BMA precipitation modeling
Fits a Bayesian Model Averaging mixture of gammas with a point mass at 0 to ensemble forecasts. Intended for predicting precipitation. Allows specification of a training rule and forecasting dates.
ensembleBMAgamma0( ensembleData, trainingDays, dates = NULL, control = controlBMAgamma0(), exchangeable = NULL)
ensembleData: An ensembleData object including ensemble forecasts with the corresponding verifying observations and their dates. Missing values (indicated by NA) are allowed.trainingDays: An integer giving the number of time steps (e.g. days) in the training period. There is no default.dates: The dates for which forecasting models are desired. By default, this will be all dates in ensembleData for which modeling is allowed given the training rule.control: A list of control values for the fitting functions. The defaults are given by the function controlBMAgamma0.exchangeable: A numeric or character vector or factor indicating groups of ensemble members that are exchangeable (indistinguishable). The models fit will have equal weights and parameters within each group. The default determines exchangeability from ensembleData.A list with the following output components: - training: A list containing information on the training length and lag and the number of instances used for training for each modeling day.
prob0coefs: The fitted coefficients in the model for the point mass at 0 (probability of zero precipitaion) for each member of the ensemble at each date.
biasCoefs: The fitted coefficients in the model for the mean of the gamma components for each member of the ensemble at each date (bias correction).
varCoefs: The fitted coefficients for the model for the variance of gamma components for each date. The coefficients are the same for all members of the ensemble.
weights: The fitted BMA weights for the gamma components for each ensemble member at each date.
power: A scalar value giving to the power by which the data was transformed to fit the models for the point mass at 0 and the bias model. The untransformed forecast is used to fit the variance model. This is input as part of control.
The output is for all of the dates in ensembleBMA, so there will be missing entries denoted by NA for dates that are too recent to be forecast with the training rule.
The following methods are available for ensembleBMAgamma0 objects: cdf, quantileForecast, modelParameters, brierScore, crps, CRPS and MAE.
ensembleData, controlBMAgamma0, fitBMAgamma0, cdf, quantileForecast, modelParameters, brierScore, crps, MAE
J. M. Sloughter, A. E. Raftery, T. Gneiting and C. Fraley, Probabilistic quantitative precipitation forecasting using Bayesian model averaging, Monthly Weather Review 135:3209--3220, 2007.
C. Fraley, A. E. Raftery, T. Gneiting and J. M. Sloughter, ensembleBMA: An R Package for Probabilistic Forecasting using Ensembles and Bayesian Model Averaging, Technical Report No. 516R, Department of Statistics, University of Washington, 2007 (revised 2010).
C. Fraley, A. E. Raftery, T. Gneiting, Calibrating Multi-Model Forecast Ensembles with Exchangeable and Missing Members using Bayesian Model Averaging, Monthly Weather Review 138:190--202, 2010.
data(ensBMAtest) ensMemNames <- c("gfs","cmcg","eta","gasp","jma","ngps","tcwb","ukmo") obs <- paste("PCP24","obs", sep = ".") ens <- paste("PCP24", ensMemNames, sep = ".") prcpTestData <- ensembleData( forecasts = ensBMAtest[,ens], dates = ensBMAtest[,"vdate"], observations = ensBMAtest[,obs], station = ensBMAtest[,"station"], forecastHour = 48, initializationTime = "00") ## Not run: # R check prcpTestFit <- ensembleBMAgamma0( prcpTestData, trainingDays = 30) ## equivalent to ## prcpTestFit <- ensembleBMA( prcpTestData, trainingDays = 30, ## model = "gamma0") ## End(Not run) # for quick run only; use more training days for forecasting prcpTestFit <- ensembleBMAgamma0( prcpTestData[3:16,], trainingDays = 6)
Useful links