mwi

The Mosquito Weather Index (MWI) combines air temperature, relative humidity and wind speed into a single number between 0 and 1 describing how favourable the weather is for adult mosquito activity, and maps that number onto five ordered activity classes. It was developed by the Benaki Phytopathological Institute and the National Observatory of Athens under the LIFE CONOPS project (LIFE12 ENV/GR/000466) and has been published operationally on meteo.gr since June 2014.

This package implements the index, the three response functions it is built from, its activity classes, and the unit conversions needed to feed it from common weather products. It has no dependencies beyond base R.

Full documentation, including a worked introduction, is at https://e4warning.github.io/mwi/.

Installation

# install.packages("pak")
pak::pak("E4Warning/mwi")

Usage

library(mwi)

mwi(temperature = 22, humidity = 70, wind_speed = 10)
#> [1] 0.5454545

The index is a product of three responses, so any one variable falling outside its favourable range sets the whole index to zero:

mwi(temperature = 22, humidity = 30, wind_speed = 10) # too dry
#> [1] 0
mwi(temperature = 35, humidity = 70, wind_speed = 10) # too hot
#> [1] 0
mwi(temperature = 22, humidity = 70, wind_speed = 30) # too windy
#> [1] 0

The individual responses are available on their own:

mwi_temperature(c(10, 17.5, 22, 27.5, 32))
#> [1] 0.0 0.5 1.0 0.5 0.0
mwi_humidity(c(30, 40, 67.5, 95))
#> [1] 0.0 0.0 0.5 1.0
mwi_wind(c(10, 21.6, 25))
#> [1] 1 1 0

Continuous values map onto the five classes used for public communication:

x <- c(0, 0.2, 0.5, 0.9, 1)
data.frame(mwi = x, class = mwi_category(x))
#>   mwi              class
#> 1 0.0        No activity
#> 2 0.2       Low activity
#> 3 0.5  Moderate activity
#> 4 0.9      High activity
#> 5 1.0 Very high activity

Computing the index over time

The index must be computed before it is averaged, not after. Because mwi() is non-linear, evaluating it on daily summaries of each weather variable is not the same as evaluating it hourly and then summarising the result — and it is much less useful. Daily maxima of temperature and of humidity typically occur many hours apart, so an index built from them describes an hour that never happened.

mwi_summarise() does it in the right order:

hours <- 0:23
temp <- 18 + 8 * sin((hours - 6) / 24 * 2 * pi)
rh <- 75 - 15 * sin((hours - 6) / 24 * 2 * pi)
wind <- rep(5, 24)

# Right: index each hour, then take the daily maximum
mwi_summarise(temp, rh, wind, by = rep("2019-07-15", 24))
#>           by       mwi  n
#> 1 2019-07-15 0.5657766 24

# Wrong: summarise the weather first, then index
mwi(max(temp), max(rh), max(wind))
#> [1] 0.7272727

Working from reanalysis output

Products such as ERA5-Land distribute temperatures in Kelvin, humidity as a dewpoint, and wind as orthogonal components. The conversions are included:

temperature <- celsius_from_kelvin(295.15)
dewpoint <- celsius_from_kelvin(288.15)
humidity <- relative_humidity_from_dewpoint(dewpoint, temperature)
wind <- wind_speed_from_components(u = 1.2, v = -0.8)

mwi(temperature, humidity, wind, units = "m/s")
#> [1] 0.4455658

Citation

If you use this package, please cite it (archived at doi:10.5281/zenodo.22101143) together with the evaluation study. See citation("mwi").

License

GPL (>= 3). The index itself was developed by the Benaki Phytopathological Institute and the National Observatory of Athens under the LIFE CONOPS project (LIFE12 ENV/GR/000466).