### Clone WeatherBench 2 Repository

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/index.md

Clone the GitHub repository to get started with WeatherBench 2. This is the initial step for setting up the project locally.

```bash
git clone git@github.com:google-research/weatherbench2.git
```

--------------------------------

### Slice dataset execution example

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/command-line-scripts.md

Example command to run slice_dataset.py with specific selection and indexing criteria.

```bash
python slice_dataset.py -- \
  --input_path=gs://weatherbench2/datasets/ens/2018-64x32_equiangular_with_poles_conservative.zarr \
  --output_path=PATH \
  --sel="prediction_timedelta_stop=15 days,latitude_start=-33.33,latitude_stop=33.33" \
  --isel="longitude_start=0,longitude_stop=180,longitude_step=40" \
  --keep_variables=geopotential,temperature
```

--------------------------------

### Install WeatherBench 2 Package

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/index.md

Install the WeatherBench 2 package using pip after cloning the repository. This command installs the package in the current environment.

```bash
cd weatherbench2
pip install .
```

--------------------------------

### Compute averages execution example

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/command-line-scripts.md

Example command to run compute_averages.py using DataflowRunner.

```bash
python compute_averages.py \
    --input_path=gs://weatherbench2/datasets/era5/1959-2022-6h-64x32_equiangular_with_poles_conservative.zarr \
    --output_path=gs://$BUCKET/datasets/era5/$USER/temperature-vertical-profile.zarr \
    --runner=DataflowRunner \
    -- \
    --project=$PROJECT \
    --averaging_dims=time,longitude \
    --variables=temperature \
    --temp_location=gs://$BUCKET/tmp/ \
    --setup_file=./setup.py \
    --requirements_file=./scripts/dataflow-requirements.txt \
    --job_name=compute-vertical-profile-$USER
```

--------------------------------

### Install WeatherBench2 in editable mode

Source: https://github.com/google-research/weatherbench2/blob/main/docs/contributing.md

Installs the project and its test dependencies in editable mode for local development.

```shell
pip install -e ".[tests]"
```

--------------------------------

### Display Help for Evaluate Script

Source: https://github.com/google-research/weatherbench2/blob/main/scripts/README.md

Use the `--help` flag to see how to invoke the `evaluate.py` pipeline with examples. This is applicable to all core pipelines.

```shell
python3 evaluate.py --help
```

--------------------------------

### Example: Compute Derived Variables

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/command-line-scripts.md

An example command to run the derived variables computation script. It specifies input and output paths, rechunking parameters, and options for renaming total precipitation variables.

```bash
python compute_derived_variables.py \
  --input_path=gs://weatherbench2/datasets/hres/2016-2022-12h-6h-0p25deg-chunk-1.zarr/ \
  --output_path=PATH \
  --working_chunks=prediction_timedelta=-1 \
  --rename_raw_tp_name=True \
  --raw_tp_name=total_precipitation_6hr \
  --rechunk_itemsize=1 \

```

--------------------------------

### Install WeatherBench 2

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/evaluation.ipynb

Install the WeatherBench 2 package from GitHub. Pip might show a warning about the Pandas version, but the notebook should still function correctly.

```python
# Pip might complain about the Pandas version. The notebook should still work as expected.
!pip install git+https://github.com/google-research/weatherbench2.git
```

--------------------------------

### Install WeatherBench 2 for Development

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/index.md

Install the WeatherBench 2 package in editable mode for active development. This allows changes to the code to be reflected immediately without reinstallation.

```bash
pip install -e .
```

--------------------------------

### Example: Compute Zonal Energy Spectra

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/command-line-scripts.md

An example command to compute zonal energy spectra. It sets the input and output paths, specifies a time range, and lists the base variables for which to compute spectra.

```bash
python compute_zonal_energy_spectrum.py \
  --input_path=gs://weatherbench2/datasets/era5/1959-2022-6h-240x121_equiangular_with_poles_conservative.zarr  \
  --output_path=PATH \
  --time_start=2020 \
  --time_stop=2020 \
  --base_variables=geopotential,specific_humidity,temperature,u_component_of_wind,v_component_of_wind,wind_speed,10m_u_component_of_wind,10m_v_component_of_wind,10m_wind_speed,2m_temperature,mean_sea_level_pressure,total_precipitation_6hr,total_precipitation_24hr

