Events (`s2stools.events`)

Finding events in the forecasts and creating composites.

ssw_composite = EventComposite(data, "path/to/eventlists*.json", descr="Sudden Warmings", model="ecmwf")

Class

class s2stools.events.EventComposite(data, events, descr, model)

__init__(data, events, descr, model)

Composite class

Generate an instance my_composite of the class EventComposite and analyse the composite data using instance my_composite.comp.

Parameters:

data (xr.Dataset) – dataset with appropriate dimensions (reftime, leadtime, hc_year, number, …)
events ((str or list)) – if str, then define the path, e.g., json/s2s_events/MODEL/ssw/* (model will automatically replaced by self.model) if list then of format [{fc: {reftime: None, hc_year: None, number: None}, days_since_init: None}, {…}]
descr (str) – event description, e.g., used for plot titles
model (str) – model name, e.g., ecmwf or ukmo

Examples

>>> ssw_composite = EventComposite(data, "path/to/eventlists*.json", descr="sudden warmings", model="ecmwf")
>>> ssw_composite
<s2stools.events.EventComposite> of 182 events of type 'sudden warmings'.
    --> see composite dataset using my_event_composite.comp
>>> ssw_composite.comp
<xarray.Dataset>
Dimensions:    (lagtime: 73, longitude: 2, latitude: 1, i: 182)
Coordinates:
  * lagtime    (lagtime) timedelta64[ns] -36 days -35 days ... 35 days 36 days
  * longitude  (longitude) float32 -180.0 -177.5
  * latitude   (latitude) float32 60.0
    number     (i) int64 0 3 4 5 6 8 10 3 8 10 ... 39 40 42 43 44 45 46 47 48 49
    reftime    (i) datetime64[ns] 2017-11-16 2017-11-16 ... 2017-11-20
    hc_year    (i) int64 -20 -20 -20 -20 -20 -20 -20 -19 -19 ... 0 0 0 0 0 0 0 0
    validtime  (i, lagtime) datetime64[ns] NaT NaT NaT NaT ... NaT NaT NaT NaT
    leadtime   (i) timedelta64[ns] 11 days 11 days 10 days ... 19 days 16 days
  * i          (i) int64 0 1 2 3 4 5 6 7 8 ... 174 175 176 177 178 179 180 181
Data variables:
    u          (i, latitude, longitude, lagtime) float32 dask.array<chunksize=(1, 1, 2, 73), meta=np.ndarray>

comp = None: Access the composite dataset using my_event_composite.comp

data = None: Obtain the original feded data using my_event_composite.data

descr = '': Access the composite description using my_event_composite.descr

Functions

s2stools.events.bootstrap_extr_prob(population, sample_size, dsi_slice, threshold, alpha=0.05, n_bootstrap_samples=1000)

Bootstrapping forecasts to compute mean and confidence interval for probability that NAM is below threshold.

Parameters:

population (DataArray) – dimensions ‘fc’ and ‘days_since_init’
sample_size (int) – size of one sample that is bootstrapped
dsi_slice (int) – at which ‘days_since_init’ to start to bootstrap
threshold (int) – NAM below which threshold?
alpha (float, optional) – alpha-level for confidence interval. Defaults to 0.05.
n_bootstrap_samples (int, optional) – times to repeat drawing samples from the population. Defaults to 1_000.

Returns:

(mean, (CI_bottom, CI_top))

Return type:

(float, (float, float))

s2stools.events.extreme_predictors(data_comp)

Compute the fraction of events with nam1000 below 0, -3 and above +3.

Parameters:: data_comp (xr.Dataset) – Dataset that contains variable nam1000 and dimensions “days_since_event” and “i”.
Returns:: Dataset with dimension “days_since_event”
Return type:: xr.Dataset

s2stools.events.find_ssw(u60_10hPa, buffer_start=10, buffer_end=10, require_westwind_start=10) → list

Find Sudden Stratospheric Warmings in S2S forecast.

Parameters:

u60_10hPa (xr.DataArray) – zonal wind at 60N 10hPa, requires dimensions reftime, hc_year, number
buffer_start (int) – When are events allowed to happen the earliest?
buffer_end (int) – When are events allowed to happen the latest?
require_westwind_start (int) – How many days should u60 start positive?

Return type:

list

Events (s2stools.events)

Class

Functions

Events (`s2stools.events`)