An R package for time series based extensions of Ensemble Model Output Statistics (EMOS) as described in the references.
It depends on the R-packages:
You can install the development version from GitHub with:
# install.packages("remotes")
remotes::install_github("jobstdavid/tsEMOS")Below is an overview of all functions contained in the R-package for model estimation and prediction:
-
semos: smooth EMOS (SEMOS). -
dar_semos: deseasonalized autoregressive smooth EMOS (DAR-SEMOS). -
dargarchmult_semos: multiplicative deseasonalized autoregressive smooth EMOS with generalized autoregressive conditional heteroscedasticity (DAR-GARCH-SEMOS ($\cdot$ )). -
dargarchadd_semos: additive deseasonalized autoregressive smooth EMOS with generalized autoregressive conditional heteroscedasticity (DAR-GARCH-SEMOS (+)). -
sar_semos: standardized autoregressive smooth EMOS (SAR-SEMOS).
# load package
library(tsEMOS)
#> Registered S3 method overwritten by 'quantmod':
#> method from
#> as.zoo.data.frame zoo
# load data for station Hannover
data(station)
# select data for lead time 24 hours
data <- station[station$lt == 24, ]
# split data in training and test data
train <- data[data$date <= as.Date("2019-12-31"), ]
test <- data[data$date > as.Date("2019-12-31"), ]fit <- semos(train = train,
test = test,
doy_col = 3,
obs_col = 9,
mean_col = 10,
sd_col = 11,
n_ahead = 0)fit <- dar_semos(train = train,
test = test,
doy_col = 3,
obs_col = 9,
mean_col = 10,
sd_col = 11,
n_ahead = 0)fit <- dargarchmult_semos(train = train,
test = test,
doy_col = 3,
obs_col = 9,
mean_col = 10,
sd_col = 11,
n_ahead = 0)fit <- dargarchadd_semos(train = train,
test = test,
doy_col = 3,
obs_col = 9,
mean_col = 10,
sd_col = 11,
n_ahead = 0)fit <- sar_semos(train = train,
test = test,
doy_col = 3,
obs_col = 9,
mean_col = 10,
sd_col = 11,
n_ahead = 0)Feel free to contact jobstd@uni-hildesheim.de if you have any questions or suggestions.
Jobst, D., Möller, A., and Groß, J. (2024). Time Series based Ensemble Model Output Statistics for Temperature Forecasts Postprocessing. https://doi.org/10.48550/arXiv.2402.00555.