value (Q=92), while spectral ratio method leads to an essential error (Q=75).

TABLE 1
Propagation time	Spectral ratio	Exponential
10 ms	12	19
20 ms	22	34
30 ms	48	77
40 ms	63	88
50 ms	75	92
60 ms	88	96
70 ms	92	98
80 ms	97	99
90 ms	98	100
100 ms	100	100

Next, interval Q estimates based on optimization approach that uses all reasonable spectra pairs are described. Consider several receivers at the depths z₁, z₂, . . . , z_n. The notation A=1/Q is used for the layered Q (cumulative Q to this layer) and α=1/Q for the interval Q. q_k denotes Q between two consecutive levels z_k−1 and z_k, as shown in FIG. 3. Q_k,m, m>k+1, denotes Q for this layer, which includes more than two receivers. The same meaning is for α and A: α_k is an interval absorption for the interval “k” between two depths z_k−1 and z_k: α_k=1/q_k. A_m,n is a layered absorption coefficient, A_m,n=1/Q_m,n. A_m is a cumulative absorption coefficient with the first/shallowest reference trace.

For two receivers at the depths z₁ and z₂, assuming homogeneous interval and vertical propagation;S₂(f)=Const×S₁(f)e^−fα2Δt2  (12)Here S₁(f) and S₂(f) are amplitude spectra of downgoing wave at the levels z₁ and z₂; Δt_1,2 is one-way time, f is the frequency in Hertz.

If A_k,m is the effective absorption coefficient between depths z₁ and z₂, then:S_m(f)=Const×S_k(f)e^{−Ak,m(Tm−Tk)}  (13)where t_m is the time from the surface to the depth z_m. On the other hand, applying eqn. (12) to the levels z_k, z_k+1, . . . , z_k+m:

$\begin{array}{c}{S}_m\left(f\right)=\mathrm{Const}\times S_{m-1}\left(f\right)e^{-f{\alpha }_m{t}_m}\\ =\mathrm{Const}\times S_{m-2}\left(f\right)e^{-f{\alpha }_m{t}_m}{e}^{-f{\alpha }_{m-1}{t}_{m-1}}\\ =\mathrm{Const}\times S_{m-2}\left(f\right)e^{-f\left({\alpha }_m{t}_m+{\alpha }_{m-1}{t}_{m-1}\right)}\\ =\dots \mathrm{Const}\times S_k\left(f\right)\exp \left(-f\sum _{j=k+1}^m{\alpha }_j{t}_j\right)\end{array}$ From this and eqn. (2) it follows that:

$\begin{array}{cc}\begin{array}{c}{A}_{k,m}\left(T_m-T_k\right)=\sum _{j=k+1}^m{\alpha }_j{t}_j\\ =\sum _{j=1}^m{\alpha }_j{t}_j-\sum _{j=1}^k{\alpha }_j{t}_j\\ =A_m{T}_m-A_k{T}_k\end{array}& \left(14\right)\end{array}$ where Q_k is effective quality factor (absorption coefficient).Eqn. (14), gives three main formulas that are useful:

$\begin{array}{cc}{A}_{k,m}=\frac{1}{\Delta T_{k,m}}\sum _{j=k+1}^m{\alpha }_j{t}_j& \left(15\right)\\ {\alpha }_k=\frac{A_k{T}_k-A_{k-1}{T}_{k-1}}{t_{k-1}}=\frac{A_k{T}_k-A_{k-1}{T}_{k-1}}{T_k-T_{k-1}}& \left(16\right)\\ A_{k,m}=\frac{A_k{T}_k-A_m{T}_m}{T_m-T_k}\Delta T_{k,m}=\sum _{j=k+1}^m{t}_j=T_m-T_k& \left(17\right)\end{array}$

Turning now to the situation with receivers and three depths: z_r<z_m<z_n, applying eqn. (17) for pairs (k,m) and (k,n) gives:

