target_plot: Target plot
In jobonaf/dartle: Air Quality Model Benchmarking

Description Usage Arguments Value References Examples

target_report: prepare input for target plot

target_report(data, obs, mod, point, pollutant = c("NO2", "O3", "PM10",
  "PM2.5"), beta = 2)

target_plot(t_rep)

`data`	a data frame including observations and model forecasts
`obs`	name of the column with observed data
`mod`	name of the column with forecasts
`point`	name of the column with station ID
`pollutant`	one of `"NO2"`, `"O3"`, `"PM10"`, `"PM2.5"`
`beta`	parameter β (default is 2)
`t_rep`	output of `target_report`

target_report returns a list of 3:

quality_points

a data frame with 9 variables for each station

Point: station ID
Obs.ave: annual mean of observations (paired with forecasts)
Mod.ave: annual mean of forecasts (paired with observations)
rmsu: RMS_U (see RMS_U_obs)
crmse_norm: CRMSE/(β RMS_U)
r: R (see R)
crmse_ratio: \frac{|NMSD|}{√{2(1-R)}} as in eq.34, p.30 in Janssen et al., 2017
bias_norm: BIAS/(β RMS_U)
mqi_ts: Model Quality Indicator RMSE/(β RMS_U) (see MQI_ts_synth)
mqi_year: Model Quality Indicator for yearly averages (see MQI_year)
n_valid: no. of valid data

quality_overall

a data frame with 3 overall indicators

mqi_ts_p90: 90th percentile of mqi_ts
mqi_year_p90: 90th percentile of mqi_year
n_points: no. of valid stations

parameters

a list of 7 parameters

U_RV95r: see params_U
alpha: see params_U
RV: see params_U
Np: see params_U
Nnp: see params_U
pollutant: same as in input, one of "NO2", "O3", "PM10", "PM2.5"
beta: same as in input

target_plot returns a target plot (object of class ggplot and gg)

Janssen et al., 2017. "Guidance Document on Modelling Quality Objectives and Benchmarking. Version 2.1"

# prepare dataset
require(dplyr)
Mod <- dMean(mod.data %>% filter(Var=="c_PM10"),
             value = "Value", time = "Time", point = "Point")
Obs <- obs.pm10 %>% mutate(Day=format(Time,"%Y-%m-%d"),
                           Point=ID)
Dat <- inner_join(Mod, Obs, by=c("Point", "Day"), suffix = c(".mod", ".obs"))

# calculate indicators
t_rep <- target_report(Dat, obs = "Value.obs", mod = "Value.mod",
                       point = "Point", pollutant = "PM10")

# plot
target_plot(t_rep)