temperature_scenario_from_records function

Make monthly temperature scenario from historic records

Produces a list of scenarios containing monthly means for Tmin and Tmax that are representative of particular years. These scenario are computed by applying linear regression to a file containing Tmin and Tmax records, and using the regression model to calculate typical values for the user-specified years.

temperature_scenario_from_records(
  weather,
  year,
  weather_start = NA,
  weather_end = NA,
  scen_type = "running_mean",
  runn_mean = 15
)

Arguments

• weather: daily weather, as produced with the fix_weather function. Can also be generated by other means, but shouold contain the columns c("Month","Day","Year","Tmin","Tmax").
• year: numeric vector of years, for which the scenario is to be produced.
• weather_start: start year of the period to be considered in calculating the regression. Defaults to NA, which means the first year of the record is used as start year.
• weather_end: end year of the period to be considered in calculating the regression. Defaults to NA, which means the last year of the record is used as end year.
• scen_type: character string, either "regression" or "running_mean", specifying how the scenario should be produced. "regression" computed the scenario based on an assumed linear trend in the data; "running_mean" uses a running mean function instead, with the length of the running mean window determined by the runn_mean parameter. The default is a running mean function, since the assumption of a linear trend often does not hold.
• runn_mean: number of vector elements to use for calculating the running mean; this is reduced, if the time series is not long enough to accommodate the specified window. Defaults to 15.

Returns

list of climate scenario objects, consisting of the following elements: 'data' = a data frame with n_intervals elements containing the absolute temperature information. 'scenario_year' = the year the scenario is representative of, i.e. the specified 'year' parameter. 'reference_year' = NA (because this is an absolute temperature scenarios, not a relative one); 'scenario_type' = 'absolute' (because this is an absolute temperature scenario, not a relative one); 'labels' = 'regression-based scenario'.

Details

This function produces outputs that can be used as input for the temperature_generation function. Sample applications are the use of the temperature_generation function for making replicate weather records for a given year for risk assessment purposes, or the generation of a weather scenario that can be compared with other datasets (e.g. climate scenarios based on the WorldClim dataset refer to a 1951-2000 baseline, so that meaningful use of such scenarios for local contexts requires consideration of a scenario that corresponds to temperatures in 1975, the central year of this period).

Examples

temperature_scenario_from_records(weather=KA_weather,year=2001,weather_start=2000,weather_end=2005)

Author(s)

Eike Luedeling