Abstract:
Precipitable Water Vapour (PWV) is an important and essential atmospheric parameter with
significant implications for weather forecasting, climate monitoring, and disaster management.
The aim of this paper is to study the correlation between the Global Navigation Satellite System
(GNSS)-derived PWV from Belihuloya, Sri Lanka, and ECMWF ERA5 reanalysis data.
Six months of data were gathered through GNSS from the CORSnet reference station located at
Sabaragamuwa university of Sri Lanka and ECMWF ERA5 data. The GNSS monthly data were
converted to daily data and processed using Canadian Spatial Reference System Precise Point
Positioning (CSRS PPP) to obtain Zenith Wet Delay (ZWD). A statistical correlation exists between
the two datasets, as evidenced by Pearson Correlation Coefficients (PCC) ranging from
0.774 in May to 0.980 in February, with an average value of about 0.960 for six months. The
multiplication factor between GNSS-derived and ERA5-derived PWV was calculated as 6.350.
The Root Mean Square Error (RMSE) was calculated as 25.346. This implies systematic differences
between the two methods due to the lack of constants used to calculate the conversion
factor (Π) defined for the study area. The correlation coefficient shows a strong relationship;
therefore, confidence can be established in GNSS-derived PWV or atmospheric monitoring,
which is highly consistent with ERA5 data, since the discrepancies were mostly attributed to
variability in the atmosphere and processing methods of data. This represents an exciting avenue
for developing meteorological applications using GNSS-derived PWV. Future accuracy
gains will come from improved data processing and integration with other atmospheric parameters.