```

--------------------------------

### Install GCP Requirements for WeatherBench 2

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/index.md

Install the additional requirements for WeatherBench 2 if you plan to launch jobs on Google Cloud Platform (GCP). This includes dependencies for GCP services like DataFlow.

```bash
pip install .[gcp]
```

--------------------------------

### Dataflow Execution with Beam DataflowRunner

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/beam-in-the-cloud.md

Run on Google Cloud Dataflow using the DataflowRunner. Ensure GCP requirements are installed and specify project, region, and temporary storage.

```bash
python scripts/evaluate.py \
  --forecast_path=gs://weatherbench2/datasets/hres/2016-2022-0012-64x32_equiangular_with_poles_conservative.zarr \
  --obs_path=gs://weatherbench2/datasets/era5/1959-2022-6h-64x32_equiangular_with_poles_conservative.zarr \
  --climatology_path=gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_64x32_equiangular_with_poles_conservative.zarr \
  --output_dir=gs://$BUCKET \
  --output_file_prefix=hres_vs_era_2020_ \
  --by_init=True \
  --input_chunks=init_time=10 \
  --eval_configs=deterministic \
  --time_start=2020-01-01 \
  --time_stop=2020-12-31 \
  --variables=geopotential \
  --use_beam=True \
  --runner=DataflowRunner \
  -- \
  --project $PROJECT \
  --region $REGION \
  --temp_location gs://$BUCKET/tmp/ \
  --setup_file=./setup.py
```

--------------------------------

### Evaluate SphericalCNN Forecasts

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/official-evaluation.md

Evaluates SphericalCNN model forecasts against ERA5 reanalysis for 2020. This example specifies the forecast path for SphericalCNN data and uses standard evaluation configurations.

```bash
python evaluate.py -- \
 --forecast_path=gs://weatherbench2/datasets/sphericalcnn/2020-240x121_equiangular_with_poles_conservative.zarr \
 --obs_path=gs://weatherbench2/datasets/era5/1959-2022-6h-240x121_equiangular_with_poles_conservative.zarr \
 --climatology_path=gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_240x121_equiangular_with_poles_conservative.zarr \
 --output_dir=$OUTDIR/240x121/deterministic/ \
 --output_file_prefix=sphericalcnn_vs_era_2020_ \
 --input_chunks=init_time=1,lead_time=10 \
 --fanout=27 \
 --regions=all \
 --eval_configs=deterministic,deterministic_temporal \
 --evaluate_climatology=False \
 --evaluate_persistence=False \
 --time_start=2020-01-01 \
 --time_stop=2020-12-31 \
 --variables=geopotential,temperature,u_component_of_wind,v_component_of_wind,specific_humidity \
 --derived_variables=wind_speed \
 --add_land_region=False \
 --use_beam=True
```

--------------------------------

### Open FuXi Zarr Dataset

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/data-guide.ipynb

Opens a FuXi model dataset from a Google Cloud Storage path using xarray. This dataset contains forecasts for 2020. Ensure you have the necessary authentication and xarray installed.

```python
xr.open_zarr('gs://weatherbench2/datasets/fuxi/2020-1440x721.zarr')
```

--------------------------------

### Read METAR Parquet Data with Pandas

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/data-guide.ipynb

Demonstrates how to open and inspect a single Parquet file containing METAR data using the pandas library. Ensure pandas is installed and you have access to the specified GCS path.

```python
import pandas as pd

# Opening a single partition
pd.read_parquet(
    'gs://weatherbench2/datasets/metar/metar-timeNominal-by-month/year=2020/month=1/2020-01.parquet'
).head()
```

--------------------------------

### Evaluate Pangu-HRES-Init vs Analysis

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/official-evaluation.md

Evaluates the Pangu-HRES-Init model against analysis data for the year 2020. This configuration disables persistence evaluation and uses Apache Beam.

```bash
python evaluate.py -- \
  --forecast_path=gs://weatherbench2/datasets/pangu_hres_init/2020_0012_64x32_equiangular_conservative.zarr \
  --obs_path=gs://weatherbench2/datasets/hres_t0/2016-2022-6h-64x32_equiangular_conservative.zarr \
  --climatology_path=gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_64x32_equiangular_conservative.zarr \
  --output_dir=$OUTDIR/64x32/deterministic/ \
  --output_file_prefix=pangu_hres_init_vs_analysis_2020_ \
  --input_chunks=init_time=1 \
  --fanout=27 \
  --regions=all \
  --eval_configs=deterministic,deterministic_temporal \
  --evaluate_climatology=False \
  --evaluate_persistence=False \
  --time_start=2020-01-01 \
  --time_stop=2020-12-31 \
  --variables=geopotential,temperature,u_component_of_wind,v_component_of_wind,specific_humidity,2m_temperature,10m_u_component_of_wind,10m_v_component_of_wind,mean_sea_level_pressure \
  --add_land_region=False \
  --use_beam=True
```

