The use of high-resolution clock file improves the performance on short time scales. However, an issue remains about the long time scales. From the plots, it appears that the long term curvature of pulsars is determined by the detrending of residual clock offsets.
We need to improve this detrending. Possible strategies:
Change the window for the interpolation used in the detrending procedure
Find a reliable way to eliminate "bad" clock measurements.
The use of high-resolution clock file improves the performance on short time scales. However, an issue remains about the long time scales. From the plots, it appears that the long term curvature of pulsars is determined by the detrending of residual clock offsets. We need to improve this detrending. Possible strategies: