b'Magnetics in the mountainsFeatureMagnetics in the mountains: An approximation for the magnetic response from topographyextended for 2.5 km at each end, beyond the area displayed in the sections. The 50 m vertical mesh discretisation of the topographic surface introduced some noise to the profile, but, for our purpose, this can be ignored. For those familiar with the highlands of PNG, the range of magnetic values arising from the topography is comparable with observed magnetic values. Indeed, for this particular area, the measured magnetic field and computed field from homogeneous ground with topography looked very similar, even down to their amplitudes.The importance of considering the topography in the interpretation is reinforced in Figure 2, which compares Kim Frankcombeobserved magnetic data against forward modelled topography Senior Consulting Geophysicist, ExploreGeofrom another project. The forward modelling was again done kim@exploregeo.com.au using uBCs MAGFOR3D.The images have purposely not been interpolated to blur the Anyone who has spent time looking at high resolutionpixel boundaries so that the resolution of the data can be more magnetic data from modern volcanic terrains will likely haveeasily compared. The observed data were collected on 100 m seen topographically induced magnetic anomalies. These arespaced flight lines by helicopter with a 30 m terrain clearance anomalies, not from a change in the magnetic properties of(60 m AGL in this area) and gridded at 20 m cell size. The the ground but from the topography alone. This is illustratedforward modelled topography was computed on a 30 m grid in Figure 1, which shows two north-south sections through awith an observation height 60 m above the ground surface with mountainous area in Papua New Guinea (PNG). a perfect drape (I=-30.7, D=-0.1, F=44300nT, k=0.01 SI). One could be forgiven for interpreting the observed data as dipolar These models were computed using uBCs MAGFOR3D. Theanomalies from bodies with some depth extent. However, modelling assumed an even susceptibility of 0.01 SI for thethe forward model suggests that the same response could be ground, an inclination of -28, declination of 5.3 and fieldcaused by a flat sheet with steep sides superimposed on a flat strength of 41,350 nT. The mesh size used was 100 x 100 x 50 m,surface. The correlation between the two is further illustrated in which under-samples the topography. However, at the time thisFigure 3, which shows a comparison between the two data sets exercise was undertaken, in 2013, the best DEM available wasalong a NS trending profile. The forward modelled topographic the 90 m SRTM so the compromise was felt justified. Our currentresponse has been re-scaled by a factor of 13 times to enable practice is to use the 30 m SRTM and 30 x 30 x 15 m voxels,it to share the Y axis with the observed data. The SRTM DEM unless a better DEM is available. topography is shown in blue for reference.The field was computed at points in the centre of each voxel.The shape and amplitude match between the two magnetic The high values in air in the north (red) and low in ground inprofiles is remarkable. In comparing the two magnetic profiles the south (blue) of the sections are edge effects due to notseveral things need to be remembered. Firstly the forward padding the model sufficiently in depth. This is in spite of themodel assumes a perfect drape while the observed data were mesh extending 8500 m from the top layer at 3500 m. Paddingacquired with the pilots best effort drape. Secondly, the SRTM Figure 1.Computed magnetic field on two south to north (left to right) sections through homogeneous ground. k=0.01 SI, I=-28 deg V:H=1.JuNE 2020 PREVIEW 34'