--------------------------------

### Evaluate Pangu-HRES Init vs Analysis

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/official-evaluation.md

Evaluates the Pangu-HRES init model against analysis data for 2020. This configuration uses a land surface model dataset and disables climatology/persistence evaluations.

```bash
python evaluate.py -- \
 --forecast_path=gs://weatherbench2/datasets/pangu_hres_init/2020_0012_240x121_equiangular_with_poles_conservative.zarr \
 --obs_path=gs://weatherbench2/datasets/hres_t0/2016-2022-6h-240x121_equiangular_with_poles_conservative.zarr \
 --lsm_dataset=gs://weatherbench2/datasets/era5/1959-2022-6h-240x121_equiangular_with_poles_conservative.zarr \
 --climatology_path=gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_240x121_equiangular_with_poles_conservative.zarr \
 --output_dir=$OUTDIR/240x121/deterministic/ \
 --output_file_prefix=pangu_hres_init_vs_analysis_2020_ \
 --input_chunks=init_time=1,lead_time=10 \
 --fanout=27 \
 --regions=all \
 --eval_configs=deterministic,deterministic_temporal \
 --evaluate_climatology=False \
 --evaluate_persistence=False \
 --time_start=2020-01-01 \
 --time_stop=2020-12-31 \
 --variables=geopotential,temperature,u_component_of_wind,v_component_of_wind,specific_humidity,2m_temperature,10m_u_component_of_wind,10m_v_component_of_wind,mean_sea_level_pressure \
 --add_land_region=False \
 --use_beam=True
```

--------------------------------

### Evaluate Pangu-HRES-Init Model vs Analysis Data

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/official-evaluation.md

Evaluates the Pangu-HRES-Init model against analysis data (HRES T0) for the year 2020. This command uses a different observation dataset path.

```bash
python evaluate.py -- \
  --forecast_path=gs://weatherbench2/datasets/pangu_hres_init/2020_0012_0p25.zarr \
  --obs_path=gs://weatherbench2/datasets/hres_t0/2016-2022-6h-1440x721.zarr \
  --lsm_dataset=gs://weatherbench2/datasets/era5/1959-2022-6h-1440x721.zarr \
  --climatology_path=gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_1440x721.zarr \
  --output_dir=$OUTDIR/1440x721/deterministic/ \
  --output_file_prefix=pangu_hres_init_vs_analysis_2020_ \
  --input_chunks=init_time=1,lead_time=1 \
  --fanout=27 \
  --regions=all \
  --eval_configs=deterministic,deterministic_temporal \
  --evaluate_climatology=False \
  --evaluate_persistence=False \
  --time_start=2020-01-01 \
  --time_stop=2020-12-31 \
  --variables=geopotential,temperature,u_component_of_wind,v_component_of_wind,specific_humidity,2m_temperature,10m_u_component_of_wind,10m_v_component_of_wind,mean_sea_level_pressure \
  --add_land_region=False \
  --use_beam=True
```

--------------------------------

### Open Spherical CNN Zarr Dataset

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/data-guide.ipynb

Opens a Spherical CNN dataset from a Google Cloud Storage path using xarray. Ensure you have the necessary authentication and xarray installed.

```python
xr.open_zarr(
    'gs://weatherbench2/datasets/sphericalcnn/2020-240x121_equiangular_with_poles.zarr'
)
```

