000			02788nam a22003137i 4500
003			UVAL
005			20240507115436.0
007			t||
008			190710s2019 ||||| |||| 00| 0 eng d
040			_aDIBRA _bspa _cUVAL _erda
084			_am
100	1		_9228446 _aDíaz Peña, Mailiu, _dautora, _eauthor.
245	1	0	_aStatistical calibration of temperature and wind speed forecasts in the central area of Chile / _cMailiu Díaz Peña.
264		1	_aValparaíso : _bUniversidad de Valparaíso, _c2019.
300			_a70 hojas.
336			_atext _btxt _2rdacontent
337			_aunmediated _bn _2rdamedia
338			_avolume _bnc _2rdacarrier
502			_6Doctor en Estadística.
520			_aThe improvement of the weather forecasts is essential around the world because it has an impact on dierent elds such as renewable energy, air quality, and other meteorological events. Ensemble prediction systems have been developed in the last years but their forecasts are often aected by biases especially when implemented on complex terrains. The main aim of this work is to improve the prediction of surface temperature and wind speed in the central zone of Chile by using statistical post-processing techniques. In order to do this, we rst generate an ensemble of nine members from WRF model using dierent initial conditions. Then, we present some statistical post-processing techniques such as the ensemble model output statistic (EMOS), the Bayesian model averaging (BMA), the quantile regression forest (QRF) and the ensemble copula coupling (ECC). The proposed calibration techniques are implemented with a regional and semi-local approach. Also, we include other variables from the WRF simulations to calibrate wind speed with the QRF model and we determine the importance of each predictor in the calibration model. As results of the two case studies (surface temperature and wind speed), we determine that EMOS with normal distribution provided better performance to calibrate the surface temperature in Santiago, and QRF including additional features is the best for calibrating the wind speed around Valparaso and Santiago de Chile. Also, both models provide better performances with a semi-local approach. Finally, some conclusions and new insights are given in order to develop bi-variate techniques.
650		4	_aTEMPERATURA ATMOSFERICA _913598.
650		4	_9121390 _aANALISIS BAYESIANO.
650		4	_aPRONOSTICO DEL TIEMPO POR ESTADISTICA _9103057.
700	1		_aNicolis, Orietta, _eProfesora guía _9203161.
700	1		_aMarín Aguado, Julio, _eProfesor advisor _9234071.
700	1		_aBaran, Sándor, _eProfesor co-advisor _9228443.
710	2		_aUniversidad de Valparaíso (Chile). _bFacultad de Ciencias. _bInstituto de Estadística _9228632.
942			_2ddc _c5
999			_c105478 _d105478