Evaluating methods for reconstructing large gaps in historic snow depth time series

<p>Historic measurements are often temporally incomplete and may contain longer periods of missing data, whereas climatological analyses require continuous measurement records. This is also valid for historic manual snow depth (HS) measurement time series, for which even whole winters can be m...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	J. Aschauer, C. Marty
Formato:	article
Lenguaje:	EN
Publicado:	Copernicus Publications 2021
Materias:	Geophysics. Cosmic physics QC801-809
Acceso en línea:	https://doaj.org/article/c75c2f3ab9c746818913e0c8a29fab2d
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:c75c2f3ab9c746818913e0c8a29fab2d
record_format	dspace
spelling	oai:doaj.org-article:c75c2f3ab9c746818913e0c8a29fab2d2021-11-24T12:44:16ZEvaluating methods for reconstructing large gaps in historic snow depth time series10.5194/gi-10-297-20212193-08562193-0864https://doaj.org/article/c75c2f3ab9c746818913e0c8a29fab2d2021-11-01T00:00:00Zhttps://gi.copernicus.org/articles/10/297/2021/gi-10-297-2021.pdfhttps://doaj.org/toc/2193-0856https://doaj.org/toc/2193-0864<p>Historic measurements are often temporally incomplete and may contain longer periods of missing data, whereas climatological analyses require continuous measurement records. This is also valid for historic manual snow depth (HS) measurement time series, for which even whole winters can be missing in a station record, and suitable methods have to be found to reconstruct the missing data. Daily in situ HS data from 126 nivo-meteorological stations in Switzerland in an altitudinal range of 230 to 2536 m above sea level are used to compare six different methods for reconstructing long gaps in manual HS time series by performing a “leave-one-winter-out” cross-validation in 21 winters at 33 evaluation stations. Synthetic gaps of one winter length are filled with bias-corrected data from the best-correlated neighboring station (BSC), inverse distance-weighted (IDW) spatial interpolation, a weighted normal ratio (WNR) method, elastic net (ENET) regression, random forest (RF) regression and a temperature index snow model (SM). Methods that use neighboring station data are tested in two station networks with different density. The ENET, RF, SM and WNR methods are able to reconstruct missing data with a coefficient of determination (<span class="inline-formula"><i>r</i><sup>2</sup></span>) above 0.8 regardless of the two station networks used. The median root mean square error (RMSE) in the filled winters is below 5 cm for all methods. The two annual climate indicators, average snow depth in a winter (HSavg) and maximum snow depth in a winter (HSmax), can be reproduced by ENET, RF, SM and WNR well, with <span class="inline-formula"><i>r</i><sup>2</sup></span> above 0.85 in both station networks. For the inter-station approaches, scores for the number of snow days with <span class="inline-formula">HS>1</span> cm (dHS1) are clearly weaker and, except for BCS, positively biased with RMSE of 18–33 d. SM reveals the best performance with <span class="inline-formula"><i>r</i><sup>2</sup></span> of 0.93 and RMSE of 15 d for dHS1. Snow depth seems to be a relatively good-natured parameter when it comes to gap filling of HS data with neighboring stations in a climatological use case. However, when station networks get sparse and if the focus is set on dHS1, temperature index snow models can serve as a suitable alternative to classic inter-station gap filling approaches.</p>J. AschauerC. MartyCopernicus PublicationsarticleGeophysics. Cosmic physicsQC801-809ENGeoscientific Instrumentation, Methods and Data Systems, Vol 10, Pp 297-312 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Geophysics. Cosmic physics QC801-809
spellingShingle	Geophysics. Cosmic physics QC801-809 J. Aschauer C. Marty Evaluating methods for reconstructing large gaps in historic snow depth time series
description	<p>Historic measurements are often temporally incomplete and may contain longer periods of missing data, whereas climatological analyses require continuous measurement records. This is also valid for historic manual snow depth (HS) measurement time series, for which even whole winters can be missing in a station record, and suitable methods have to be found to reconstruct the missing data. Daily in situ HS data from 126 nivo-meteorological stations in Switzerland in an altitudinal range of 230 to 2536 m above sea level are used to compare six different methods for reconstructing long gaps in manual HS time series by performing a “leave-one-winter-out” cross-validation in 21 winters at 33 evaluation stations. Synthetic gaps of one winter length are filled with bias-corrected data from the best-correlated neighboring station (BSC), inverse distance-weighted (IDW) spatial interpolation, a weighted normal ratio (WNR) method, elastic net (ENET) regression, random forest (RF) regression and a temperature index snow model (SM). Methods that use neighboring station data are tested in two station networks with different density. The ENET, RF, SM and WNR methods are able to reconstruct missing data with a coefficient of determination (<span class="inline-formula"><i>r</i><sup>2</sup></span>) above 0.8 regardless of the two station networks used. The median root mean square error (RMSE) in the filled winters is below 5 cm for all methods. The two annual climate indicators, average snow depth in a winter (HSavg) and maximum snow depth in a winter (HSmax), can be reproduced by ENET, RF, SM and WNR well, with <span class="inline-formula"><i>r</i><sup>2</sup></span> above 0.85 in both station networks. For the inter-station approaches, scores for the number of snow days with <span class="inline-formula">HS>1</span> cm (dHS1) are clearly weaker and, except for BCS, positively biased with RMSE of 18–33 d. SM reveals the best performance with <span class="inline-formula"><i>r</i><sup>2</sup></span> of 0.93 and RMSE of 15 d for dHS1. Snow depth seems to be a relatively good-natured parameter when it comes to gap filling of HS data with neighboring stations in a climatological use case. However, when station networks get sparse and if the focus is set on dHS1, temperature index snow models can serve as a suitable alternative to classic inter-station gap filling approaches.</p>
format	article
author	J. Aschauer C. Marty
author_facet	J. Aschauer C. Marty
author_sort	J. Aschauer
title	Evaluating methods for reconstructing large gaps in historic snow depth time series
title_short	Evaluating methods for reconstructing large gaps in historic snow depth time series
title_full	Evaluating methods for reconstructing large gaps in historic snow depth time series
title_fullStr	Evaluating methods for reconstructing large gaps in historic snow depth time series
title_full_unstemmed	Evaluating methods for reconstructing large gaps in historic snow depth time series
title_sort	evaluating methods for reconstructing large gaps in historic snow depth time series
publisher	Copernicus Publications
publishDate	2021
url	https://doaj.org/article/c75c2f3ab9c746818913e0c8a29fab2d
work_keys_str_mv	AT jaschauer evaluatingmethodsforreconstructinglargegapsinhistoricsnowdepthtimeseries AT cmarty evaluatingmethodsforreconstructinglargegapsinhistoricsnowdepthtimeseries
_version_	1718415074372091904

Evaluating methods for reconstructing large gaps in historic snow depth time series

Ejemplares similares