b'Magnetics in the mountainsFeatureDEM typically does a poor job of mapping the true shape of valleys, particularly where they are steep sided and finally, this is a profile through a three dimensional model so there will be contributions from off-profile sources.Re-scaling the forward model by a factor of 13 implies a ground susceptibility of 0.13 SI, which is within the range of values we could expect for a recent andesite flow in this area.As a result of this lesson, we always run a forward model of topography before interpreting new survey data in areas of high relief. Typically the run time for this might be of the order of a week to a month, which does not impact on overall production if the modelling starts at the same time as the survey design. However, a recent survey of over 45 000 line kms in Indonesia caused a rethink of this practice, as it was clear that it would take between 6 and 8 months to run the forward models, which would have to be run in panels because of the way the code uses memory.Integral equation based methods such as used by Potent were of no assistance, and although Parker had published a Figure 2.Comparison of observed magnetic data (left) and forward modelledFourier based method to compute the magnetic response topography (right). Images use a linear colour stretch and the graticule has 1over topography, it required a plane observation surface, km sides. The anomalies being discussed are truncated as they lie at the edge ofnot a draped one. While that could have been extended to the heliborne survey. compute multiple plane surfaces and so build a draped surface Figure 3.Comparison of the two data sets in Figure 2 along a NS profile. Observed data (red) and re-scaled forward modelled topography response (gold), both on the left axis in nT, and SRTM topography (blue). Right axis in metres and X axis in metres35 PREVIEW JuNE 2020'