$\begin{array}{cc}\Delta T_{r,m}{A}_{r,m}\sum _{j=r+1}^m{\alpha }_j{t}_j\Delta T_{r,n}{A}_{r,n}\sum _{j=r+1}^n{\alpha }_j{t}_j& \left(18\right)\end{array}$ In eqn. (18), z_r may be considered as a reference depth, which may be used to calculate Q for two intervals [z_r,z_m] and [z_r,z_n].Then

$\begin{array}{c}\Delta T_{r,n}{A}_{r,n}-\Delta T_{r,m}{A}_{r,m}=\sum _{j=r+1}^n{\alpha }_j{t}_j-\sum _{j=r+1}^m{\alpha }_j{t}_j\\ =\sum _{j=m+1}^n{\alpha }_j{t}_j\\ =\Delta T_{m,n}{A}_{m,n}\end{array}$ This gives the results:

$\begin{array}{cc}{A}_{m,n}=\frac{\Delta T_{r,n}{A}_{r,n}-\Delta T_{r,m}{A}_{r,m}}{T_n-T_m}& \left(19\right)\end{array}$ where z_r is any reference level above the segment [z_m,z_n]. It should be noted that in eqn. (19), reference trace k may be not only above the segment [z_m,z_n] (z_r<z_m<z_n), but also below: z_m<z_n<z_r.

Eqn (19) implies that, in addition to eqn (6), any level r (r<m<n) may be used to estimate Q_m,n between levels m and n, that is, for the segment [z_m,z_n] shown in FIG. 4. For each segment [z_m,z_n], not only may the amplitude spectra Q_m and Q_n eqn (17) be used, but also different reference traces above and below this segment that we use to calculate Q_m,n through eqn. (19). Using many different reference traces to estimate segment Q makes it more stable.

To find an initial approximation, eqns. (6) and (8) may be used. To determine this approximation, all reasonable spectra pairs (eqn. (6)), and also all possible triples spectra at the level z_r, z_m and z_n where z_r<z_k<z_m may be used. Here z_r is a reference depth, and Q(m,n) may be calculated using two Q values: Q_r,m and Q_r,n. If the trace window is defined as the receivers located between two receivers at the depths z_m and z_n and the amplitude spectra at these two receivers at the depths z_m and z_n are used, then it is possible to determine average Q in this window, that is, Q_m,n. Moving the trace window along the well and average Q_s for a set of overlapping windows can be calculated. Windows with a reasonable Q (e.g. the values inside a range from 20 to 200) are retained. By averaging these window Qs and calculating interval Q, that is, Q between each consecutive receivers, an initial approximation may be estimated. The initial approximation may be further improved by minimizing objective function, which is the squared average difference between calculated Qs and may be determined from trace windows.

To determine interval absorption coefficients α_k, the objective function F with respect to α_j is minimized:

$\begin{array}{cc}F\left(\alpha \right)=\sum _{m>k=1}^N{u_{\mathrm{km}}\left[A_{k,m}-\frac{1}{T_m-T_k}\sum _{j=k+1}^m{\alpha }_j\left(T_j-T_{j-1}\right)\right]}^2+\sum _{j=2}^N{w_j\left({\alpha }_j-{\beta }_j\right)}^2& \left(20\right)\end{array}$ where β_j are initial approximations for α_j. In eqn (20), weights w_j depend on quality of Q_k,m estimates. The weights u_k,m are used according to:

$\begin{array}{cc}{u}_{m,n}=\frac{1}{{\sigma }_{m,n}+{\sigma }_{\mathrm{Aver}}}& \left(21\right)\end{array}$ where σ_m,n is a standard deviation of Q estimate for the layer [z_k,z_m]; σ_Aver is an average standard deviation over all pairs. Weights w_j are for regularization purpose to keep solutions within given range. The optimization of F(α) is done iteratively: at the first iteration, all the weights w_j are the same. After minimization, the coefficients α_j may be checked against a given range, and the weights w_j of α_j that are outside the given range may be increased. In a second iteration, F(α) may be optimized by applying new weights w_j and again checking the output absorption coefficients α. If some of the coefficients α_j are outside the range, we increased corresponding weights and continue iterations until all α_j are within a given range. The function in eqn. (20), may be minimized by solving the linear system of equations:

