This study aimed to evaluate the field scale potential of Microbially Induced Desaturation (MID) to improve the compactibility of silty sands. Conventional soil compaction techniques for saturated silty sands often require either high compaction effort or even fail to reach aspired relative densities. This thesis is a follow-up study based on the promising small scale results of Stals 2020. The proposed solution consists of a two-stage method in which, firstly, a soil sample was desaturated by means of Microbially Induced Desaturation (MID) through denitrification and, secondly, dynamically compacted. The objective of this, and previous, research was to assess whether biogenic gas formation could be used as a pretreatment to increase the effectivity of soil compaction techniques. Tests with a height of 1m and 2.5m were developed, opposed to the previous 15cm small scale tests. Series of tests were conducted on a silt-sand mixture of 16% fines. Each series was compacted with both a different compaction method and energy input. The investigated scale effects included: the desaturation stage, bubble growth, bubble distribution, gas migration and final degree of compaction. All treated soils were able to desaturate to, and even below the optimum water content. In this research the in-situ shear strength was linked to swelling and bubble production. Based on a qualitative inspection, the gas distribution in the sample was found to be homogeneous over depth. In order to estimate swelling and venting of the sample during the desaturation stage, a model was produced based on a saturation dependent gas conductivity. It was suggested that a continuous gas phase was responsible for the highest amounts of compactive strain. With a non-intrusive compaction method, more compactive energy led to faster, but not more, release of gas. All 2.5m tests ended with a residual amount of gas. A 1m test series, consisting of three tests resulted in higher degrees of compaction for the treated tests opposed to the untreated tests. The final relative density of the treated tests was at least two times higher than the final relative density of the untreated test. A 2.5m test, which included surcharge, had a final RD 2.8 times higher opposed to a similar untreated test without surcharge. Taking into account the uncertainties of the results, it was concluded that a field trial is feasible. A hypothetical field case was set up of 500 m3. The costs of the proposed method were a factor 2.2 times cheaper opposed to the current soil improvement for silty
sands, namely a stone column vibroreplacement technique.