test_thresholdsensitivity: Test threshold sensitivity
In NaiaraLopezRojo/globalIRmap: What the Package Does (One Line, Title Case)

Description Usage Arguments Value

Test sensitivity of model predictions to flow intermittence probability threshold used to classify global river reaches into non-perennial and perennial classes.

test_thresholdsensitivity(
  in_gpredsdt,
  in_rivpred,
  threshrange_gauges = seq(0.3, 0.7, 0.01),
  threshrange_network = seq(0.45, 0.55, 0.01),
  mincutoff = 0.1,
  gaugescol = "IRpredprob_CVnosp",
  netcol = "predbasic800"
)

`in_gpredsdt`	data.table of model predictions for gauging stations. Here, output from bind_gaugepreds.
`in_rivpred`	data.table of model predictions for global river network. Here, output from netpredformat.
`threshrange_gauges`	numerical vector of probability threshold values for which to assess and plot model predictions and performance for gauging stations.
`threshrange_network`	numerical vector of probability threshold values for which to produce predictions of flow intermittence for the global network (set narrower than `thresrange_gauges` as each one takes much longer to compute).
`mincutoff`	minimum long-term mean annual flow (MAF) to include in assessment of gauges and network (i.e., mincutoff == 0.1 means that only gauging stations and river reaches with a WaterGAP estimated MAF >= 0.1 are included in the sensitivity analysis)
`gaugescol`	name of the column in `in_gpredsdt` containing the predicted probability of flow intermittence.
`netcol`	name of the column in `in_rivpred` containing the predicted probability of flow intermittence.

list with three elements:

A ggplot of the probability threshold ranges for gauges >= 10m3/s and gauges < 10 m3/s that maximize balanced accuracy, bias, raw accuracy, and |sensitivity-specificity|.
A ggplot of the predicted global prevalence of IRES as a function of probability thresholds for two gauge size classes (<10 and >=10 m3/s)
data.table containing the predicted global prevalence of IRES based on the range of probability threshold threshrange_network.