We analyse ambient-noise seismic data from 23 three-component seismic nodes to study firn velocity structure and seismic anisotropy near the EastGRIP camp along the Northeast Greenland Ice Stream (NEGIS). Using nine-component correlation tensors, we derive dispersion curves of Rayleigh and Love wave group velocities from 3 to 40 Hz. These velocity distributions exhibit anisotropy along and across the flow. To assess these variations, we invert dispersion curves for shear wave velocities (Vsh and Vsv) in the top 150 m of the NEGIS using a Markov chain Monte Carlo approach. The reconstructed 1-D shear velocity model reveals radial anisotropy in the firn, with Vsh 12 %–15 % greater than Vsv, peaking at the critical density (550 kg m‑3). We combine density data from firn cores drilled in 2016 and 2018 to create a new density parameterisation for the NEGIS, serving as a reference for our results. We link seismic anisotropy in the NEGIS to effective and intrinsic causes. Seasonal densification, wind crusts, and melt layers induce effective anisotropy, leading to faster Vsh waves. Changes in firn recrystallisation cause intrinsic anisotropy, altering the Vsv / Vsh ratio. We observe a shallower firn–ice transition across the flow (≈ 50 m) compared with along the flow (≈ 60 m), suggesting increased firn compaction due to the predominant wind direction and increased deformation towards the shear margin. We demonstrate that short-duration (9 d minimum), passive, seismic deployments and noise-based analysis can determine seismic anisotropy in firn, and we reveal 2-D firn structure and variability.