Spatiotemporal reconstruction of global ocean surface pCO2 based on optimized random forest

The Science of The Total Environment, Год журнала: 2023, Номер 912, С. 169209 - 169209

Опубликована: Дек. 11, 2023

The partial pressure of ocean surface CO2 (pCO2) plays an important role in quantifying the carbon budget and assessing acidification. For such a vast complex system as global ocean, most current research practices tend to study into regions. In order reveal overall characteristics avoid mutual influence between zones, holistic method was used detect correlation twelve predictive factors, including chlorophyll concentration (Chlor_a), diffuse attenuation coefficient at 490 nm (Kd_490), density mixed layer thickness (Mlotst), eastward velocity (East), northward (North), salinity (Sal), temperature (Temp), dissolved iron (Fe), silicate (Si), nitrate (NO3), potential hydrogen (pH), phosphate (PO4), scale. Based on measured data from Global Surface pCO2 (LDEO) database, combined with National Aeronautics Space Administration (NASA) Ocean Color satellite Copernicus reanalysis data, improved optimized random forest (ORF) is proposed for reconstruction pCO2, compared various machine learning methods. results indicate that ORF accurate modeling scale (mean absolute error 6.27μatm, root mean square 15.34μatm, R2 0.92). independent observations LDEO dataset time series observation stations, model further validated, distribution map 0.25° × 2010 2019 reconstructed, which significance cycle assessment.

Язык: Английский

---

The magnitude and potential of the sedimentary carbon sink in the Eastern China Marginal Seas

Palaeogeography Palaeoclimatology Palaeoecology, Год журнала: 2024, Номер 655, С. 112482 - 112482

Опубликована: Сен. 4, 2024

Язык: Английский

---

Climate Change Drove the Decline in Yangtze Estuary Net Primary Production Over the Past Two Decades

Environmental Science & Technology, Год журнала: 2024, Номер 58(43), С. 19305 - 19314

Опубликована: Окт. 15, 2024

Net primary productivity (NPP) is highly sensitive to multiple stressors under progressive and intensifying climate change anthropogenic impacts. The importance of understanding spatiotemporal distribution patterns the associated driving factors that govern estuary NPP paramount for regional carbon (C) budget assessments. Using a combined remote sensing machine learning (ML) approach, average Yangtze Estuarine–offshore continuum (YEOC) was measured at 273.19 ± 21.26 mgC m–2 day–1 over past two decades. Temporally, exhibited significant downward trend between 2002 2022. Climate (climate fluctuations, sea level rise, discharge) drove phytoplankton biomass (Chl-a) while light conditions (PAR Kd490) affected photosynthesis rates. Together, they can explain 65% variation. Anthropogenic disturbances (i.e., damming nutrient emissions) were not significant. Additionally, changes in decreased C sequestration rates from 11.9 10.4 Tg year–1, reducing estuary's sink capacity, which relies on biological fixation. This study highlights climate's influence transformation YEOC enhancing our response EOC budgets activities.

Язык: Английский

---

Research on Atlantic Surface Pco2 Reconstruction Based on Machine Learning

Опубликована: Янв. 1, 2024

Язык: Английский

---

Interannual variability of air-sea CO2 exchange in the Northern Yellow Sea and its underlying mechanisms

Frontiers in Marine Science, Год журнала: 2024, Номер 11

Опубликована: Июль 1, 2024

A three-dimensional (3-D) physical-biogeochemical-carbon cycle coupled model is used to investigate the interannual variability of air-sea carbon dioxide (CO 2 ) flux ( F CO in Northern Yellow Sea (NYS) from 2009 2018. The verification indicate that simulation results for multiple variables exhibit consistency and fit well with observed data. study show although multi-year average NYS close source-sink balance, there are obvious differences between different years. In particular, a relatively strong source atmospheric (1.0 mmol m –2 d –1 exhibited 2014, while sink (–0.7 emerges 2016. Mechanism analysis indicates abnormally high temperature main controlling factor efflux rate low dissolved inorganic (DIC) concentration contributing influx Further reveals primary reason DIC since onset winter 2016 net decrease 2015, influenced by community production summer advection processes during autumn. 2015 excessive reduction through biological processes. addition, due northeasterly wind event November low-concentration-DIC water (YS) extends into Bohai (BS). This further leads higher BS upper mixed layer increases inflow Southern (SYS) NYS. These ultimately result abnormal autumn 2015. enriches our understanding NYS, which will not only help reveal variations under human activities climate change, but also provide useful information guiding comprehensive assessment budget.

Язык: Английский

---

Satellite Estimation of Global Sea-Air Carbon Dioxide Fugacity from 2000 to 2020 based on Machine Learning Models

Authorea (Authorea), Год журнала: 2024, Номер unknown

Опубликована: Сен. 2, 2024

Based on over 160,000 quality-controlled measurements of surface ocean carbon dioxide fugacity from 2000 to 2020, and employing machine learning methods, a satellite-based assessment model for sea-air (fCO2) has been developed, aiming reveal global changes in the past 20 years. Examining factors affecting fCO2, this study encompasses satellite data coordinates, basic seawater parameters such as salinity ,temperature, wind speed, acidity alkalinity, velocity, geostrophic sea water partial pressure water, downward mass flux expressed carbon, well concentrations dissolved inorganic phosphate, nitrate concentration, thickness marine mixed layer, total silicate influencing solubility, chlorophyll concentration indicating biological activity, oxygen concentration. A comparative analysis was conducted various including XGBoost, Random Forest, Light Gradient Booster, Feedforward Neural Network, Convolutional Backpropagation Network. XGBoost algorithm chosen construction based best performance. The results independent field validation indicate that low root mean square error (RMES=18.08μatm) absolute percentage (MAPE=1.1%) R-squared (R2=0.91). Finally, distribution at resolution 0.25°×0.25° 2020 reconstructed. oceans shown slow upward trend, years, increased by 6.7%.

Язык: Английский

---

Climate change in interaction with global carbon cycle

Elsevier eBooks, Год журнала: 2024, Номер unknown, С. 227 - 257

Опубликована: Окт. 11, 2024

Язык: Английский

---

Seasonal variation of CO2 air-sea flux and effects of warming in the Kuroshio current of the East China Sea

Marine Chemistry, Год журнала: 2024, Номер 267, С. 104469 - 104469

Опубликована: Ноя. 1, 2024

Язык: Английский

---