Secondly, utilizing Landsat TM/ETM+ images, during the course of the coastline secular variation, the questions of the spatial data organization and processing have been studied by the means of RS and GIS.

In this paper the ILS and BIH polar coordinates in the period from 1962.0 to 1979.0 are compared, and the secular drifts of ILS origin and ILS stations with respect to the BIH system in this period are derived.

The SLR derived secular change in the Moon's mean motion caused by the total tidal dissipation is -24.78 arc sec/century 2,agreeing very well with the result ((-24.9±1.0) arc sec/century 2) from the analysis of the lunar laser ranging data.

The principle, basic performances and specifications of variety of two-wavelenght EDM instruments in the world, and examples of determing secular strain changes of fault and monitoring activities of earthquake are given in this paper.

The frequency and damping parameter in the resonance excitation equation become as time-dependent parameters. Parameter resonance medel can provide a revision of visco-elasticity for computing Q. Then the mean value of Q is 63, which is increasing slowly at the rate of 0. 8 every secular Peried.

We assessed the secular trends in the serum lipid levels in Shanghai residents from 1973 to 1999.

Secular evolution of rotary motion of a charged satellite in a decaying orbit

Secular perturbations of the orbit are taken into account: they are caused by the second zonal harmonic of the geopotential.

Using the method of averaging, basic regularities of the secular evolution of rotary motion of a screened satellite are revealed.

Using double numerical averaging, the equations are constructed that describe the secular evolution of eccentricity and perihelion longitude of the asteroid orbit.

The local deviations from the reference system of BIH(1968)in the observed UT1 of 18 observatories during 1950-1977 are used to compute the secular variations of longitudes of all the stations and to determine the angular velocities of American plates relative to Eurasian. The coordinates of the centre of rotation have been given by Le Pichon (1968, 1971).

The local systematic correction of time observations in 17 observing stations for BIH system has been analysed by means of AR spectral technique. The results have showed that except annual and semi-annual term, there exist the periods of 7, 6, 3, 2 years in longitude variations. It should be considered when we want to keep long-term stability of universal time and to study secular variations of longitude.

The existense of the non—polar local latitude variations of the five I L S stations is proved by analysing the residuals mathematically. And the uncertainty of the conventional international origin (CIO) caused by such variations is also demonstrated, it is pointed out that the uncertain solutions of polar motion and the drifts of the stations are led by the uncertainty of CIO. As a results,the key to this problem is to determine or choosc a proper origin of rhe reference system. 1. Some results of matrix operation...

The existense of the non—polar local latitude variations of the five I L S stations is proved by analysing the residuals mathematically. And the uncertainty of the conventional international origin (CIO) caused by such variations is also demonstrated, it is pointed out that the uncertain solutions of polar motion and the drifts of the stations are led by the uncertainty of CIO. As a results,the key to this problem is to determine or choosc a proper origin of rhe reference system. 1. Some results of matrix operation For convenience, matrix symbols are used in this paper. And theorems of matrix opcration are introduced. The major one is theorem 1.3, that is, if A∈R~(m·n) and r(A)=n, we have r(I-A(A~Т A)~(-1) A)=m-n, where R~(m×n) is the mxn—dimension Euclidean space, A is a matrix with m columns and n lines, and r(A) stands for thc rank of A. 2. Errors analysis The formula L(t)=AX(t)+E(t)is often used for observational error eq- uations, where t is a time variable, L(t)∈R~m is the measured value taken the form of m—dimentional vector, E(t)∈R~m stands for the error of L(t) (include random and systematic ones), X(t)(R~n is the parameter to be determined, and A∈R~(m×n) is an known coefficient matrix. Let V(t)∈R~m express the residuals of the solutions solved with the least square method. (For brevity the variable t is often omitted in the formulas.) Then we have theorem 2.1 V=(I-A (A~Т A)~(-1)A~Т)E and theorem 2.2 if E is a m—dimentional normal distribution N(M, Б), then V is in the same form N (M_v,Б), and M_v=(I-A(A~Т A)~(-1) A )M, Б_v=(I-A(A~T A)~(-1) A~Т)Б(I-A(A~Т A)~(-1) A~Т), in which M, M_v∈R~m are the mathematical expectations, Б, Б_v∈R~(m×m) are the covariance matrixs and I∈R~(m×m) is a unit matrix. 3. The proof of the local non—polar latitude systematic variation With respect to the seriese Δφ(t) of latitude variations publishedby the residuals V(t) are obtained by using the least square method to the equation L(t)=Δφ(t)=AX, where The root—mean—square errors of yearly average in the observational latitude are reasonable, considered to be less than 0″.005, since these of monthly average are estimated to be better than 0″.01. Thus, suppose the covariance matrix of E, Б=0.005~2I. The test of significance is taken for V in the light of theorem 2.2. Consequently the level of significance is much less than 1% so that Mv=(?) (the average of V(t) ) is negated. It is confirmed that E contain the non—polar local systematic errors M(t) which vary with the time variable. We can regard the total M(t) as the drifts of the stations. 4. The uncertainty of CIO Because of M(t), the way to keep CIO by the five ILS stations can not be relized actually. Furthermore, the determination of CIO depends on that of M(t), and inversely, the ascertainment of M(t) is based on that of CIO. Therefore, CIO defined in such a way is undeterminable. 5. Solution and its reference system For the equations L=AX+M, the number of unknown quantities is three more than that of equations owing to M(t). The solutions are not able to be determined then. In accordence with the equations L=AX+M, it is easy to know that if M is certain due to the origin to be certain, we can obtain the solution X. It is the same the other way round. From known X. M and so the origin of the reference system can be gained. Thereby, referring to solutions of polar motion we must point out exactly which point is adopted as a origin. Formerly, the solutions X solved with the least square method were related to the origin which is based on that the drifts of stations M=V are certain. Yymi and Okuda and others tried to solve the problem of secular polar motion by separating M from V. However, as they failed to catch the essence of the problems, their works in fact have only been to transform the solution from the foregoing reference system to those ones in which other points are referred to be the origin but are not any better than the original one. Thus, the problem is the question of choosing a proper reference system actually, that is, defining a suitable origin. It is unable that to define a fixed point related to the earth surface because of the relative drifts among the respective parts of the earth crust.