EmissV : an R package to create vehicular and other emissions for air quality models

Air quality models need input data containing information about atmosphere (such as temperature, wind, humidity), terrestrial data (such as terrain, land use, soil types) and emissions. Therefore, the emission inventories are easily seen as the scapegoat if a mismatch is found between modelled and observed concentrations of air pollutants (Pulles & Heslinga, 2010). The anthropogenic emissions, especially vehicular emissions, are highly dependent on human activity and constantly changing due to various factors ranging from economic (such as the state of conservation of the fleet, renewal of the fleet and the price of fuel) to legal aspects (such as the vehicle routing).


Summary
Air quality models need input data containing information about atmosphere (such as temperature, wind, humidity), terrestrial data (such as terrain, land use, soil types) and emissions.Therefore, the emission inventories are easily seen as the scapegoat if a mismatch is found between modelled and observed concentrations of air pollutants (Pulles & Heslinga, 2010).The anthropogenic emissions, especially vehicular emissions, are highly dependent on human activity and constantly changing due to various factors ranging from economic (such as the state of conservation of the fleet, renewal of the fleet and the price of fuel) to legal aspects (such as the vehicle routing).
The EmissV is an R package that estimates vehicular emissions by a top-down approach, the emissions are calculated using the statistical description of the fleet at available level (National, State, City, etc).The following steps show an example of the workflow for calculating vehicular emissions, this emissions are initially temporally and spatially disaggregated, and then distributed spatially and temporally to be used as input in numeric air quality models such WRF-Chem (Grell et al., 2005).I. Total: emission of pollutants is estimated from the fleet (number, type and year of vehicles), vehicular activity (km/day) and emission factors (g/km) by pollutant for each interest area (cities, states, countries, etc) or alternatively the totals of some inventory can be used.

II.
Spatial distribution: the package has functions to read information from tables, georeferenced images (tiff), shapefiles (sh), openstreetmap data (osm), global inventories in NetCDF format (nc) to calculate point, line and area sources.

III. Emission calculation: calculates the final emission from all different sources and converts to model units and resolution.
IV. Temporal distribution: a set of hourly profiles that represents the mean activity (by hour and day of the week) calculated from traffic counts of toll stations located at São Paulo city available for apply in the emissions.
The package has additional functions for creating emissions from individual sources (including plume rise parameterizations) and to estimate the vehicular emissions of volatile organic compounds from exhaust (through the exhaust pipe), liquid (carter and evaporative) and vapor (fuel transfer operations).

Functions and data
EmissV counts with the folllwing functions:

Examples
The following example creates an area source for São Paulo State (Brasil).The vehicles function creates a data.framewith information about the São Paulo Fleet using data from (DETRAN, 2015), the emissionFactors creates a data.framewith emission factors for CO and PM (CETESB, 2015).The totalEmission calculates the total emissions of CO for these vehicles and this emission factors.The next 3 lines opens different data: wrf file, a raster and the area shapefiles.These data are the input for areaSouce that creates an area source based on an image of persistent lights of the Defense Meteorological Satellite Program (DMSP) for São Paulo and Minas Gerais (Brasil) states and finally the function emission calculates the CO emissions.
The R package EmissV is available at the repository https://github.com/atmoschem/EmissV.And this installation is tested automatically on Linux via TravisCI and Windows via Appveyor continuous integration systems.Also, EmissV is already on CRAN https: //CRAN.R-project.org/package=EmissV.
by area emission Emissions in the format for atmospheric models emissionFactors Tool to set-up vehicle emission factors gridInfo Read grid information from a NetCDF file lineSource Distribution of emissions by streets perfil Dataset with temporal profile for vehicular emissions plumeRise Calculate plume rise height pointSource Emissions from point sources rasterSource Distribution of emissions by a georeferenced image read Read NetCDF data from global inventories streedDist Distribution by OpenStreetMap street totalEmission Calculate total emissions totalVOC Calculate total VOCs emissions vehicles Tool to set-up vehicle data frame Sao_Paulo <-areaSource(shape,raster,grid,name = "Sao Paulo") # processing Sao Paulo area ... # fraction of Sao Paulo area inside the domain = 0.473260323300595 sp::spplot(raster::merge(drop_units(TOTAL$CO[[1]]) * Sao_Paulo, drop_units(TOTAL$CO[[2]]) * Minas_Gerais), scales = list(draw=TRUE),ylab="Lat",xlab="Lon",

Figure 1 :
Figure 1: Emissions of CO using nocturnal lights.

Figure 2 :
Figure 2: CO emissions ready for use in air quality model.