Open-Source Toolkit for Arctic d-excess and Sea-Ice Interaction Research: Lagged Correlation, Causal Connectivity, and Assimilated Model Evaluation

Fuente: Zenodo
Salvato in:
Dettagli Bibliografici
Autori principali: Heydarizad, Mojtaba, Liu, Zhongfang, Parker, Mason, Mora, Thiago, Thakur, Edward
Natura: Recurso digital
Lingua:inglese
Pubblicazione: Zenodo 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866901983736102912
author Heydarizad, Mojtaba
Liu, Zhongfang
Parker, Mason
Mora, Thiago
Thakur, Edward
author_facet Heydarizad, Mojtaba
Liu, Zhongfang
Parker, Mason
Mora, Thiago
Thakur, Edward
contents <p>This archive contains the analysis and figure-generation code used to study links between Arctic sea-ice variability and precipitation d-excess at multiple stations. The toolkit computes best-lag correlations, extracts top-K “parent” sea-ice regions per station (with sign and lag), visualizes radial edge networks, and compares stations using an <em>F1</em> similarity metric with a ±1-month lag tolerance. It also includes baseline modeling utilities (RFE, LASSO), a simple neural-network + genetic algorithm example, SHAP-based feature importance, an entropy model-averaging (EMA) routine, and a scalar Kalman filter demonstration for data assimilation.</p>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_17411473
institution Zenodo
language eng
publishDate 2025
publisher Zenodo
record_format zenodo
spellingShingle Open-Source Toolkit for Arctic d-excess and Sea-Ice Interaction Research: Lagged Correlation, Causal Connectivity, and Assimilated Model Evaluation
Heydarizad, Mojtaba
Liu, Zhongfang
Parker, Mason
Mora, Thiago
Thakur, Edward
Supervised Machine Learning
Arctic region
Stable isotopes in the atmosphere
sea ice
<p>This archive contains the analysis and figure-generation code used to study links between Arctic sea-ice variability and precipitation d-excess at multiple stations. The toolkit computes best-lag correlations, extracts top-K “parent” sea-ice regions per station (with sign and lag), visualizes radial edge networks, and compares stations using an <em>F1</em> similarity metric with a ±1-month lag tolerance. It also includes baseline modeling utilities (RFE, LASSO), a simple neural-network + genetic algorithm example, SHAP-based feature importance, an entropy model-averaging (EMA) routine, and a scalar Kalman filter demonstration for data assimilation.</p>
title Open-Source Toolkit for Arctic d-excess and Sea-Ice Interaction Research: Lagged Correlation, Causal Connectivity, and Assimilated Model Evaluation
topic Supervised Machine Learning
Arctic region
Stable isotopes in the atmosphere
sea ice
url https://doi.org/10.5281/zenodo.17411473