Skip to contents

Predict DO from a fitted model.

Source: R/metab_model_interface.R, R/metab_Kmodel.R, R/metab_model.predict_DO.R, and 2 more
predict_DO.Rd

A function in the metab_model_interface. Returns predictions of dissolved oxygen.

Usage

predict_DO(
  metab_model,
  date_start = NA,
  date_end = NA,
  ...,
  attach.units = deprecated(),
  use_saved = TRUE
)

# S3 method for metab_Kmodel
predict_DO(metab_model, date_start = NA, date_end = NA, ..., use_saved = TRUE)

# S3 method for metab_model
predict_DO(
  metab_model,
  date_start = NA,
  date_end = NA,
  ...,
  attach.units = deprecated(),
  use_saved = TRUE
)

# S3 method for metab_night
predict_DO(metab_model, date_start = NA, date_end = NA, ..., use_saved = TRUE)

# S3 method for metab_sim
predict_DO(metab_model, date_start = NA, date_end = NA, ...)

Arguments

metab_model

A metabolism model, implementing the metab_model_interface, to use in predicting metabolism

date_start

Date or a class convertible with as.Date. The first date (inclusive) for which to report DO predictions. If NA, no filtering is done.

date_end

Date or a class convertible with as.Date. The last date (inclusive) for which to report DO predictions. If NA, no filtering is done.

...

Other arguments passed to class-specific implementations of predict_DO

attach.units

(deprecated, effectively FALSE in future) logical. Should units be attached to the output?

use_saved

logical. Is it OK to use predictions that were saved with the model?

Value

A data.frame of dissolved oxygen predictions at the temporal resolution of the input data

Methods (by class)

  • metab_Kmodel: Throws an error because models of type 'Kmodel' can't predict DO. metab_Kmodel predicts K at daily timesteps and usually knows nothing about GPP or ER. So it's not possible to predict DO from this model. Try passing the output to metab_mle and THEN predicting DO.

  • metab_model: This implementation is shared by many model types

  • metab_night: Generate nighttime dissolved oxygen predictions from a nighttime regression model. metab_night only fits ER and K, and only for the darkness hours, so predictions are only generated for those hours.

  • metab_sim: Simulate values for DO.obs (with process and observation error), DO.mod (with process error only), and DO.pure (with no error). The errors are randomly generated on every new call to predict_DO.

See also

Other metab_model_interface: get_data_daily(), get_data(), get_fitting_time(), get_fit(), get_info(), get_param_names(), get_params(), get_specs(), get_version(), predict_metab()

Examples

dat <- data_metab('3', day_start=12, day_end=36)
mm <- metab_night(specs(mm_name('night')), data=dat)
preds <- predict_DO(mm, date_start=get_fit(mm)$date[3])
head(preds)
#>         date          solar.time DO.obs   DO.sat depth temp.water light
#> 1 2012-09-20 2012-09-20 17:55:58   8.07 7.769228  0.16       9.76     0
#> 2 2012-09-20 2012-09-20 18:00:58   8.06 7.782325  0.16       9.69     0
#> 3 2012-09-20 2012-09-20 18:05:58   8.00 7.797340  0.16       9.61     0
#> 4 2012-09-20 2012-09-20 18:10:58   7.94 7.810518  0.16       9.54     0
#> 5 2012-09-20 2012-09-20 18:15:58   7.91 7.825624  0.16       9.46     0
#> 6 2012-09-20 2012-09-20 18:20:58   7.87 7.840781  0.16       9.38     0
#>     DO.mod
#> 1 8.070000
#> 2 7.998760
#> 3 7.934597
#> 4 7.877057
#> 5 7.825408
#> 6 7.779299