Probabilistic precipitation downscaling for ungauged mountain sites: a pilot study for the Hindu Kush Himalaya
<p>This study introduces a novel approach to post-processing (i.e. downscaling and bias-correcting) reanalysis-driven regional climate model daily precipitation outputs that can be generalised to ungauged mountain locations by leveraging sparse in situ observations and a probabilistic regres...
Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Copernicus Publications
2025-07-01
|
| Series: | Hydrology and Earth System Sciences |
| Online Access: | https://hess.copernicus.org/articles/29/3073/2025/hess-29-3073-2025.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849429457863966720 |
|---|---|
| author | M. Girona-Mata M. Girona-Mata A. Orr M. Widmann D. Bannister G. H. Dars S. Hosking S. Hosking J. Norris D. Ocio T. Phillips J. Steiner J. Steiner R. E. Turner |
| author_facet | M. Girona-Mata M. Girona-Mata A. Orr M. Widmann D. Bannister G. H. Dars S. Hosking S. Hosking J. Norris D. Ocio T. Phillips J. Steiner J. Steiner R. E. Turner |
| author_sort | M. Girona-Mata |
| collection | DOAJ |
| description | <p>This study introduces a novel approach to post-processing (i.e. downscaling and bias-correcting) reanalysis-driven regional climate model daily precipitation outputs that can be generalised to ungauged mountain locations by leveraging sparse in situ observations and a probabilistic regression framework. We call this post-processing approach generalised probabilistic regression (GPR) and implement it using both generalised linear models and artificial neural networks (i.e. multi-layer perceptrons). By testing the GPR post-processing approach across three Hindu Kush Himalaya (HKH) basins with varying hydro-meteorological characteristics and four experiments, which are representative of real-world scenarios, we find it performs consistently much better than both raw regional climate model output and deterministic bias correction methods for generalising daily precipitation post-processing to ungauged locations. We also find that GPR models are flexible and can be trained using data from a single region or multiple regions combined together, without major impacts on model performance. Additionally, we show that the GPR approach results in superior skill for post-processing entirely ungauged regions, by leveraging data from other regions as well as ungauged high-elevation ranges. This suggests that GPR models have potential for extending post-processing of daily precipitation to ungauged areas of HKH. Whilst multi-layer perceptrons yield marginally improved results overall, generalised linear models are a robust choice, particularly for data-scarce scenarios, i.e. post-processing extreme precipitation events and generalising to completely ungauged regions.</p> |
| format | Article |
| id | doaj-art-dfb71a5e0f6d43f29e9feecf081270ea |
| institution | Kabale University |
| issn | 1027-5606 1607-7938 |
| language | English |
| publishDate | 2025-07-01 |
| publisher | Copernicus Publications |
| record_format | Article |
| series | Hydrology and Earth System Sciences |
| spelling | doaj-art-dfb71a5e0f6d43f29e9feecf081270ea2025-08-20T03:28:21ZengCopernicus PublicationsHydrology and Earth System Sciences1027-56061607-79382025-07-01293073310010.5194/hess-29-3073-2025Probabilistic precipitation downscaling for ungauged mountain sites: a pilot study for the Hindu Kush HimalayaM. Girona-Mata0M. Girona-Mata1A. Orr2M. Widmann3D. Bannister4G. H. Dars5S. Hosking6S. Hosking7J. Norris8D. Ocio9T. Phillips10J. Steiner11J. Steiner12R. E. Turner13British Antarctic Survey, UK Research and Innovation, Cambridge, UKDepartment of Engineering, University of Cambridge, Cambridge, UKBritish Antarctic Survey, UK Research and Innovation, Cambridge, UKSchool of Geography, Earth and Environmental Sciences, University of Birmingham, Birmingham, UKWTW Research Network, WTW, London, UKU.S.