New versions of the BDS/GNSS zenith tropospheric delay model IGGtrop

The initial IGGtrop model proposed for Chinese BDS (BeiDou System) is not very suitable for BDS/GNSS research and application due to its large data volume while it shows a global mean accuracy of 4 cm. New versions of the global zenith tropospheric delay (ZTD) model IGGtrop are developed through further investigation on the spatial and temporal characteristics of global ZTD. From global GNSS ZTD observations and weather reanalysis data, new ZTD characteristics are found and discussed in this study including: small and inconsistent seasonal variation in ZTD between (Formula presented.) and stable seasonal variation outside; weak zonal variation in ZTD at higher latitudes (north of (Formula presented.) and south of (Formula presented.) ) and at heights above 6 km, etc. Based on these analyses, new versions of IGGtrop, named (Formula presented.) , are established through employing corresponding strategies: using a simple algorithm for equatorial ZTD; generating an adaptive spatial grid with lower resolutions in regions where ZTD varies little; and creating a method for optimized storage of model parameters. Thus, the (Formula presented.) models require much less parameters than the IGGtrop model, nearly 3.1–21.2 % of that for the IGGtrop model. The three new versions are validated  by five years of GNSS-derived ZTDs at 125 IGS sites, and it shows that: (Formula presented.) demonstrates the highest ZTD correction performance, similar to IGGtrop; (Formula presented.) requires the least model parameters; (Formula presented.) is moderate in both zenith delay prediction performance and number of model parameters. For the (Formula presented.) model, the biases at those IGS sites are between (Formula presented.) and 4.3 cm with a mean value of (Formula presented.) cm and RMS errors are between 2.1 and 8.5 cm with a mean value of 4.0 cm. Different BDS and other GNSS users can choose a suitable model according to their application and research requirements.