google-research weatherbench2 issues

google-research / weatherbench2

A benchmark for the next generation of data-driven global weather models.

https://weatherbench2.readthedocs.io

Apache License 2.0

416 stars 42 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

No public description

#186 copybara-service[bot] opened 1 week ago
0
Clarify computation of accumulated precipitation in docstring and tests.

#185 copybara-service[bot] closed 1 week ago
0
RMSE scores on website differ from those on the corresponding Google Cloud dataset

#184 cindyl983 opened 2 weeks ago
1
Misc. fixes for newer versions of Xarray

#183 copybara-service[bot] closed 4 weeks ago
0
Misc. fixes for newer versions of Xarray

#182 copybara-service[bot] opened 4 weeks ago
0
No public description

#181 copybara-service[bot] opened 4 weeks ago
0
* Allow `--variables` in `compute_ensemble_mean`

#180 copybara-service[bot] closed 1 month ago
0
`resample_in_time` `BUGFIX`. Since the input dataset is modified (e.g. some variables dropped), some `input_chunks` will no longer be valid. This CL removes those dims from `input_chunks` to prevent missing dims errors.

#179 copybara-service[bot] closed 1 month ago
0
Evaluate Persistence does not work

#178 bar35643 opened 1 month ago
0
What region is data stored in?

#177 jrbourbeau opened 2 months ago
2
BUGFIX `compute_zonal_energy_spectrum.py` so it uses `BASE_VARIABLES` to filter the `source` Dataset right from the start. This means the dimensions in (the filtered) `source` will be the same as in the output.

#176 copybara-service[bot] closed 2 months ago
0
[numpy] Fix users of NumPy APIs that are removed in NumPy 2.0.

#175 copybara-service[bot] closed 2 months ago
0
The production of the low resolution prediction results

#174 rs-lsl opened 2 months ago
3
Possible bug with xarray dataset selection within beam

#173 jdwillard19 closed 2 months ago
2
Possible bug in schema.apply_time_conventions (?)

#172 timothyas opened 2 months ago
0
conservative regridding with land-sea mask

#171 markmbaum opened 3 months ago
0
Allow later compatible versions of xarray and pandas

#170 peter-dudbridge opened 3 months ago
0
Fix bug in time convention application

#169 peter-dudbridge opened 3 months ago
1
NaN values in relative humidity and vertical velocity

#168 SauryChen opened 3 months ago
0
No public description

#167 copybara-service[bot] opened 3 months ago
0
Use a faster version of rankdata in CRPS. This is identical to the V1.13 scipy version with `method='ordinal'`

#166 copybara-service[bot] closed 3 months ago
0
Speed up CRPS for small ensemble sizes by looping rather than sorting. For M ensemble members, this is `O(M²)` compute and `O(M)` memory, which beats sorting for smaller M.

#165 copybara-service[bot] closed 3 months ago
0
Speed up CRPS evaluation by caching spread computation (it sorts large arrays). This CL adds `@utils.dataset_safe_lru_cache`, which is similar to `functools.lru_cache`, but handles `xarray.Dataset` in addition to `Hashable`.

#164 copybara-service[bot] closed 3 months ago
0
No public description

#163 copybara-service[bot] opened 4 months ago
0
era5/1959-2023_01_10-full_37-1h-1440x721.zarr cant be opened

#162 markmbaum opened 4 months ago
0
ERA5 sea surface temperatures all NaN

#161 markmbaum opened 4 months ago
0
No public description

#160 copybara-service[bot] opened 4 months ago
0
IFS HRES t=0 “Analysis” dataset total_precipitation_6hr parameter is always 0

#159 agrigoriev opened 4 months ago
0
Allow NaN to propagate in EnsembleBrierScores and other Metrics. Previously, `NaN` forecasts ended up with a terrible Brier score. Now, they will get `NaN`. This happened for two reasons.

#158 copybara-service[bot] closed 4 months ago
0
Regional models not supported

#157 sehoffmann opened 4 months ago
0
Correct `DebiasedEnsembleBrierScore` docstring. Previously, the limiting (as ensemble size --> infinity) score was incorrect.

#156 copybara-service[bot] closed 4 months ago
0
Unit label incorrect for specific humidity plots

#155 oliverwm1 opened 4 months ago
1
No public description

#154 copybara-service[bot] opened 4 months ago
0
Use `Dataset.drop_vars` to implement `--drop_variables` flag in `slice_dataset.py`. This allows dropping of coordinates in addition to data vars.

#153 copybara-service[bot] closed 4 months ago
0
Use neuralGCM local output data as input gets error

#152 weatherforecasterwhai opened 4 months ago
1
Bump jinja2 from 3.1.2 to 3.1.4 in /docs

#151 dependabot[bot] opened 5 months ago
0
No public description

#150 copybara-service[bot] opened 5 months ago
0
Downloading data for one selected pressure level.

#149 ChenyuDongNUS opened 5 months ago
1
Ran build_cleaner. Also some internal dependency changes.

#148 copybara-service[bot] opened 5 months ago
0
No public description

#147 copybara-service[bot] opened 5 months ago
0
No public description

#146 copybara-service[bot] opened 5 months ago
0
Add `--num_threads` flag to weatherbench scripts. Added to all `DatasetToChunks` or `ChunksToZarr` call. Previously, some scripts had it and some didn't. Sometimes `num_threads=16` was hardcoded.

#145 copybara-service[bot] closed 5 months ago
0
Require `--quantiles` set if `'quantile' in --stats`. This prevents a mysterious `ZeroDivisionError`.

#144 copybara-service[bot] closed 5 months ago
0
Adjust windows so that `method=resample` and `method=rolling` align. Also set `skipna=False` when computing statistics. This is anticipation of the following development workflow:

#143 copybara-service[bot] closed 5 months ago
0
BUGFIX: Remove hard-coded time dimension in `resample_in_time.py`. Previously, we used `TIME_DIM.value` in some places, and `time` in others.

#142 copybara-service[bot] closed 5 months ago
0
`slice_dataset.py` added to `weatherbench2/scripts/`. This writes a new dataset that is a slice of the old one.

#141 copybara-service[bot] closed 5 months ago
0
Flexible time resampling added in `resample_in_time.py`.

#140 copybara-service[bot] closed 5 months ago
0
Fix URLs for Keisler data

#139 copybara-service[bot] closed 5 months ago
0
Run build_cleaner on weatherbench2/scripts

#138 copybara-service[bot] opened 5 months ago
0
Use `datetime64[ns]` dtypes rather than `datetime64[D]`. Day resolution is not fully supported in Xarray. See https://github.com/pydata/xarray/issues/7493

#137 copybara-service[bot] closed 5 months ago
0