源于美国scripps海洋研究所，可处理海洋可控源的正反演问题。

C----------------------------------------------------------------------- program runocc C----------------------------------------------------------------------- c C OCCAM 2.0: Steven Constable IGPP/SIO La Jolla CA 92093-0225 c Program Revision 2.01, 13 Jan 1993 c c runocc is the calling program for OCCAM v2.0. C C REFERENCES: CONSTABLE, PARKER & CONSTABLE, 1987: GEOPHYSICS 52, 289-300. C DEGROOT-HEDLIN & CONSTABLE, 1990: GEOPHYSICS 55, 1613-1624. C CONSTABLE, 1991: GEOPHYS. J. INT. 106, 387-388. C CONSTABLE, 1992: OCCAM DISTRIBUTION NOTES. C C REVISION HISTORY: C MARCH 1986, (VERSION 1.2) C OCTOBER 1987, (VERSION 1.3) C AUGUST 1988, (VERSION 1.4) C JANUARY 1989 (WITH THANKS TO C.DEGROOT-HEDLIN), (VERSION 1.5) C SEPTEMBER 1989, (VERSION 1.5.2) C APRIL 1992 (MORE THANKS TO CATHERINE AND ALSO WSE) (VERSION 2.0) C C IF YOU OBTAIN THIS CODE FROM A THIRD PERSON, PLEASE SEND YOUR NAME AND ADDRESS C TO S. CONSTABLE. YOU WILL THEN RECEIVE UPDATES, NEWS ON BUGS, ETC. C C VERSION 2.0 IS A RE-WRITE TO MAKE OCCAM INDEPENDENT OF THE DIMENSIONALITY C OF THE FORWARD PROBLEM, AND TO MAKE MAXIMUM USE OF DYNAMIC MEMORY ALLOCATION, C AS THE ARRAY SIZES FOR THE 2D PROBLEM ARE NOW GETTING TOO LARGE FOR COMFORT. C INCLUDE FILES ARE NOW USED TO STREAMLINE CHANGES OF DIMENSION. C IMPLICIT DOUBLE PRECISION HAS BEEN REMOVED, AGAIN FOR STORAGE REASONS. C CALLING PROGRAM IS NOW AN INTEGRAL PART OF THE PACKAGE AS ALL THE MODEL- C DEPENDENT STUFF HAS BEEN SPUN OFF. C C IF YOU OBTAIN THIS CODE FROM A THIRD PERSON, PLEASE SEND YOUR NAME AND ADDRESS C TO S. CONSTABLE. YOU WILL THEN RECEIVE UPDATES, NEWS ON BUGS, ETC. C C SUBROUTINES WHICH MUST BE SUPPLIED BY THE USER: C C FORMOD(NP,ND,PM,DP,DM), COMPUTES THE FORWARD FUNCTION FOR MODEL PM() AT C THE DATA PARAMETERS DP() AN

大地电磁反演程序fortran，key kerry写的

occam反演方法，计算速度快捷，方法通俗易懂。Fortran 程序

kerry-key编写的大地电磁Occam二维反演程序，程序简明易懂，是很好的参考模板，经测试可以使用，可以供相关工作人员使用学习

constalb编写的大地电磁Occam反演程序，程序简明易懂，是很好的参考模板，经测试可以使用，可以供相关工作人员使用学习

电磁响应的数据很难得到一个唯一解，但是我们又不可避免要找到一个简单模型来解释观测资料。我们建议这个模型尽可能简单、平滑，以便降低过度解释的可能，并且消除在简单模型中的造成武断不连续性。