SAR&GPS delay comparision

[mehdidar...@gmail.com]

Dear David,

I estimated delay using ERA-I, ERA5, MERRA2 and WRFmodels (SAR correction not InSAR) with TRAIN (throught xyz files and before applying slant and conversion (pseudo-range to phase delay) factors) for S1 data. I am trying to compar ZWD and ZHD derived from SAR with with GPS counterparts. Initial results presented an high underestimation of ZWD and ZHD values for SAR with respect to GPS values. Concerning your previous post here, I increased the zref parameter from its default value of 15000 to 30000 in the aps_models_SAR.m files and reprocessed the data. The outcomes are as follows:

 ZWD:                       GPS          ERA-I          ERA5          MERRA2          WRF                   ZHD:                    GPS          ERA-I          ERA5          MERRA2          WRF     

              Std(cm)        2.6              3                2.5                 2.7                2.9                             Std(cm)        0.93          0.99               1                 9.6                1.4               

           

          RMSE(cm)         -               1.5               0.5                1.8                 3.3                          RMSE(cm)         -               1.9               0.6                1.8               1.5     

               Corr(R)         -               0.85             0.97               0.8                 0.3                              Corr(R)         -               0.97             0.96               0.93             0.19

I wanted to know your opinion that to what extetn do you think the results are valid and reseanable? In case WRF, setting zref=30000 provided an overestimation for ZWD and ZHD, and above results for WRF obtained by zref=26000.

Since the results are zref-parameter dependent, finding the proper height is challenging, especially when I am going to cross-validate the results with GPS.

I compared GACOS (ZTD) product with my GPS (ZTD) and SAR (ZTD), showed a good agreement, but I do not know what zref parameter has been used to produce the GACOS (ZTD).

Thanks for your suggestions and recommendations in advance.

Regards,

Mehdi