--------------------------------

### Evaluate Pangu-HRES-Init vs ERA5

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/official-evaluation.md

Evaluates the Pangu-HRES-Init model against ERA5 observations for the year 2020. This configuration disables persistence evaluation and uses Apache Beam.

```bash
python evaluate.py -- \
  --forecast_path=gs://weatherbench2/datasets/pangu/pangu_hres_init/2020_0012_64x32_equiangular_conservative.zarr \
  --obs_path=gs://weatherbench2/datasets/era5/1959-2022-6h-64x32_equiangular_conservative.zarr \
  --climatology_path=gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_64x32_equiangular_conservative.zarr \
  --output_dir=$OUTDIR/64x32/deterministic/ \
  --output_file_prefix=pangu_hres_init_vs_era_2020_ \
  --input_chunks=init_time=1 \
  --fanout=27 \
  --regions=all \
  --eval_configs=deterministic,deterministic_temporal \
  --evaluate_climatology=False \
  --evaluate_persistence=False \
  --time_start=2020-01-01 \
  --time_stop=2020-12-31 \
  --variables=geopotential,temperature,u_component_of_wind,v_component_of_wind,specific_humidity,2m_temperature,10m_u_component_of_wind,10m_v_component_of_wind,mean_sea_level_pressure \
  --add_land_region=False \
  --use_beam=True
```

--------------------------------

### Import Libraries for WeatherBench 2

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/evaluation.ipynb

Import necessary libraries for WeatherBench 2, including apache_beam and xarray. Apache Beam needs to be imported separately to prevent a TypingError.

```python
import apache_beam   # Needs to be imported separately to avoid TypingError
import weatherbench2
import xarray as xr
```

--------------------------------

### Open GraphCast Operational Zarr Dataset

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/data-guide.ipynb

Use this snippet to open a GraphCast operational forecast dataset stored in Zarr format from a Google Cloud Storage path. Ensure you have the xarray library installed.

```python
xr.open_zarr(
    'gs://weatherbench2/datasets/graphcast_hres_init/2020/date_range_2019-11-16_2021-02-01_12_hours_derived.zarr'
)
```

--------------------------------

### Evaluate Pangu-HRES-Init Model vs ERA5

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/official-evaluation.md

Evaluates the Pangu-HRES-Init model against ERA5 data for the year 2020. This configuration also disables persistence evaluation.

```bash
python evaluate.py -- \
  --forecast_path=gs://weatherbench2/datasets/pangu_hres_init/2020_0012_0p25.zarr \
  --obs_path=gs://weatherbench2/datasets/era5/1959-2022-6h-1440x721.zarr \
  --climatology_path=gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_1440x721.zarr \
  --output_dir=$OUTDIR/1440x721/deterministic/ \
  --output_file_prefix=pangu_hres_init_vs_era_2020_ \
  --input_chunks=init_time=1,lead_time=1 \
  --fanout=27 \
  --regions=all \
  --eval_configs=deterministic,deterministic_temporal \
  --evaluate_climatology=False \
  --evaluate_persistence=False \
  --time_start=2020-01-01 \
  --time_stop=2020-12-31 \
  --variables=geopotential,temperature,u_component_of_wind,v_component_of_wind,specific_humidity,2m_temperature,10m_u_component_of_wind,10m_v_component_of_wind,mean_sea_level_pressure \
  --add_land_region=False \
  --use_beam=True
```

--------------------------------

### Define input dataset paths

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/evaluation.ipynb

Set the Google Cloud Storage paths for forecast, observation, and climatology Zarr datasets.

```python
forecast_path = 'gs://weatherbench2/datasets/hres/2016-2022-0012-64x32_equiangular_conservative.zarr'
obs_path = 'gs://weatherbench2/datasets/era5/1959-2022-6h-64x32_equiangular_conservative.zarr'
climatology_path = 'gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_64x32_equiangular_conservative.zarr'
```

--------------------------------

### Open IFS HRES t=0 Zarr Archive

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/data-guide.ipynb

Opens a Zarr archive containing IFS HRES t=0 analysis data. This is used as ground truth for forecast evaluation. Ensure you have the xarray library installed.

```python
xr.open_zarr('gs://weatherbench2/datasets/hres_t0/2016-2022-6h-1440x721.zarr')
```

