DeepPrecip: a deep neural network for precipitation retrievals
<p>Remotely-sensed precipitation retrievals are critical for advancing our understanding of global energy and hydrologic cycles in remote regions. Radar reflectivity profiles of the lower atmosphere are commonly linked to precipitation through empirical power laws, but these relationships are...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Copernicus Publications
2022-10-01
|
| Series: | Atmospheric Measurement Techniques |
| Online Access: | https://amt.copernicus.org/articles/15/6035/2022/amt-15-6035-2022.pdf |
| _version_ | 1811337518815117312 |
|---|---|
| author | F. King G. Duffy G. Duffy L. Milani L. Milani C. G. Fletcher C. Pettersen K. Ebell |
| author_facet | F. King G. Duffy G. Duffy L. Milani L. Milani C. G. Fletcher C. Pettersen K. Ebell |
| author_sort | F. King |
| collection | DOAJ |
| description | <p>Remotely-sensed precipitation retrievals are critical for advancing our understanding of global energy and hydrologic cycles in remote regions. Radar reflectivity profiles of the lower atmosphere are commonly linked to precipitation through empirical power laws, but these relationships are tightly coupled to particle microphysical assumptions that do not generalize well to different regional climates. Here, we develop a robust, highly generalized precipitation retrieval algorithm from a deep convolutional neural network (DeepPrecip) to estimate 20 min average surface precipitation accumulation using near-surface radar data inputs. DeepPrecip displays a high retrieval skill and can accurately model total precipitation accumulation, with a mean square error (MSE) 160 % lower, on average, than current methods. DeepPrecip also outperforms a less complex machine learning retrieval algorithm, demonstrating the value of deep learning when applied to precipitation retrievals. Predictor importance analyses suggest that a combination of both near-surface (below 1 km) and higher-altitude (1.5–2 km) radar measurements are the primary features contributing to retrieval accuracy. Further, DeepPrecip closely captures total precipitation accumulation magnitudes and variability across nine distinct locations without requiring any explicit descriptions of particle microphysics or geospatial covariates. This research reveals the important role for deep learning in extracting relevant information about precipitation from atmospheric radar retrievals.</p> |
| first_indexed | 2024-04-13T17:56:08Z |
| format | Article |
| id | doaj.art-51713e4b0c114b3eabc9d61a1419227f |
| institution | Directory Open Access Journal |
| issn | 1867-1381 1867-8548 |
| language | English |
| last_indexed | 2024-04-13T17:56:08Z |
| publishDate | 2022-10-01 |
| publisher | Copernicus Publications |
| record_format | Article |
| series | Atmospheric Measurement Techniques |
| spelling | doaj.art-51713e4b0c114b3eabc9d61a1419227f2022-12-22T02:36:31ZengCopernicus PublicationsAtmospheric Measurement Techniques1867-13811867-85482022-10-01156035605010.5194/amt-15-6035-2022DeepPrecip: a deep neural network for precipitation retrievalsF. King0G. Duffy1G. Duffy2L. Milani3L. Milani4C. G. Fletcher5C. Pettersen6K. Ebell7Dept. of Geography & Environmental Management, University of Waterloo, 200 University Ave W, Waterloo, Ontario, CanadaJet Propulsion Laboratory, NASA, 4800 Oak Grove Dr, Pasadena, 91109, California, USAEarth and Environmental Sciences, University of Syracuse, 900 South Crouse Ave, Syracuse, New York, USAGoddard Space Flight Center, NASA, 8800 Greenbelt Rd, Greenbelt, Maryland, USAEarth System Science Interdisciplinary Center, University of Maryland, 5825 University Research Ct suite 4001, College Park, Maryland, USADept. of Geography & Environmental Management, University of Waterloo, 200 University Ave W, Waterloo, Ontario, CanadaClimate and Space Sciences and Engineering, University of Michigan, Climate and Space Research Building, 2455 Hayward St, Ann Arbor, Michigan, USAInstitute for Geophysics and Meteorology, University of Cologne, Albertus-Magnus-Platz, Cologne, Germany<p>Remotely-sensed precipitation retrievals are critical for advancing our understanding of global energy and hydrologic cycles in remote regions. Radar reflectivity profiles of the lower atmosphere are commonly linked to precipitation through empirical power laws, but these relationships are tightly coupled to particle microphysical assumptions that do not generalize well to different regional climates. Here, we develop a robust, highly generalized precipitation retrieval algorithm from a deep convolutional neural network (DeepPrecip) to estimate 20 min average surface precipitation accumulation using near-surface radar data inputs. DeepPrecip displays a high retrieval skill and can accurately model total precipitation accumulation, with a mean square error (MSE) 160 % lower, on average, than current methods. DeepPrecip also outperforms a less complex machine learning retrieval algorithm, demonstrating the value of deep learning when applied to precipitation retrievals. Predictor importance analyses suggest that a combination of both near-surface (below 1 km) and higher-altitude (1.5–2 km) radar measurements are the primary features contributing to retrieval accuracy. Further, DeepPrecip closely captures total precipitation accumulation magnitudes and variability across nine distinct locations without requiring any explicit descriptions of particle microphysics or geospatial covariates. This research reveals the important role for deep learning in extracting relevant information about precipitation from atmospheric radar retrievals.</p>https://amt.copernicus.org/articles/15/6035/2022/amt-15-6035-2022.pdf |
| spellingShingle | F. King G. Duffy G. Duffy L. Milani L. Milani C. G. Fletcher C. Pettersen K. Ebell DeepPrecip: a deep neural network for precipitation retrievals Atmospheric Measurement Techniques |
| title | DeepPrecip: a deep neural network for precipitation retrievals |
| title_full | DeepPrecip: a deep neural network for precipitation retrievals |
| title_fullStr | DeepPrecip: a deep neural network for precipitation retrievals |
| title_full_unstemmed | DeepPrecip: a deep neural network for precipitation retrievals |
| title_short | DeepPrecip: a deep neural network for precipitation retrievals |
| title_sort | deepprecip a deep neural network for precipitation retrievals |
| url | https://amt.copernicus.org/articles/15/6035/2022/amt-15-6035-2022.pdf |
| work_keys_str_mv | AT fking deepprecipadeepneuralnetworkforprecipitationretrievals AT gduffy deepprecipadeepneuralnetworkforprecipitationretrievals AT gduffy deepprecipadeepneuralnetworkforprecipitationretrievals AT lmilani deepprecipadeepneuralnetworkforprecipitationretrievals AT lmilani deepprecipadeepneuralnetworkforprecipitationretrievals AT cgfletcher deepprecipadeepneuralnetworkforprecipitationretrievals AT cpettersen deepprecipadeepneuralnetworkforprecipitationretrievals AT kebell deepprecipadeepneuralnetworkforprecipitationretrievals |