⬡

climdata

A unified interface for climate data extraction, bias adjustment & downscaling, extreme-index computation, and visualisation.

Python 3.11+ xarray xclim conda-forge MIT License

✨ Key Features

Provider-agnostic extraction — MSWX, CMIP, POWER, DWD, HYRAS, ERA5, W5E5 and more
Point, region, and shapefile extraction modes with curvilinear grid support
Unit normalisation via xclim
Extreme climate indices — Tx10p, Tn90p, CDD, WSDI, …
xarray.Dataset → long-form pandas.DataFrame with one call
Bias Adjustment & Statistical Downscaling (BCSD, QDM, DQM, QM)
Fully reproducible runs via Hydra configuration

⚡ Quick Install

conda create -n climdata python=3.11 -y
conda activate climdata
pip install climdata

Or from source: git clone <repo> && pip install -e .

🚀 Quick Example

from climdata import ClimData

extractor = ClimData(overrides=[
    "dataset=mswx",
    "lat=52.5",
    "lon=13.4",
    "time_range.start_date=2014-01-01",
    "time_range.end_date=2014-12-31",
    "variables=[tasmin,tasmax,pr]",
    "data_dir=/path/to/data",
])

ds  = extractor.extract()           # returns xarray.Dataset
df  = extractor.to_dataframe(ds)    # long-form DataFrame
extractor.to_csv(df)

🔁 Workflow Runner

result = extractor.run_workflow(
    actions=["extract", "calc_index", "to_dataframe", "to_csv"])

WorkflowResult contains produced datasets, dataframes, and output filenames.

🗂 Getting Started in this App

Download — Select a point or bounding box on the map, configure a dataset and date range, then click Run. The raster overlay is rendered automatically after a successful pipeline run.

BASD — Apply bias adjustment to model output using an observational reference. Choose a method (BCSD, QDM, DQM, or QM) and a target variable, then run.

Comparison — Compare two climate datasets side-by-side across variables, time periods, or sources to visualise differences.

🔧 Bias Adjustment & Statistical Downscaling

Correct systematic biases in climate model output using observational reference data.

Reference (Observations)

Dataset

Start

End

Target (Model Output)

Dataset

Start

End

Method & Variable

Bias correction method

Variable

Output

Output directory

default

Format

ℹ️

About BASD

Bias Adjustment & Statistical Downscaling corrects systematic errors in climate model projections by comparing them to a historical observational reference.

Output is bias-corrected data at the spatial resolution of the reference dataset.