--------------------------------

### Run Probabilistic Evaluation at 1440x721 Resolution

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/official-evaluation.md

Executes probabilistic evaluation scripts for weather datasets. Requires setting the OUTDIR environment variable before execution.

```bash
python evaluate.py -- \
  --forecast_path=gs://weatherbench2/datasets/ens/2018-2022-1440x721.zarr \
  --obs_path=gs://weatherbench2/datasets/hres_t0/2016-2022-6h-1440x721.zarr \
  --lsm_dataset=gs://weatherbench2/datasets/era5/1959-2022-6h-1440x721.zarr \
  --climatology_path=gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_1440x721.zarr \
  --output_dir=$OUTDIR/1440x721/probabilistic/ \
  --output_file_prefix=ens_vs_analysis_2020_ \
  --input_chunks=init_time=1,lead_time=1 \
  --fanout=27 \
  --regions=all \
  --eval_configs=probabilistic \
  --evaluate_climatology=False \
  --evaluate_persistence=False \
  --time_start=2020-01-01 \
  --time_stop=2020-12-31 \
  --variables=geopotential,temperature,u_component_of_wind,v_component_of_wind,specific_humidity,2m_temperature,10m_u_component_of_wind,10m_v_component_of_wind,mean_sea_level_pressure,10m_wind_speed,wind_speed \
  --use_beam=True
```

```bash
python evaluate.py -- \
  --forecast_path=gs://weatherbench2/datasets/ens/2018-2022-1440x721.zarr \
  --obs_path=gs://weatherbench2/datasets/era5/1959-2022-6h-1440x721.zarr \
  --climatology_path=gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_1440x721.zarr \
  --output_dir=$OUTDIR/1440x721/probabilistic/ \
  --output_file_prefix=ens_vs_era_2020_ \
  --input_chunks=init_time=1,lead_time=1 \
  --fanout=27 \
  --regions=all \
  --eval_configs=probabilistic \
  --evaluate_climatology=False \
  --evaluate_persistence=False \
  --time_start=2020-01-01 \
  --time_stop=2020-12-31 \
  --variables=geopotential,temperature,u_component_of_wind,v_component_of_wind,specific_humidity,2m_temperature,10m_u_component_of_wind,10m_v_component_of_wind,mean_sea_level_pressure,total_precipitation_6hr,total_precipitation_24hr,10m_wind_speed,wind_speed \
  --use_beam=True
```

```bash
python evaluate.py -- \
  --forecast_path=gs://weatherbench2/datasets/ens/2018-2022-1440x721.zarr \
  --obs_path=gs://weatherbench2/datasets/era5/1959-2022-6h-1440x721.zarr \
  --climatology_path=gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_1440x721.zarr \
  --output_dir=$OUTDIR/1440x721/probabilistic/ \
  --output_file_prefix=climatology_vs_era_2020_ \
  --input_chunks=init_time=1,lead_time=1 \
  --fanout=27 \
  --regions=all \
  --eval_configs=probabilistic \
  --evaluate_probabilistic_climatology=True \
  --probabilistic_climatology_start_year=1990 \
  --probabilistic_climatology_end_year=2019 \
  --time_start=2020-01-01 \
  --time_stop=2020-12-31 \
  --variables=geopotential,temperature,u_component_of_wind,v_component_of_wind,specific_humidity,2m_temperature,10m_u_component_of_wind,10m_v_component_of_wind,mean_sea_level_pressure,total_precipitation_6hr,total_precipitation_24hr,10m_wind_speed,wind_speed \
  --use_beam=True
```

--------------------------------

### Evaluate Pangu-HRES Init vs ERA5

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/official-evaluation.md

Evaluates the Pangu-HRES init model against ERA5 data for 2020. This script disables climatology and persistence evaluations and land region calculations.

```bash
python evaluate.py -- \
 --forecast_path=gs://weatherbench2/datasets/pangu_hres_init/2020_0012_240x121_equiangular_with_poles_conservative.zarr \
 --obs_path=gs://weatherbench2/datasets/era5/1959-2022-6h-240x121_equiangular_with_poles_conservative.zarr \
 --climatology_path=gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_240x121_equiangular_with_poles_conservative.zarr \
 --output_dir=$OUTDIR/240x121/deterministic/ \
 --output_file_prefix=pangu_hres_init_vs_era_2020_ \
 --input_chunks=init_time=1,lead_time=10 \
 --fanout=27 \
 --regions=all \
 --eval_configs=deterministic,deterministic_temporal \
 --evaluate_climatology=False \
 --evaluate_persistence=False \
 --time_start=2020-01-01 \
 --time_stop=2020-12-31 \
 --variables=geopotential,temperature,u_component_of_wind,v_component_of_wind,specific_humidity,2m_temperature,10m_u_component_of_wind,10m_v_component_of_wind,mean_sea_level_pressure \
 --add_land_region=False \
 --use_beam=True
```