$\begin{array}{cc}\frac{\partial F\left(\alpha \right)}{\partial {\alpha }_j}=0.& \left(22\right)\end{array}$

In eqn. (20), only those A_km, that are within a given range may be retained. If all Q_km are within a reasonable range, then there would be N(N−1)/2 values. Normally, about ⅓ of this number are within a reasonable range, that is, about 0.15N². For N=100 traces, there would be about 1500 values. Taking into account that different reference numbers that are used to estimate layered Q, there may be about 10000 input Qs to estimate 99 unknown intervals q_j. This provides sufficient statistics for a stable Q estimation. Resolution of Q estimation may depend on the length of intervals [z_k, z_m] that provide reasonable Q values.

Turning now to FIG. 5, some of the steps of one method 500 according to an embodiment of the present disclosure are illustrated. In step 501, Zero- or small-offset VSP data may be acquired and a wavefield separation may be used to separate the downgoing wavefield from the upgoing wavefield using known prior art techniques. For the purposes of the present disclosure, the term “zero-offset” includes data acquired using a source as close to the wellhead as practical. It should be noted that the upgoing wavefield may be a compressional wavefield or a shear wavefield.

In step 503, the amplitude spectra S_j(f) for all receivers, j=1, 2, . . . , N; (where N is the number of receivers) may be calculated for the downgoing wavefield. In step 505, an effective layered absorption coefficient A_m,n corresponding to the depth interval [z_m, z_n] may be estimated for each of the receiver pairs at depths z_m-z_n. This absorption coefficient A_m,n may be calculated by minimizing eqn. (11), where A_m,n=α. In step 507, another effective absorption coefficient A_m,n may be estimated using all possible triple depths z_r, z_n and z_m and eqn. (19).

In step 509, From calculated layered, effective absorption coefficients that are outside input range [A_min, A_max] of the calculated layered A_m,n may be dropped. It may be assumed that a reasonable range for the Q-factor has been selected, thus a range for effective absorption coefficient α may be calculated using eqn. (3):A_min=π/Q_max,A_max=π/Q_min.A reasonable range for the Q factor can be obtained from published literature and knowledge of the lithology of the subsurface that is expected.

In step 511, the above remaining effective absorption coefficients A_m,n may be used to calculate initial values for interval effective absorption coefficients α_k. For this, the average remaining layered effective absorption coefficients A_m,n with respect to common depth interval [z_k−1,z_k] may be used interval effective absorption coefficients α_k, corresponding to the interval [z_k−1,z_k], may be obtained.

In step 513, the initial values may be improved by minimizing objective eqn. (20) iteratively. In the first iteration, all the weights w_j may be the same. After minimization, α_j may be checked, and, for those α_j that are outside given range, the weights w_j may be increased for those α_j that are outside the given range. In the second iteration, eqn. (20) may be optimized with new weights w_j and the output absorption coefficients α_j may be checked again. If some of the coefficients α_j are outside the range, the corresponding weights may be increased and the iterations may continue until all α_j are within a given range.

The processing methodology described above may be implemented on a general purpose digital computer. As would be known to those versed in the art, instructions for the computer reside on a computer-readable memory device such as ROMs, EPROMs, EAROMs, Flash Memories and Optical disks. These may be part of the computer or may be linked to the computer by suitable communication channels, and may be even at a remote location. These are all examples of non-transitory computer-readable media. Similarly, seismic data of the type discussed above may be stored on the computer or may be linked through suitable communication channels to the computer. The communication channels may include the Internet, enabling a user to access data from one remote location and get the instructions from another remote location to process the data. The instructions on the computer-readable memory device enable the computer to access the multicomponent data and process the data according to the method described above.

While the foregoing disclosure is directed to the specific embodiments of the invention, various modifications will be apparent to those skilled in the art. It is intended that all such variations within the scope and spirit of the appended claims be embraced by the foregoing disclosure.

