Caravan MultiMet: Weather Nowcasts and Forecasts for 22,000 Caravan Catchments
- Provider
- Google Research and Caravan community (Shalev, Kratzert et al.)
- License
- CC-BY-4.0 (source)
- Format
- Zarr (in sampled files)
- Stored at
- Google Cloud Storage, gs://caravan-multimet/v1.1 (us)
- Access
- Free and public. No account with the data provider is needed.
- DOI
- 10.5281/zenodo.14161235
- Category
- Climate & Weather
- Source
- https://arxiv.org/abs/2411.09459
About Caravan MultiMet: Weather Nowcasts and Forecasts for 22,000 Caravan Catchments
Caravan MultiMet extends Caravan, the global community dataset for large-sample hydrology (Kratzert et al., Scientific Data 2023), with catchment-averaged meteorological nowcasts and forecasts for every Caravan basin, so rainfall-runoff and streamflow forecasting models (for example LSTMs) can be trained and evaluated on real forecast inputs. Measured 2026-11-16 in the v1.1 folder: one timeseries.zarr per product, indexed by basin id (e.g. camelsaus_102101A, covering CAMELS-US, -GB, -AUS, -BR, -CL, LamaH and HYSETS basins) and date. Nowcasts: CPC precipitation 1979-2024, CHIRPS 1981-2024, IMERG v07 Early 2000-2024. Forecasts with lead_time dimension: ECMWF IFS HRES (10 days, 2016-2024, 5 variables incl. 2 m temperature, total precipitation, surface pressure, net solar and thermal radiation), GraphCast (10 days, 2016-2023) and CHIRPS-GEFS (16 days, 2000-2024). ERA5-Land (1950-2024, 15 variables) is included in UTC to match. Up to 22,492 basins per product; all data in UTC. Open with xarray.open_zarr('gs://caravan-multimet/v1.1/HRES/timeseries.zarr'). Licensed CC BY 4.0 (Zenodo record); cite Shalev et al. 2024 and Kratzert et al. 2023, and the source products.
How to load this dataset in Python
import os
import xarray as xr
ds = xr.open_zarr(os.path.join(os.environ['SCIGANTIC_MOUNT_PATH'], 'CHIRPS/timeseries.zarr'))
print(ds)Run in a Scigantic notebook on 2026-11-17 against CHIRPS/timeseries.zarr: xarray Dataset dims [basin=18655, date=15917], 1 variables e.g. ['chirps_precipitation'].
From any machine
import gcsfs
fs = gcsfs.GCSFileSystem(token='anon')
fs.ls('caravan-multimet/v1.1/', detail=False)[:10]Lists the files with anonymous access. Runs on any machine, with no cloud account.
Data structure
File types in sampled files: Zarr
Files
Example file: v1.1/CHIRPS/timeseries.zarr/.zattrs
Analyze this dataset without downloading it
On Scigantic, Caravan MultiMet: Weather Nowcasts and Forecasts for 22,000 Caravan Catchments opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.
This page is also available as Markdown and as JSON.
Related Climate & Weather datasets
- DWD ICON-EU
- CACES Mobile Aerosol Mass Spectrometer (AMS) Data, Pittsburgh 2016-2017: Mobile PM Composition and Source Factors
- AI Weather Prediction (AIWP) Model Reforecasts
- CMIP6 Climate Model Intercomparison Data
- GPM IMERG Early Precipitation L3 1 day 0.1 degree x 0.1 degree V07 (GPM_3IMERGDE) at GES DISC
- Pangeo: GFDL CM2.6 High-Resolution Climate Model
- Nonlocal Gravity Wave Parameterization Training Data from ERA5
- Greenland Ice Sheet flow - 1972 to 2026