Fast domain-aware neural network emulation of a planetary boundary layer parameterization in a numerical weather forecast model
<p>Parameterizations for physical processes in weather and climate models are computationally expensive. We use model output from the Weather Research Forecast (WRF) climate model to train deep neural networks (DNNs) and evaluate whether trained DNNs can provide an accurate alternative to the...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Copernicus Publications
2019-10-01
|
| Series: | Geoscientific Model Development |
| Online Access: | https://www.geosci-model-dev.net/12/4261/2019/gmd-12-4261-2019.pdf |
| _version_ | 1818888376663670784 |
|---|---|
| author | J. Wang P. Balaprakash R. Kotamarthi |
| author_facet | J. Wang P. Balaprakash R. Kotamarthi |
| author_sort | J. Wang |
| collection | DOAJ |
| description | <p>Parameterizations for physical processes in weather and climate
models are computationally expensive. We use model output from the Weather
Research Forecast (WRF) climate model to train deep neural networks (DNNs)
and evaluate whether trained DNNs can provide an accurate alternative to the
physics-based parameterizations. Specifically, we develop an emulator using
DNNs for a planetary boundary layer (PBL) parameterization in the WRF model.
PBL parameterizations are used in atmospheric models to represent the
diurnal variation in the formation and collapse of the atmospheric boundary
layer – the lowest part of the atmosphere. The dynamics and turbulence, as
well as velocity, temperature, and humidity profiles within the boundary
layer are all critical for determining many of the physical processes in the
atmosphere. PBL parameterizations are used to represent these processes that
are usually unresolved in a typical climate model that operates at
horizontal spatial scales in the tens of kilometers. We demonstrate that a
domain-aware DNN, which takes account of underlying domain structure (e.g.,
nonlocal mixing), can successfully simulate the vertical profiles within the
boundary layer of velocities, temperature, and water vapor over the entire
diurnal cycle. Results also show that a single trained DNN from one location
can produce predictions of wind speed, temperature, and water vapor profiles
over nearby locations with similar terrain conditions with correlations
higher than 0.9 when compared with the WRF simulations used as the training
dataset.</p> |
| first_indexed | 2024-12-19T16:52:09Z |
| format | Article |
| id | doaj.art-c225447a5f85495f8bae16556cec3909 |
| institution | Directory Open Access Journal |
| issn | 1991-959X 1991-9603 |
| language | English |
| last_indexed | 2024-12-19T16:52:09Z |
| publishDate | 2019-10-01 |
| publisher | Copernicus Publications |
| record_format | Article |
| series | Geoscientific Model Development |
| spelling | doaj.art-c225447a5f85495f8bae16556cec39092022-12-21T20:13:31ZengCopernicus PublicationsGeoscientific Model Development1991-959X1991-96032019-10-01124261427410.5194/gmd-12-4261-2019Fast domain-aware neural network emulation of a planetary boundary layer parameterization in a numerical weather forecast modelJ. Wang0P. Balaprakash1R. Kotamarthi2Environmental Science Division, Argonne National Laboratory, 9700 South Cass Avenue,Lemont, IL 60439, USAMathematics and Computer Science Division, Argonne National Laboratory, 9700 South Cass Avenue, Lemont, IL 60439, USAEnvironmental Science Division, Argonne National Laboratory, 9700 South Cass Avenue,Lemont, IL 60439, USA<p>Parameterizations for physical processes in weather and climate models are computationally expensive. We use model output from the Weather Research Forecast (WRF) climate model to train deep neural networks (DNNs) and evaluate whether trained DNNs can provide an accurate alternative to the physics-based parameterizations. Specifically, we develop an emulator using DNNs for a planetary boundary layer (PBL) parameterization in the WRF model. PBL parameterizations are used in atmospheric models to represent the diurnal variation in the formation and collapse of the atmospheric boundary layer – the lowest part of the atmosphere. The dynamics and turbulence, as well as velocity, temperature, and humidity profiles within the boundary layer are all critical for determining many of the physical processes in the atmosphere. PBL parameterizations are used to represent these processes that are usually unresolved in a typical climate model that operates at horizontal spatial scales in the tens of kilometers. We demonstrate that a domain-aware DNN, which takes account of underlying domain structure (e.g., nonlocal mixing), can successfully simulate the vertical profiles within the boundary layer of velocities, temperature, and water vapor over the entire diurnal cycle. Results also show that a single trained DNN from one location can produce predictions of wind speed, temperature, and water vapor profiles over nearby locations with similar terrain conditions with correlations higher than 0.9 when compared with the WRF simulations used as the training dataset.</p>https://www.geosci-model-dev.net/12/4261/2019/gmd-12-4261-2019.pdf |
| spellingShingle | J. Wang P. Balaprakash R. Kotamarthi Fast domain-aware neural network emulation of a planetary boundary layer parameterization in a numerical weather forecast model Geoscientific Model Development |
| title | Fast domain-aware neural network emulation of a planetary boundary layer parameterization in a numerical weather forecast model |
| title_full | Fast domain-aware neural network emulation of a planetary boundary layer parameterization in a numerical weather forecast model |
| title_fullStr | Fast domain-aware neural network emulation of a planetary boundary layer parameterization in a numerical weather forecast model |
| title_full_unstemmed | Fast domain-aware neural network emulation of a planetary boundary layer parameterization in a numerical weather forecast model |
| title_short | Fast domain-aware neural network emulation of a planetary boundary layer parameterization in a numerical weather forecast model |
| title_sort | fast domain aware neural network emulation of a planetary boundary layer parameterization in a numerical weather forecast model |
| url | https://www.geosci-model-dev.net/12/4261/2019/gmd-12-4261-2019.pdf |
| work_keys_str_mv | AT jwang fastdomainawareneuralnetworkemulationofaplanetaryboundarylayerparameterizationinanumericalweatherforecastmodel AT pbalaprakash fastdomainawareneuralnetworkemulationofaplanetaryboundarylayerparameterizationinanumericalweatherforecastmodel AT rkotamarthi fastdomainawareneuralnetworkemulationofaplanetaryboundarylayerparameterizationinanumericalweatherforecastmodel |