--------------------------------

### Compute Zonal Energy Spectrum for Keisler Dataset

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/official-evaluation.md

Use this script to compute the zonal energy spectrum for the Keisler dataset. Specify the input Zarr path, output directory, and time range. This example uses a different spatial resolution (360x181).

```bash
python compute_zonal_energy_spectrum.py -- \
 --input_path=gs://weatherbench2/datasets/keisler/2020-360x181.zarr  \
 --output_path=$OUTDIR/highest_res/spectra/keisler_2020.zarr \
 --time_start=2020 \
 --time_stop=2020 \
 --base_variables=geopotential,specific_humidity,temperature,u_component_of_wind,v_component_of_wind,wind_speed \

```

--------------------------------

### List Dataflow User Metrics

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/beam-in-the-cloud.md

Use the beta gcloud CLI commands to list user metrics for a specific Dataflow job.

```shell
JOBID=<enter job id here>
gcloud beta dataflow metrics list $JOBID --source=user
```

--------------------------------

### Run Probabilistic Evaluation Commands

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/official-evaluation.md

Execute evaluation scripts using the evaluate.py entry point with specific dataset paths, time ranges, and evaluation configurations.

```bash
python evaluate.py -- \
  --forecast_path=gs://weatherbench2/datasets/ens/2018-2022-64x32_equiangular_conservative.zarr \
  --obs_path=gs://weatherbench2/datasets/hres_t0/2016-2022-6h-64x32_equiangular_conservative.zarr \
  --lsm_dataset=gs://weatherbench2/datasets/era5/1959-2022-6h-64x32_equiangular_conservative.zarr \
  --climatology_path=gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_64x32_equiangular_conservative.zarr \
  --output_dir=$OUTDIR/64x32/probabilistic/ \
  --output_file_prefix=ens_vs_analysis_2020_ \
  --input_chunks=init_time=1,lead_time=1 \
  --fanout=27 \
  --regions=all \
  --eval_configs=probabilistic \
  --evaluate_climatology=False \
  --evaluate_persistence=False \
  --time_start=2020-01-01 \
  --time_stop=2020-12-31 \
  --variables=geopotential,temperature,u_component_of_wind,v_component_of_wind,specific_humidity,2m_temperature,10m_u_component_of_wind,10m_v_component_of_wind,mean_sea_level_pressure,10m_wind_speed,wind_speed \
  --use_beam=True
```

```bash
python evaluate.py -- \
  --forecast_path=gs://weatherbench2/datasets/ens/2018-2022-64x32_equiangular_conservative.zarr \
  --obs_path=gs://weatherbench2/datasets/era5/1959-2022-6h-64x32_equiangular_conservative.zarr \
  --climatology_path=gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_64x32_equiangular_conservative.zarr \
  --output_dir=$OUTDIR/64x32/probabilistic/ \
  --output_file_prefix=ens_vs_era_2020_ \
  --input_chunks=init_time=1,lead_time=1 \
  --fanout=27 \
  --regions=all \
  --eval_configs=probabilistic \
  --evaluate_climatology=False \
  --evaluate_persistence=False \
  --time_start=2020-01-01 \
  --time_stop=2020-12-31 \
  --variables=geopotential,temperature,u_component_of_wind,v_component_of_wind,specific_humidity,2m_temperature,10m_u_component_of_wind,10m_v_component_of_wind,mean_sea_level_pressure,total_precipitation_6hr,total_precipitation_24hr,10m_wind_speed,wind_speed \
  --use_beam=True
```

```bash
python evaluate.py -- \
  --forecast_path=gs://weatherbench2/datasets/ens/2018-2022-64x32_equiangular_conservative.zarr \
  --obs_path=gs://weatherbench2/datasets/era5/1959-2022-6h-64x32_equiangular_conservative.zarr \
  --climatology_path=gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_64x32_equiangular_conservative.zarr \
  --output_dir=$OUTDIR/64x32/probabilistic/ \
  --output_file_prefix=climatology_vs_era_2020_ \
  --input_chunks=init_time=1,lead_time=1 \
  --fanout=27 \
  --regions=all \
  --eval_configs=probabilistic \
  --evaluate_probabilistic_climatology=True \
  --probabilistic_climatology_start_year=1990 \
  --probabilistic_climatology_end_year=2019 \
  --time_start=2020-01-01 \
  --time_stop=2020-12-31 \
  --variables=geopotential,temperature,u_component_of_wind,v_component_of_wind,specific_humidity,2m_temperature,10m_u_component_of_wind,10m_v_component_of_wind,mean_sea_level_pressure,total_precipitation_6hr,total_precipitation_24hr,10m_wind_speed,wind_speed \
  --use_beam=True
```

