Uncertainty in the wet tropospheric path delay correction (WPD) has a one-to-one mapping to sea surface height uncertainty. For SWOT, the WPD correction is performed using a dual-beam radiometer operating between 18-34 GHz. The two radiometer beams are pointed a few degrees off-nadir to either side of the ground track and roughly centered in each of the two KaRIn swaths. The radiometer provides path delay only at these two discrete locations in the swath. A separate algorithm is required to map those two measurements into the best representation of the water vapor field over the swath. We will discuss a performance evaluation of several algorithms to map the two measurements to a 2D swath product. We assess the 2D SWOT path delay performance relative to other co-incidence observations from imaging sensors.

We first inter-compare two simple interpolation approaches, linear and nearest neighbor interpolation, which are the highest order interpolators that can be considered with just two beams. The nearest neighbor approach simply assigns the along-track WPD from each radiometer beam to the respective 60km swaths on either side of nadir. The other approach is to linearly interpolate across-track between the WPD samples from each of the two swaths that are spatially closest. The linear algorithm requires extrapolation at the swath edge beyond the boresight of the radiometer, roughly from 30-60km across-track.

The interpolation approaches are only able to remove linear variations in path delay which can led to cm-level errors in the presence of certain atmospheric features. More advanced approaches are required to reduce uncertainties when the cross-track WPD is highly non-linear. Morphing algorithms are class of algorithm that have been developed to fill in data gaps between satellites observations globally using some knowledge of how the particular quantity of interest transforms in time and space. Water vapor is particularly well suited for a morphing algorithm as it is nearly a conserved quantity on hourly times scales, meaning it advects with the mean atmospheric flow. This is particularly true for larger scale structures. We show a comparison of the advective morphing approach of Wimmers and Velden (2011) to the WPD from the simple interpolation algorithms.