Emissions are the pollutant mass released into the atmosphere [@pulles2010art]. The origin of the emissions can be human-made or anthropogenic or biogenic. The consequences of this pollution are complex affecting the atmosphere, human health, ecosystems, and infrastructure [@SeinfeldPandis2016]. In fact, pollution caused 9 million premature deaths in 2015, 16% of all deaths worldwide [@landrigan2017lancet].

An important tool for policy decision is air quality models. They have been used not only to study the impact of different emissions scenarios for policy making but also to understand the dynamics of air pollutants in various parts of the world [@Andradeetal2015]. The inputs for an air quality models are meteorology and emissions. Currently, there are tools for developing emissions inventories such as the VEIN [@vein] and the EmissV models [@emissv]. However, the existing tools for inputting the emissions into the air quality models are not written in a user-friendly way, such as PREP-Chem written in Fortran and C [@freitas2011prep]. Also, as the R language has a growing community, including statistical R packages for model evaluation and validation, such as openair [], air quality modelers already familiar with R would tend to use 'eixport' for inputing data into the appropiate format for each model without the need of deep knowledge in a specific model language. Therefore, we developed eixport, a tool for inputting data into atmospheric models using R [@R].

eixport imports functions form the R packages sf [@sf] which provides functions for spatial vector data, providing bindings to the GDAL, GEOS, and Proj.4 C++ libraries. Also, eixport import functions from the package ncdf4 [@ncdf4], which interface to Unidata netCDF Format Data Files, and from the raster package [@raster], which provides functions to gridded data.

eixport count with the following functions:

The following example creates a directory EMISS and then create a wrfchem input in that file. The package already counts with wrfinput files required to run eixport and create inputs for WRF-Chem. The line data(Lights) load a matrix of night light to spatially distribute the emissions. The perfil argument is used to temporally distribute the emissions. Lastly, the function to_wrf in one line reads the 1521983 $t \cdot y^{-1}$ of CO, spatially distribute it using nighttime traffic matrix Lights, temporally distribute it with the perfil, injecting the array of emissions directly into the wrfchemi file. The colour palette is "mpl_inferno" from the R package cptcity [@cpt].

library(eixport)
dir.create(file.path(tempdir(), "EMISS"))
wrf_create(wrfinput_dir = system.file("extdata", package = "eixport"),
          wrfchemi_dir = file.path(tempdir(), "EMISS"))

# get the name of created file
files <- list.files(path = file.path(tempdir(), "EMISS"),
                   pattern = "wrfchemi",
                   full.names = TRUE)

# load end write some data in this emission file
data(Lights)
to_wrf(Lights, files[1], total = 1521983, names = "E_CO")

wrf_plot(files[1], "E_CO")
# [1] "EMISS/wrfchemi_d01_2011-08-01_00:00:00"
# [1] "E_CO"
# [1] "Max value: 26.6966304779053, Min value: 0"

The resulting plot can be seen in the Fig. 1.

The R package eixport is available at the repository https://github.com/atmoschem/eixport. To ensure the usability of the package, in any commit to GitHub, eixport is installed in Ubuntu via Travis-CI (https://travis-ci.org/atmoschem/eixport) and Windows via Appveyor (https://ci.appveyor.com/project/Schuch666/eixport). Also, eixport is already on CRAN https://CRAN.R-project.org/package=eixport. Moreover, this packages tests functions with the suite CodeCov (https://codecov.io/) and the r package covr [@covr], achieving 89% of coverage (https://codecov.io/github/atmoschem/eixport).

\pagebreak

The development of eixport was supported by postdoc grans fro the Fundação de Universidade de São Paulo and Fundação Coordenação de Aperfeiçoamento de Pessoal de Nível Superior.