--------------------------------

### Evaluate Forecasts with WeatherBench2

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/command-line-scripts.md

Use this script to evaluate forecast data against observations. Specify input dataset paths, output directory, and evaluation configurations.

```bash
python evaluate.py \
  --forecast_path=gs://weatherbench2/datasets/hres/2016-2022-0012-64x32_equiangular_with_poles_conservative.zarr \
  --obs_path=gs://weatherbench2/datasets/era5/1959-2022-6h-64x32_equiangular_with_poles_conservative.zarr \
  --climatology_path=gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_64x32_equiangular_with_poles_conservative.zarr \
  --output_dir=./ \
  --output_file_prefix=hres_vs_era_2020_ \
  --input_chunks=init_time=1 \
  --eval_configs=deterministic,deterministic_spatial,deterministic_temporal \
  --time_start=2020-01-01 \
  --time_stop=2020-12-31 \
  --variables=geopotential,temperature,u_component_of_wind,v_component_of_wind,specific_humidity,2m_temperature,10m_u_component_of_wind,10m_v_component_of_wind,mean_sea_level_pressure,total_precipitation_6hr,total_precipitation_24hr,10m_wind_speed,wind_speed
```

--------------------------------

### Import Evaluation Functions

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/evaluation.ipynb

Import necessary functions for running evaluations, either in-memory or via Apache Beam.

```python
from weatherbench2.evaluation import evaluate_in_memory, evaluate_with_beam
```

