Helioseismic kernels to infer large scale flows in the Sun

 The inference of internal properties of the Sun from surface measurements of seismic wave travel times is the goal of time-distance helioseismology. Wave travel times, as measured on the surface, will change if the wave encounters perturbations e.g. in the sound speed or flows, as it passes through the solar interior. The perturbation that is causing that travel-time shift can be imaged using functions called sensitivity kernels that relate shifts in measurements to underlying anomalies. The inference of large-scale structures e.g meridional circulation requires computing sensitivity kernels in spherical geometry. Though helioseismic computations in spherical geometry are expensive, I shall show in this talk how we overcome this limitation by performing the computation in parallel, thereby achieving a significant speedup in wall-clock time. Validation tests show that kernels can be computed very accurately using our method. This gives us confidence that these accurate sensitivity kernels might allow us to infer large-scale subsurface structure accurately.