Banca de QUALIFICAÇÃO: AUGUSTO DE RUBIM COSTA GURGEL

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
STUDENT : AUGUSTO DE RUBIM COSTA GURGEL
DATE: 30/06/2022
TIME: 09:00
LOCAL: VODEO CONFERENCIA
TITLE:

Analysis of wind energy from ensembles of regional climate models for northeastern Brazil

 

KEY WORDS:

Wind Speed; ECMWF-ERA5 reanalysis; Regional Climate Models; CORDEX; wind potential; bias correction; ensemble; Principal Component Analysis; Weighted Averages.

 

PAGES: 64
BIG AREA: Ciências Exatas e da Terra
AREA: Geociências
SUBÁREA: Meteorologia
SPECIALTY: Meteorologia Aplicada
SUMMARY:

The significant increase in energy demand has caused clean energy sources to be further exploited worldwide. In Brazil, in particular, the development of wind and solar energy was observed mainly in the Northeast region (NEB). Thus, the objective is to generate ensembles from high resolution Regional Climate Models (RCM) to project the trend of wind speed and wind potential related to the climates of the recent past (1986-2005), near future (2041-2060) and distant future (2080-2099) for NEB. The data came from Xavier et al. (2016) (wind speed and precipitation), ECMWF reanalysis (ERA 5) (wind speed and precipitation) and MCR belonging to the CORDEX (wind speed) project. Thus, the reanalysis of the ECMWF (ERA 5) from Xavier et al. (2016) was evaluated through statistical metrics for wind speed and precipitation, as well as MCR and its ensembles generated by arithmetic means that were evaluated from the ECMWF (ERA 5) for wind speed data. In addition, this variable was analyzed for some subareas of NEB, as well as hydrowind complementarity. Two Representative Concentration Pathways (RCP), RCP 2.6 (greater mitigation) and RCP 8.5 (lower mitigation) were used to project wind speed to near and far future climates, from Mann-Kendall, Theil-Sen and Pettitt tests. Differences in wind speed were made between the climates of the near and distant future with the climate of the recent past. Bias correction techniques (bias correction), Principal Component Analysis and Weighted Averages will be used to generate more robust ensembles than those generated by arithmetic means. The reanalysis of the ECMWF (ERA 5) satisfactorily represented the data of Xavier et al. (2016) for wind speed, with an average correlation between 0.4 and 0.8, bias below 1m/s in almost any region and Root of Mean Quadratic Error below 2.1 m/s. In addition, good results were observed for precipitation, with correlation above 0.8, bias below 200mm annually and Root of Mean Quadratic Error below 15mm annually for most NEB. Ensembles generated by arithmetic means when compared to the individual simulations of each MCR satisfactorily represented the reanalysis of the ECMWF (ERA 5) for wind speed. However, they overestimated wind speed, mainly in the second half of the year (June to December) with bias higher than 2m/s for almost the entire NEB. For some subareas of the NEB, the ensembles present a good correlation (0.8), standard deviations close to the reanalysis, with a maximum value of 2.4 m/s when compared to 1.41 m/s of the reanalysis and the Root of the Mean Quadratic Error maximum of 1.7m/s. The four subareas analyzed in the NEB have average velocities higher than 5m/s, being considered good for wind power generation. A characteristic of hydro wind complementation was observed, because the periods of lower precipitation contrast with periods of maximum wind speeds. There was a positive trend for wind speed in all subareas studied with both RCP and speed differences that may be greater than 15% when compared to the climates of the near and distant future with recent past, therefore, a continuity in good wind energy production is expected.

BANKING MEMBERS:
Presidente - 1914304 - KELLEN CARLA LIMA
Externo à Instituição - DOMINGO CASSAIN SALES - FUNCEME
Externa à Instituição - MAYTÊ DUARTE LEAL COUTINHO - INMET