Analysis of datum-instability effect on calculated results of data from Longmen Mountain r

Analysis of datum-instability effect on calculated results of data from Longmen Mountain regional

gravity network

Sun Shaoan1，2and Zhou Xin31Institute of seismology，China Earthquake Administration，Wuhan 430071，China2Crustal Movement Laboratory，Wuhan 430071，China3Laboratory of Computational Geodynamics，College of Earth Science，Graduate University of Chinese Academy of Sciences，Beijing 100049，China

【摘 要】Abstract:A statistical correlation method is used to study the effect of instability of the calculation datum (used in traditional method of indirect adjustment)on calculated gravity results，using data recorded by Longmen Mountain regional gravity network during 1996－2007.The result shows that when this effect is corrected，anomalous gravity changes before the 2008 Wenchuan Ms8.0 earthquake become obvious and characteristically distinctive.Thus the datum-stability problem must be considered when processing and analyzing data recorded by a regional gravity network. 【期刊名称】大地测量与地球动力学（英文版） 【年(卷),期】2011(002)004 【总页数】6

【关键词】Key words:regional gravity network;classic indirect adjustment;gravity datum;Wenchuan Ms8.0 earthquake

1 Introduction

Extracting precursory information related to earthquakepreparation process from a large amount of repeatedlymeasured gravimetric data is the prerequisite for using gravity data to predict earthquakes.Gravity difference between survey segments and Gravity Relative Values (GRV)of the stations are two basic quantities from which the extraction is made.Difference between Survey Segments(DSS)usually refers to the difference of gravity values between two adjacent stations in the gravity network.Being differential in nature，it is less easily disturbed，and is thus mainly applied to fault-activity monitoring and recognition.Gravity Relative Values(GRV)refers to the difference between the gravity observed at a station and the gravity datum in the network.By using GRV，it is not only possible to study temporal variation at the station，but also may study the temporal and spatial gravity variation in a region by combining the observations at many stations located in that region.This method is presently preferred in studying the temporal and spatial variations of a regional gravity field［1，2］.However，GRV is easily disturbed by such factors as datum instability，size and shape of the gravity network，system error，and error propagation.The problem with datum instability comes from the fact that a relative stable station selected as the reference is assumed to have no temporal gravity change，which is not true in reality.In this paper，we present a method

to minimize the disturbance caused by the instability of datum in a gravity network in order to discover true regional gravity changes.

2 Longmen Mountain regional gravity network

The Longmen Mountain regional gravity network con-sists of stations along a closed loop and a line.The stations in the loop are distributed along both sides of Longmenshan fault between Beichuan in the northeast and Yingxiu and Wenchuan in the southwest.The line connects the loop with the gravity-datum station in Chengdu(Fig.1). This network was established and first surveyed in 1996.It consisted of 29 stations initially，and some stations were added later without changing the network shape.Gravity measurements were made annually，except in 2004 when two measurements were made.A total of 13 surveys were made up to the end of 2007，after which the network was seriously damaged by the Ms8.0 Wenchuan earthquake. The

measurement

was

made

with

LaCoste-Romberg

G(LCR-

G)gravimeters.The measurement precision is better than 15×10－8ms－2and the mean station gravity precision is around 15×10－8ms－2，thus capable of detecting non-tidal gravity changes caused by tectonic activities［3］.

3 Stability analysis of gravity datum and results

To extract information of gravity-field changes from gravity data may be done in two steps:1)pre-processing and relative-gravity adjustment for

the data;2) minimizing various disturbances on gravity data，establishing gravity datum for the network，and calculating temporal gravity changes at individual sites during individual surveys.Step 1 is carried out by using the standard LGADJ software package［4，5］.The pre-processing includes such corrections as solid-earth-tide gravity variation，atmosphere-pressure variation，linear scale，height of gravimeter，incidentals，and gravimeter drift.The relative-gravity network adjustment is done by using DSS data of individual stations as elements and by using classical indirect adjustment.In doing so，there is a need of designating a certain station in the network with known gravity value as the datum.If absolute gravity value is not known at the datum station，it is usually assumed to be 0.This is the main method of data adjustment for almost all regional relative gravity networks，since there is no simultaneous absolute-gravity measurement with relative-gravity observation at the present.This practice would undoubtedly generate disturbance on the data of gravityfield changes，because the earth’s gravity field is a changing parameter especially in tectonically active regions.However，this problem is rarely mentioned in the literature.It will be discussed in detail here.

Regarding the DSS results of a gravity network，the gravity value at any station is related to the value at the datum station Gas follows，

where Gis the gravity value of the mth station for a certain survey and

Δg，0is the DSS between the mth station and the datum station.It is obvious that gravity values of all stations depend on the datum’s gravity value.The effect of ignoring datum’s variation(by setting the value to 0)should not be ignored in search of true gravity changes in a regional gravity network.

Now let us use the data of Longmen Mountain relative-gravimetric network as an example to illustrate the effect of datum variation on relative-gravity values at individual stations of the network through a simple correlation analysis.Firstly，we adjust all gravity data of the 13 periods from 1996 to 2007，based on No.1 station(located about 20 km from Chengdu)as datum of the gravity network(with no gravity change during measurement periods).Secondly，we use the gravity values of the first survey in 1996 as the reference to calculate gravity variations at the individual stations and their DSS in subsequent periods.

Figure 2 shows，as examples，individual gravity values in 1997 and 1998 relative to 1996，respectively.In this figure，the abscissa denotes the gravity value of every station relative to the datum in 1996.From figure 2，we may see that the gravity differences basically follow a zonal distribution parallel to the X axis，and there is a linear relation between two periods.From equation(1)，we may derive a relation between the inter-survey gravity changes at a station and the datum station as follow，