-Pakistan Center for Advanced Studies in Water, Mehran University of Engineering and Technology, Jamshoro, PakistanBritish Antarctic Survey, UK Research and Innovation, Cambridge, UKThe Alan Turing Institute, London, UKAtmospheric and Oceanic Sciences, University of California Los Angeles, Los Angeles, CA, USAMott MacDonald, Cambridge, UKBritish Antarctic Survey, UK Research and Innovation, Cambridge, UKInstitute of Geography and Regional Science, University of Graz, Graz, AustriaHimalayan University Consortium, Lalitpur, NepalDepartment of Engineering, University of Cambridge, Cambridge, UK<p>This study introduces a novel approach to post-processing (i.e. downscaling and bias-correcting) reanalysis-driven regional climate model daily precipitation outputs that can be generalised to ungauged mountain locations by leveraging sparse in situ observations and a probabilistic regression framework. We call this post-processing approach generalised probabilistic regression (GPR) and implement it using both generalised linear models and artificial neural networks (i.e. multi-layer perceptrons). By testing the GPR post-processing approach across three Hindu Kush Himalaya (HKH) basins with varying hydro-meteorological characteristics and four experiments, which are representative of real-world scenarios, we find it performs consistently much better than both raw regional climate model output and deterministic bias correction methods for generalising daily precipitation post-processing to ungauged locations. We also find that GPR models are flexible and can be trained using data from a single region or multiple regions combined together, without major impacts on model performance. Additionally, we show that the GPR approach results in superior skill for post-processing entirely ungauged regions, by leveraging data from other regions as well as ungauged high-elevation ranges. This suggests that GPR models have potential for extending post-processing of daily precipitation to ungauged areas of HKH. Whilst multi-layer perceptrons yield marginally improved results overall, generalised linear models are a robust choice, particularly for data-scarce scenarios, i.e. post-processing extreme precipitation events and generalising to completely ungauged regions.</p>https://hess.copernicus.org/articles/29/3073/2025/hess-29-3073-2025.pdf |
| spellingShingle | M. Girona-Mata M. Girona-Mata A. Orr M. Widmann D. Bannister G. H. Dars S. Hosking S. Hosking J. Norris D. Ocio T. Phillips J. Steiner J. Steiner R. E. Turner Probabilistic precipitation downscaling for ungauged mountain sites: a pilot study for the Hindu Kush Himalaya Hydrology and Earth System Sciences |
| title | Probabilistic precipitation downscaling for ungauged mountain sites: a pilot study for the Hindu Kush Himalaya |
| title_full | Probabilistic precipitation downscaling for ungauged mountain sites: a pilot study for the Hindu Kush Himalaya |
| title_fullStr | Probabilistic precipitation downscaling for ungauged mountain sites: a pilot study for the Hindu Kush Himalaya |
| title_full_unstemmed | Probabilistic precipitation downscaling for ungauged mountain sites: a pilot study for the Hindu Kush Himalaya |
| title_short | Probabilistic precipitation downscaling for ungauged mountain sites: a pilot study for the Hindu Kush Himalaya |
| title_sort | probabilistic precipitation downscaling for ungauged mountain sites a pilot study for the hindu kush himalaya |
| url | https://hess.copernicus.org/articles/29/3073/2025/hess-29-3073-2025.pdf |
| work_keys_str_mv | AT mgironamata probabilisticprecipitationdownscalingforungaugedmountainsitesapilotstudyforthehindukushhimalaya AT mgironamata probabilisticprecipitationdownscalingforungaugedmountainsitesapilotstudyforthehindukushhimalaya AT aorr probabilisticprecipitationdownscalingforungaugedmountainsitesapilotstudyforthehindukushhimalaya AT mwidmann probabilisticprecipitationdownscalingforungaugedmountainsitesapilotstudyforthehindukushhimalaya AT dbannister probabilisticprecipitationdownscalingforungaugedmountainsitesapilotstudyforthehindukushhimalaya AT ghdars probabilisticprecipitationdownscalingforungaugedmountainsitesapilotstudyforthehindukushhimalaya AT shosking probabilisticprecipitationdownscalingforungaugedmountainsitesapilotstudyforthehindukushhimalaya AT shosking probabilisticprecipitationdownscalingforungaugedmountainsitesapilotstudyforthehindukushhimalaya AT jnorris probabilisticprecipitationdownscalingforungaugedmountainsitesapilotstudyforthehindukushhimalaya AT docio probabilisticprecipitationdownscalingforungaugedmountainsitesapilotstudyforthehindukushhimalaya AT tphillips probabilisticprecipitationdownscalingforungaugedmountainsitesapilotstudyforthehindukushhimalaya AT jsteiner probabilisticprecipitationdownscalingforungaugedmountainsitesapilotstudyforthehindukushhimalaya AT jsteiner probabilisticprecipitationdownscalingforungaugedmountainsitesapilotstudyforthehindukushhimalaya AT returner probabilisticprecipitationdownscalingforungaugedmountainsitesapilotstudyforthehindukushhimalaya |