Exploring the potential of GNSS radio occultation for the analysis of tropical width metrics

The tropics are expanding poleward as a result of anthropogenic climate change, this in turn has great implications on the temperature and precipitation patterns in the subtropical regions. Previous studies have found varying widening trends, most of which have been derived using reanalysis and climate model data including complex error characteristics due to their inherent choices. These discrepancies in the widening trends underline the interest for studies using alternative datasets with reduced uncertainty.

Here, we explore the potential of GNSS radio occultation (RO) data for the analysis of the tropical width, as an independent observational source of information. RO data has many advantages characteristics such as high accuracy, global availability and long-term consistency. Another valuable attribute is the high vertical resolution in the upper troposphere. These specific advantages make it highly interesting for the study of tropical atmospheric phenomena. The RO processing provides a high-quality temperature record over the period September 2002 to December 2020, with dense horizontal coverage since 2006.

To evaluate the skill of RO data to accurately capture the expansion of the tropics, we look at four different metrics which are commonly analyzed in studies of the tropical width: the tropopause break, the eddy driven jet and the subtropical jet. To calculate these metrics temperature and wind data is necessary. Synoptic-scale climatic winds from RO geopotential height records are being derived in the Wegener Center research group. These monthly wind records include as a novelty higher-order solutions of the wind-field equations. The metric results are compared to three state-of-the-art reanalysis datasets (i.e., ERA5, MERRA-2, and JRA-3Q). In this process, zonal patterns and the regional structure of the tropospheric features are investigated to further test the utility of RO in respect to its spatial robustness. Moreover, we aim to provide a perspective on the needed record length to calculate trends for these metrics.

Comparison to reanalyses shows good agreement between the different datasets for zonal average monthly mean values over the available RO record period. As for the analysis of the longitudinally resolved metrics, results from reanalyses and RO align well with some exceptions over the northern hemisphere. While the RO record 2006 to 2020 is yet too short for trend studies on tropical widening, we find overall encouraging results that suggest RO to be an alternative observation-based dataset to modern reanalysis. RO data will be valuable for future trend studies of tropical features, especially due to its long-term consistency and high accuracy.