**5. Conclusions**

The combined analysis of PWV using GPS-based estimations, MODIS satellite products, and atmospheric sounding in the Pacific region of Costa Rica provides the first comparison between di fferent water vapor calculation techniques for the Central American region. The evaluation of GPS-based estimates of PWV confirms the good performance of these estimations in comparison to the traditional and standard technique based on radiosondes, with no significant di fferences during the dry and wet seasons. These first results demonstrate the feasibility of incorporating GPS-based meteorological applications in order to improve the study of moisture inputs on the seasonal water vapor distribution in Central America. However, the performed evaluation identified significant biases between the GPS PWV estimates and the MODIS Aqua PWV estimations under both dry and wet season conditions and only the MODIS Terra PWV estimations recorded during the dry season were not significantly biased relating to the GPS PWV calculations. These results open the opportunity to evaluate other satellite products that provide higher spatial and temporal resolutions in order to give better insights into the causes of disagreements. Our analysis was also able to identify the influence of the main circulation patterns in Costa Rica, namely the trade wind regime and the ITCZ passage on PWV variability, which resulted in the relatively greater variability of the smaller PWV values during the dry season in comparison to the relatively smaller variability of the greater PWV values observed during the wet season. The influence of these moisture transport patterns was identified using the HYSPLIT analysis done for the Central Valley of Costa Rica. The multiple linear regression model successfully applied to this region can simulate the seasonal PWV variations using major meteorological variables, namely the mean daily air temperature, the mean daily relative humidity, the mean daily downward solar radiation flux, and the mean daily atmospheric pressure. We consider that a further analysis based on hourly GPS data could better analyze these relations between water vapor and HYSPLIT calculations and could refine the mathematical modeling presented in this work.

**Author Contributions:** Methodology, G.E.-H., J.F.V-C., and J.M.-Z.; software, J.F.V-C., S.R.-R., and J.M.-Z.; validation, P.C.-A. and S.R.-R.; writing—original draft, P.C.-A. and G.E.-H.; writing—review and editing, R.S.-M. and J.B.

**Funding:** This research received no external funding.

**Acknowledgments:** Germain Esquivel-Hernández, José Francisco Valverde-Calderón, and Ricardo Sánchez-Murillo thank the Research Office of the National University of Costa Rica through Grant SIA-0457-16.

**Conflicts of Interest:** The authors declare no conflict of interest.
