Exploring the Potential of Long Short-Term Memory Networks for Predicting Net CO2 Exchange Across Various Ecosystems With Multi-Source Data DOI Open Access

Chengcheng Huang,

Wei He, Jinxiu Liu

et al.

Authorea (Authorea), Journal Year: 2023, Volume and Issue: unknown

Published: Nov. 14, 2023

Upscaling flux tower measurements based on machine learning (ML) algorithms is an essential approach for large-scale net ecosystem CO2 exchange (NEE) estimation, but existing ML upscaling methods face some challenges, particularly in capturing NEE interannual variations (IAVs) that may relate to lagged effects. With the capacity of characterizing temporal memory effects, Long Short-Term Memory (LSTM) networks are expected help solve this problem. Here we explored potential LSTM predicting across various ecosystems using data over 82 sites North America. The model with differentiated plant function types (PFTs) demonstrates capability explain 79.19% (R2 = 0.79) monthly within testing set, RMSE and MAE values 0.89 0.57 g C m-2 d-1 respectively (r 0.89, p < 0.001). Moreover, performed robustly cross-site variability, 67.19% can be predicted by both models without distinguished PFTs showing improved predictive ability. Most importantly, IAV highly correlated observations 0.81, 0.001), clearly outperforming random forest -0.21, 0.011). Among all nine PFTs, solar-induced chlorophyll fluorescence, downward shortwave radiation, leaf area index most important variables explaining variations, collectively accounting approximately 54.01% total. This study highlights great improving carbon multi-source remote sensing data.

Language: Английский

Weaker regional carbon uptake albeit with stronger seasonal amplitude in northern mid-latitudes estimated by higher resolution GEOS-Chem model DOI Creative Commons
Zhiqiang Liu, Ning Zeng, Yun Liu

et al.

The Science of The Total Environment, Journal Year: 2023, Volume and Issue: 912, P. 169477 - 169477

Published: Dec. 22, 2023

Terrestrial ecosystem in the Northern Hemisphere is characterized by a substantial carbon sink recent decades. However, inferred from atmospheric CO2 data usually larger than process- and inventory-based estimates, resulting release or near-neutral exchange tropics. The approach known to be uncertain due systematic biases of coarse transport model simulation. Compared coarse-resolution inverse estimate at 4° × 5° using GEOS-Chem integrated region N. America, E. Asia, Europe 2015 2018, annual native high-resolution 0.5° 0.625° reduced −3.0±0.08 gigatons per year (GtC yr−1) −2.15±0.08 GtC yr−1 prominent more during non-growing seasons. major reductions concentrate mid-latitudes (20°N–45°N), where mean land sinks China USA are 0.64±0.03 0.35±0.02 0.14±0.03 0.15±0.02 yr−1, respectively. tends trap both uptake signal within planetary boundary layer, weaker estimates biosphere seasonal strength. Since strong fossil fuel emissions persistently released surface, trapped leads stronger uptakes. These results suggest that inversion with accurate vertical meridional urgently needed targeting national neutrality.

Language: Английский

Citations

4
Estimation of carbon emissions in various clustered regions of China based on OCO-2 satellite XCO2 data and random forest modelling DOI

Yibing Tan,

Shanshan Wang, Ruibin Xue

et al.

Atmospheric Environment, Journal Year: 2024, Volume and Issue: 338, P. 120860 - 120860

Published: Oct. 8, 2024

Language: Английский

Citations

1
Improved estimates of net ecosystem exchanges in mega-countries using GOSAT and OCO-2 observations DOI Creative Commons
Lingyu Zhang, Fei Jiang, Wei He

et al.

Communications Earth & Environment, Journal Year: 2024, Volume and Issue: 5(1)

Published: Nov. 25, 2024

Accurate national terrestrial net ecosystem exchange estimates are crucial for the global stocktake. Net from different inversion models vary greatly at scale, and relative impacts of prior fluxes observations on these inversions remain unclear. Here we estimate 51 land regions 2017-2019 period, focusing 10 largest countries, using 12 biosphere XCO2 retrievals GOSAT OCO-2 satellites as constraints. The average uncertainty reduction countries increases 37% with 45% to 50% combined observations, indicating a trend towards robust estimates. At finer spatial scales, even is only 33%, i.e., flux dominates This finding underscores critical importance integrating multi-source refining improve accuracy carbon Choice model input satellite data has significant impact modelled dioxide its associated large according atmospheric data.

Language: Английский

Citations

1
Spatiotemporal Variations in Carbon Sources and Sinks in National Park Ecosystem and the Impact of Tourism DOI Open Access

Quanxu Hu,

Jinhe Zhang,

Huaju Xue

et al.

Sustainability, Journal Year: 2024, Volume and Issue: 16(18), P. 7895 - 7895

Published: Sept. 10, 2024

The capacity of carbon sinks varies among the different types ecosystems, and whether national parks, as an important type nature reserve, have a high sink (CSC) eco-tourism in parks affects their CSC are main scientific issues discussed. Using MODIS Net Primary Production (NPP) product data, this study analysed spatiotemporal variation sources (CSSs) ecosystem Huangshan National Park from 2000 to 2020, well impact tourism on these sinks. findings indicate that, while ecosystems generally strong CSC, they may not always function sinks, during period, served source for four years. Temporally, CSSs park exhibit cyclical pattern change with four-year cycle seasonality, spring autumn functioning summer winter sources. Spatially, exhibited vertical band spectrum spatial distribution, showed trend gradual enhancement low altitude altitude. Tourism is major factor that has ecosystems.

Language: Английский

Citations

0
Development of a regional carbon assimilation system and its application for estimating fossil fuel carbon emissions in the Yangtze River Delta, China DOI

Zhengqi Zhang,

Shuzhuang Feng, Y. Chen

et al.

The Science of The Total Environment, Journal Year: 2024, Volume and Issue: 957, P. 177720 - 177720

Published: Nov. 27, 2024

Language: Английский

Citations

0
Exploring the Potential of Long Short-Term Memory Networks for Predicting Net CO2 Exchange Across Various Ecosystems With Multi-Source Data DOI Open Access

Chengcheng Huang,

Wei He, Jinxiu Liu

et al.

Authorea (Authorea), Journal Year: 2023, Volume and Issue: unknown

Published: Nov. 14, 2023

Upscaling flux tower measurements based on machine learning (ML) algorithms is an essential approach for large-scale net ecosystem CO2 exchange (NEE) estimation, but existing ML upscaling methods face some challenges, particularly in capturing NEE interannual variations (IAVs) that may relate to lagged effects. With the capacity of characterizing temporal memory effects, Long Short-Term Memory (LSTM) networks are expected help solve this problem. Here we explored potential LSTM predicting across various ecosystems using data over 82 sites North America. The model with differentiated plant function types (PFTs) demonstrates capability explain 79.19% (R2 = 0.79) monthly within testing set, RMSE and MAE values 0.89 0.57 g C m-2 d-1 respectively (r 0.89, p < 0.001). Moreover, performed robustly cross-site variability, 67.19% can be predicted by both models without distinguished PFTs showing improved predictive ability. Most importantly, IAV highly correlated observations 0.81, 0.001), clearly outperforming random forest -0.21, 0.011). Among all nine PFTs, solar-induced chlorophyll fluorescence, downward shortwave radiation, leaf area index most important variables explaining variations, collectively accounting approximately 54.01% total. This study highlights great improving carbon multi-source remote sensing data.

Language: Английский

Citations

0