Historical Data

The Hindcast class provides access to historical weather data (known as hindcasts) from Jua's weather models. Hindcasts are past model runs that allow you to analyze historical weather patterns, evaluate model performance, or test your pipelines.

Hindcasts differ from forecasts in that they provide access to a large archive of past weather predictions, while forecasts focus on future predictions from recent model runs. With hindcasts, you can:

• Access years of historical weather data

• Compare model predictions with observed weather

• Train and validate weather-dependent applications

• Study past weather events with high-resolution data

For more details see Historical Data.

Hindcast data is provided for the following models:

• EPT 2

• EPT 1.5 and 1.5 Early

• AIFS

⚠️ Loading hindcast data might take long. Make sure to select only the data that you need.

Multiple points & times

from jua.client import JuaClient
from jua.types.geo import LatLon
from jua.weather import Models, Variables

client = JuaClient()
model = client.weather.get_model(Models.EPT1_5)

sites = [
    # label is optional
    LatLon(55.06, 13.00, label="Kriegers Flak"),
    LatLon(54.04, 5.96, label="Gemini Wind Farm"),
    LatLon(51.71, 2.91, label="Borssele III & IV"),
    LatLon(53.89, 1.79, label="Hornsea Two"),
]

variables = [
    Variables.WIND_SPEED_AT_HEIGHT_LEVEL_100M,
]
site_hindcast = model.hindcast.get_hindcast(
    # First 7 days of May 2024
    init_time=slice("2024-05-01", "2024-05-08"),
    variables=variables,
    points=sites,
    # Select the first 24 hours
    max_lead_time=24,
    method="nearest",
)

print(site_hindcast.to_xarray())

[########################################] | 100% Completed | 15.52 s
<xarray.Dataset> Size: 14kB
Dimensions:                          (points: 4, time: 32,
                                      prediction_timedelta: 25)
Coordinates:
    latitude                         (points) float32 16B 55.07 54.02 ... 53.85
    longitude                        float32 4B -64.05
  * prediction_timedelta             (prediction_timedelta) timedelta64[ns] 200B ...
  * time                             (time) datetime64[ns] 256B 2024-05-01 .....
  * points                           (points) <U17 272B 'kriegers_flak' ... '...
Data variables:
    wind_speed_at_height_level_100m  (points, time, prediction_timedelta) float32 13kB ...

European Data

from jua.client import JuaClient
from jua.weather import Models, Variables

client = JuaClient()
model = client.weather.get_model(Models.EPT1_5)

# Define lat & lon grid for Europe
latitudes = slice(36, 71)
longitudes = slice(-15, 50)

# Define lead times
lead_times = [6, 12, 24, 72]

variables = [
    Variables.WIND_SPEED_AT_HEIGHT_LEVEL_100M,
]
site_hindcast = model.hindcast.get_hindcast(
    # First 7 days of May 2024
    init_time=slice("2024-05-01", "2024-05-08"),
    variables=variables,
    latitude=latitudes,
    longitude=longitudes,
    prediction_timedelta=lead_times,
    method="nearest",
)

print(site_hindcast.to_xarray())

[########################################] | 100% Completed | 54.43 s
<xarray.Dataset> Size: 164MB
Dimensions:                          (time: 32, prediction_timedelta: 4,
                                      latitude: 432, longitude: 741)
Coordinates:
  * latitude                         (latitude) float32 2kB 70.95 ... 36.02
  * longitude                        (longitude) float32 3kB -15.0 ... 45.0
  * prediction_timedelta             (prediction_timedelta) timedelta64[ns] 32B ...
  * time                             (time) datetime64[ns] 256B 2024-05-01 .....
Data variables:
    wind_speed_at_height_level_100m  (time, prediction_timedelta, latitude, longitude) float32 164MB ...