On using self‐organizing maps and discretized Sammon maps to study links between atmospheric circulation and weather extremes

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F68378289%3A_____%2F23%3A00566010" target="_blank" >RIV/68378289:_____/23:00566010 - isvavai.cz</a>

Výsledek na webu

<a href="https://rmets.onlinelibrary.wiley.com/doi/10.1002/joc.7996" target="_blank" >https://rmets.onlinelibrary.wiley.com/doi/10.1002/joc.7996</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1002/joc.7996" target="_blank" >10.1002/joc.7996</a>

Jazyk výsledku

angličtina

Název v původním jazyce

On using self‐organizing maps and discretized Sammon maps to study links between atmospheric circulation and weather extremes

Popis výsledku v původním jazyce

Classifications by self-organizing maps (SOMs) have been used to study links between extreme weather events and atmospheric circulation. However, SOMs may often provide a too generalized image of circulation, which is too coarse to properly resolve meteorological extremes. Several approaches have been used to circumvent this limitation, including an increase in the number of nodes and/or relaxation of the stress on topological ordering of nodes, through adjusting the neighbourhood radius. Here, we (1) test whether weighting of training fields based on extremity of their circulation could lead to SOMs that would better represent circulation extremes without impairing the topological ordering of nodes, and (2) we investigate whether Sammon maps (SM), which have so far been limited to validation of SOMs in climatology, could provide a viable alternative to SOMs in studying weather extremes. We show that removing a certain percentage of circulation fields closest to the climatological mean has a greater effect on the representation of circulation extremes than more than doubling the SOM array size. Additionally, this step allows for decreasing the neighbourhood radius, the parameter responsible for topological ordering, which would not otherwise be possible without impairing the topology of the map. The SOMs are compared with two newly developed methods that discretize SMs based on arbitrarily chosen and distribution-based bins. We demonstrate the usefulness of both methods in climate model validation and show that both methods are able to capture circulation variability in its extremes in more detail than the SOMs. The distribution-based method does not suffer from sample-size issues and outperforms all other classifications in extracting ERA5 patterns leading to extreme temperature minima over central Europe. Our results suggest that classifications based on SM could be a viable alternative to SOMs when SOMs lack the ability to discriminate circulation outliers leading to temperature extremes.

Název v anglickém jazyce

On using self‐organizing maps and discretized Sammon maps to study links between atmospheric circulation and weather extremes

Popis výsledku anglicky

Classifications by self-organizing maps (SOMs) have been used to study links between extreme weather events and atmospheric circulation. However, SOMs may often provide a too generalized image of circulation, which is too coarse to properly resolve meteorological extremes. Several approaches have been used to circumvent this limitation, including an increase in the number of nodes and/or relaxation of the stress on topological ordering of nodes, through adjusting the neighbourhood radius. Here, we (1) test whether weighting of training fields based on extremity of their circulation could lead to SOMs that would better represent circulation extremes without impairing the topological ordering of nodes, and (2) we investigate whether Sammon maps (SM), which have so far been limited to validation of SOMs in climatology, could provide a viable alternative to SOMs in studying weather extremes. We show that removing a certain percentage of circulation fields closest to the climatological mean has a greater effect on the representation of circulation extremes than more than doubling the SOM array size. Additionally, this step allows for decreasing the neighbourhood radius, the parameter responsible for topological ordering, which would not otherwise be possible without impairing the topology of the map. The SOMs are compared with two newly developed methods that discretize SMs based on arbitrarily chosen and distribution-based bins. We demonstrate the usefulness of both methods in climate model validation and show that both methods are able to capture circulation variability in its extremes in more detail than the SOMs. The distribution-based method does not suffer from sample-size issues and outperforms all other classifications in extracting ERA5 patterns leading to extreme temperature minima over central Europe. Our results suggest that classifications based on SM could be a viable alternative to SOMs when SOMs lack the ability to discriminate circulation outliers leading to temperature extremes.

Druh

J_imp - Článek v periodiku v databázi Web of Science

CEP obor

—

OECD FORD obor

10510 - Climatic research

Projekt

<a href="/cs/project/GJ19-24425Y" target="_blank" >GJ19-24425Y: Nebezpečné zimní počasí ve výstupech klimatických modelů</a><br>

Návaznosti

I - Institucionalni podpora na dlouhodoby koncepcni rozvoj vyzkumne organizace

Rok uplatnění

2023

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název periodika

International Journal of Climatology

ISSN

0899-8418

e-ISSN

1097-0088

Svazek periodika

43

Číslo periodika v rámci svazku

6

Stát vydavatele periodika

GB - Spojené království Velké Británie a Severního Irska

Počet stran výsledku

21

Strana od-do

2678-2698

Kód UT WoS článku

000916901300001

EID výsledku v databázi Scopus

2-s2.0-85146657177