THE APPLICATION OF THE UPWARD CONTINUATION METHOD IN THE CALCULATION OF THE VERTICAL DERIVATIVE TO IDENTIFY THE STRUCTURAL BOUNDARIES USING GRAVITY DATA

• Main Author: Toifur, Ahmad
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/71300
• Institution: Institut Teknologi Bandung

In processing gravity data, techniques that can facilitate the interpretation process are needed. One technique that can be used is the vertical derivative. This technique can be used to enhance the perception of shallow anomalies and explain their boundaries. The derivative in the vertical direction can be calculated through two events, first by utilizing the Laplace equation if only CBA maps are available. However, if gravity values for the same point at different elevations are available, the derivative in the vertical direction can be calculated directly. To obtain gravity values at different elevations, two events can be done, namely calculating its gravity response directly at different elevations or using upward continuation. In this research, experiments were conducted with several synthetic models that have variations in horizontal position, depth, and density contrast. Then, an analysis and comparison were made between the vertical derivative using direct response data and using upward continuation. The result, both ways produced maps with patterns that can show the deviation of the body. However, the second derivative calculated using direct response produced a map that was less good than using upward continuation, this was caused by rounding errors in the calculation using direct response. In addition, the effect of data density variations and the addition of noise on the vertical derivative map was also tested. On the map with a spacing of 400 m, an error of 67% was found in identifying the body boundary compare to the map with a spacing 100 m. Meanwhile, on the vertical derivative with the addition of normal distributed noise with a standard deviation of 0.005 mGal, an error of 30% was obtained after filtering or smoothing using a moving average with a window width of 5x5