| dc.description.abstract | In marine CSEM it is known that the real areal extent of a hydrocarbon reservoir will not necessarily be mapped correctly during inversion. Generally, the inversion will cause the reservoir to be undersized, as a result of loss of frequencies in the subsurface. The error margin of the under sizing, and how this depends on burial depth will therefore be investigated in this study. This was done by studying the significance level of the mean squared error for a 2D model where length and burial depth of the reservoir was varied. The mean square error was found by using forward modeling where data from a set reservoir length was compared to the data for reservoirs with varied length. The data was then inverted by using Gauss Newton, where the length of the imaged reservoir was compared to the result of the forward modeling. It was found by using a statistical significance level of 10 standard deviations the change in the data fitted well with the change in the imaged model. | |