--------------------------------

### Evaluate Weather Forecasts with WeatherBench2

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/official-evaluation.md

These commands execute the evaluation pipeline for different forecast models against ERA5 or analysis data. Ensure the OUTDIR environment variable is set before running.

```bash
python evaluate.py -- \
  --forecast_path=gs://weatherbench2/datasets/hres/2016-2022-0012-1440x721.zarr \
  --obs_path=gs://weatherbench2/datasets/era5/1959-2022-6h-1440x721.zarr \
  --climatology_path=gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_1440x721.zarr \
  --output_dir=$OUTDIR/1440x721/deterministic/ \
  --output_file_prefix=hres_vs_era_2020_ \
  --input_chunks=init_time=1,lead_time=1 \
  --fanout=27 \
  --regions=all \
  --eval_configs=deterministic,deterministic_temporal \
  --evaluate_climatology=False \
  --evaluate_persistence=False \
  --time_start=2020-01-01 \
  --time_stop=2020-12-31 \
  --variables=geopotential,temperature,u_component_of_wind,v_component_of_wind,specific_humidity,2m_temperature,10m_u_component_of_wind,10m_v_component_of_wind,mean_sea_level_pressure,total_precipitation_6hr,total_precipitation_24hr,10m_wind_speed,wind_speed \
  --compute_seeps=True \
  --use_beam=True
```

```bash
python evaluate.py -- \
  --forecast_path=gs://weatherbench2/datasets/hres/2016-2022-0012-1440x721.zarr \
  --obs_path=gs://weatherbench2/datasets/hres_t0/2016-2022-6h-1440x721.zarr \
  --lsm_dataset=gs://weatherbench2/datasets/era5/1959-2022-6h-1440x721.zarr \
  --climatology_path=gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_1440x721.zarr \
  --output_dir=$OUTDIR/1440x721/deterministic/ \
  --output_file_prefix=hres_vs_analysis_2020_ \
  --input_chunks=init_time=1,lead_time=1 \
  --fanout=27 \
  --regions=all \
  --eval_configs=deterministic,deterministic_temporal \
  --evaluate_climatology=False \
  --evaluate_persistence=False \
  --time_start=2020-01-01 \
  --time_stop=2020-12-31 \
  --variables=geopotential,temperature,u_component_of_wind,v_component_of_wind,specific_humidity,2m_temperature,10m_u_component_of_wind,10m_v_component_of_wind,mean_sea_level_pressure,total_precipitation_6hr,10m_wind_speed,wind_speed \
  --use_beam=True
```

```bash
python evaluate.py -- \
  --forecast_path=gs://weatherbench2/datasets/ens/2018-2022-1440x721_mean.zarr/ \
  --obs_path=gs://weatherbench2/datasets/era5/1959-2022-6h-1440x721.zarr \
  --climatology_path=gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_1440x721.zarr \
  --output_dir=$OUTDIR/1440x721/deterministic/ \
  --output_file_prefix=ens_vs_era_2020_ \
  --input_chunks=init_time=1,lead_time=1 \
  --fanout=27 \
  --regions=all \
  --eval_configs=deterministic,deterministic_temporal \
  --evaluate_climatology=False \
  --evaluate_persistence=False \
  --time_start=2020-01-01 \
  --time_stop=2020-12-31 \
  --variables=geopotential,temperature,u_component_of_wind,v_component_of_wind,specific_humidity,2m_temperature,10m_u_component_of_wind,10m_v_component_of_wind,mean_sea_level_pressure,total_precipitation_6hr,total_precipitation_24hr,10m_wind_speed,wind_speed \
  --compute_seeps=True \
  --use_beam=True
```

```bash
python evaluate.py -- \
  --forecast_path=gs://weatherbench2/datasets/ens/2018-2022-1440x721_mean.zarr/ \
  --obs_path=gs://weatherbench2/datasets/hres_t0/2016-2022-6h-1440x721.zarr \
  --lsm_dataset=gs://weatherbench2/datasets/era5/1959-2022-6h-1440x721.zarr \
  --climatology_path=gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_1440x721.zarr \
  --output_dir=$OUTDIR/1440x721/deterministic/ \
  --output_file_prefix=ens_vs_analysis_2020_ \
  --input_chunks=init_time=1,lead_time=1 \
  --fanout=27 \
  --regions=all \
  --eval_configs=deterministic,deterministic_temporal \
  --evaluate_climatology=False \
  --evaluate_persistence=False \
  --time_start=2020-01-01 \
  --time_stop=2020-12-31 \
  --variables=geopotential,temperature,u_component_of_wind,v_component_of_wind,specific_humidity,2m_temperature,10m_u_component_of_wind,10m_v_component_of_wind,mean_sea_level_pressure,total_precipitation_6hr,10m_wind_speed,wind_speed \
  --use_beam=True
```

```bash
python evaluate.py -- \
  --forecast_path=gs://weatherbench2/datasets/hres/2016-2022-0012-1440x721.zarr \
  --obs_path=gs://weatherbench2/datasets/era5/1959-2022-6h-1440x721.zarr \
  --climatology_path=gs://weatherbench2/datasets/era5-hourly-climatology/1990-2019_6h_1440x721.zarr \
  --output_dir=$OUTDIR/1440x721/deterministic/ \
  --output_file_prefix=climatology_vs_era_2020_ \
  --input_chunks=init_time=1,lead_time=1 \
  --fanout=27 \
  --regions=all \
  --eval_configs=deterministic,deterministic_temporal \
  --evaluate_climatology=True \
  --evaluate_persistence=False \
  --time_start=2020-01-01 \
  --time_stop=2020-12-31 \
  --variables=geopotential,temperature,u_component_of_wind,v_component_of_wind,specific_humidity,2m_temperature,10m_u_component_of_wind,10m_v_component_of_wind,mean_sea_level_pressure,total_precipitation_6hr,total_precipitation_24hr,10m_wind_speed,wind_speed \
  --compute_seeps=True \
  --use_beam=True
```

```bash
python evaluate.py -- \
  --forecast_path=gs://weatherbench2/datasets/hres/2016-2022-0012-1440x721.zarr
```

--------------------------------

### Import WeatherBench2 Configuration Module

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/evaluation.ipynb

Imports the necessary configuration module from the weatherbench2 library. This is a prerequisite for defining configurations.

```python
from weatherbench2 import config
```

--------------------------------

### Open Pangu-Weather Dataset

Source: https://github.com/google-research/weatherbench2/blob/main/docs/source/data-guide.ipynb

Opens the Pangu-Weather forecast dataset from Google Cloud Storage using xarray. This dataset includes various meteorological variables.

```python
xr.open_zarr('gs://weatherbench2/datasets/pangu/2018-2022_0012_0p25.zarr')
```