Evaluation of the physical parametrizations of the WRF model for the atmospheric layer of the Metropolitan Region of the Great Victory.
Name: Juan Felipe Medina Velásquez
Type: MSc dissertation
Publication date: 05/04/2017
Advisor:
Name |
Role |
|---|---|
| Jane Meri Santos | Advisor * |
| Neyval Costa Reis Jr. | Co-advisor * |
| Taciana Toledo de Almeida Albuquerque | Advisor * |
Examining board:
| Name |
Role |
|---|---|
| Davidson Martins Moreira | Internal Examiner * |
| Jane Meri Santos | Advisor * |
| Luiz Claudio Gomes Pimentel | External Examiner * |
| Neyval Costa Reis Jr. | Co advisor * |
| Taciana Toledo de Almeida Albuquerque | Advisor * |
Summary: The main objective of this work is to evaluate the performance of the different PBL physical parameterizations available in the Weather Research and Forecasting model (WRF) 3.6.1., in order to identify which one best represents the meteorological conditions of the Metropolitan Region of Grande Vitória (RMGV) in the two most representative periods of the region, winter and summer. The achieve this goal, were made a total of 34 simulations, 12 for the winter period (07/2010) e 12 for the summer period (02/2016), with which all the PBL parameterizations were evaluated, except QNSE (Quasi-normal Scale Elimination) and the MYNN (Mellor-Yamada Nakanishi Niino) level 2.5 and 3, with their respective CLS parameterization available in the model. For these simulations, tow nested domains were used WHERE the major domain has a spatial resolution of 5 km, forming a 5 x 5 km domain with 49 x 49 cells covering the entire state of Espírito Santo, part of the Minas Gerais, Rio de Janeiro and Bahia and the smaller domain has a spatial resolution of 1 km, forming a 1 x 1 km domain with 120 x 120 cells comprising the entire RMGV. both with a vertical structure represented by 21 vertical layers and centered at the coordinates 20,25 ° S and 40,29 ° W.
In order to achieve the main goal, the model data obtain with each parameterization were compared with the data measured in the stations belonging to RMAQAR and the airport of surface temperature (2 m), wind and speed direction (10 m), using the statistical parameters mentioned in subsection 4.3 of this dissertation. The results showed that the parameterization that best represented the values of the meteorological variables previously mentioned for the summer period was the parameterization used in the modeling M_1 that corresponds to the YSU schemes for CLA and the improved MM5 for the CLS, with the Carapina station presenting simulated values closer to the real values observed. On the other hand, for the winter period, the parameterization that best represented the values of the mentioned meteorological variables was the parameterization used in the modeling M_12 that corresponds to the schemes UW for the CLA and the MM5 for the CLS, being Cariacica and airport the stations that presented an accuracy between the simulated data and the real data observed. The results presented by both parameterizations show that the best results are presented for wind speed, followed by surface temperature and wind direction. These results suggest the need to test the other physical parameterizations available in the model in order to improve the prediction results of the meteorological variables for the RMGV and thus to have better results at the moment of using this data in dispersion models. Key words: WRF, Planetary Boundary Layer, RMGV, physical parameterizations.