Claims

1. A method of evaluating an earth formation, the method comprising: acquiring seismic data using a plurality of seismic detectors, each at one of a plurality of spaced apart locations in a wellbore, responsive to activation of a seismic source near a top of the wellbore;estimating a spectrum of a down going wave field of the acquired seismic data at each of the plurality of seismic detectors by separating the down going wave field from an up going wave field of the acquired seismic data;estimating an absorption coefficient of the earth formation for at least one pair of the plurality of seismic detectors by minimizing an objective function based on an exponential relation between the spectra of the at least one pair of seismic detectors; andevaluating the earth formation using the estimated absorption coefficient to indicate at least one of (i) a lithology and (ii) a fluid content of the earth formation.

2. The method of claim 1 wherein the at least one pair of seismic detectors further comprises (i) a plurality of pairs of the plurality of seismic detectors at the spaced apart locations, and (ii) a plurality of pairs defined by a triplet of the plurality of seismic detectors.

3. The method of claim 2 further comprising eliminating a pair of seismic detectors that gives an absorption coefficient outside a predefined range.

4. The method of claim 1 further comprising using the estimated absorption coefficient as a starting value for obtaining an improved estimated absorption coefficient by minimizing an objective function F of the form:

5. The method of claim 1 further comprising using the estimated absorption coefficient for further processing of the acquired seismic data in formation evaluation.

6. The method of claim 1 wherein activation of the seismic source generates a seismic wave, the seismic wave being one of: (i) a compressional wave, and (ii) a shear wave.

7. A system configured to evaluate an earth formation, the system comprising: a plurality of seismic detectors, each positioned at a plurality of spaced apart locations in a wellbore, configured to acquire seismic data comprising a signal responsive to activation of a seismic source near a top of the wellbore; andat least one processor configured to: estimate a spectrum of the acquired signal of a down going wave field of the acquired seismic data at each of the plurality of seismic detectors by separating the down going wave field from an up going wave field of the acquired seismic data;estimate an absorption coefficient of the earth formation for at least one pair of the plurality of seismic detectors by minimizing an objective function based on an exponential relation between the spectra of the at least one pair of seismic detectors; andevaluate the earth formation using the estimated absorption coefficient to indicate at least one of (i) a lithology and (ii) a fluid content of the earth formation.

8. The system of claim 7 wherein the at least one pair of seismic detectors further comprises (i) a plurality of pairs of detectors at the spaced apart locations, and (ii) a plurality of pairs defined by a triplet of the plurality of seismic detectors.

9. The system of claim 8 wherein the processor is further configured further eliminate a pair of seismic detectors that gives an absorption coefficient outside a predefined range.

10. The system of claim 7 wherein the processor is further configured to use the estimated absorption coefficient as a starting value for obtaining an improved estimated absorption coefficient by minimizing an objective function F of the form:

11. The system of claim 7 wherein the processor is further configured to use the estimated absorption coefficient for further processing of the acquired seismic data in formation evaluation.

12. The system of claim 7 wherein the source is configured to generate a seismic wave selected from: (i) a compressional wave, and (ii) a shear wave.

13. A non-transitory computer-readable medium product having stored thereon instructions that when read by at least one processor cause the at least one processor to execute a method, the method comprising: using seismic data acquired by a plurality of seismic detectors, each at one of a plurality of spaced apart locations in a wellbore, responsive to activation of a seismic source near a top of the wellbore for estimating a spectrum of a down going wave field of the acquired seismic data at each of the plurality of seismic detectors by separating the down going wave field from an up going wave field of the acquired seismic data;estimating an absorption coefficient of the earth formation for at least one pair of the plurality of seismic detectors by minimizing an objective function based on an exponential relation between the spectra of the at least one pair of seismic detectors; andevaluating the earth formation using the estimated absorption coefficient to indicate at least one of (i) a lithology and (ii) a fluid content of the earth formation.

14. The non-transitory computer-readable medium product of claim 13 further comprising at least one of: (i) a ROM, (ii) an EPROM, (iii) an EAROM, (iv) a flash memory, and (v) an Optical disk.