The spatial and environmental distribution of the global eddy covariance tower network

International Journal of Biometeorology · 2026-03-30

Eddy covariance has revolutionized our understanding of ecosystem-atmosphere interactions. Multiple studies have characterized the climate space occupied by flux tower networks, but none to our knowledge have characterized if eddy covariance sites represent the global distribution of soil characteristics that are critical for determining ecosystem function or studied the distances between towers to apply ‘paired’ tower studies. Of 1233 global eddy covariance towers explored here, half had a nearest neighbor within 10 km. Soil database pixels with towers have nearly 20% more silt and 8% less sand than the global soil texture distribution, with more soil N (0.58 g/kg vs. 0.38 g/kg) and organic C (8.3 g/kg vs. 5.4 g/kg), and 10% greater cation exchange capacity in upper layers than pixels without towers. Global syntheses of eddy covariance towers should be cognizant that tower networks capture more fertile soils than the terrestrial surface on average. A logical way to improve global representativeness is to further build collaborations and invest in underrepresented regions.

Chlorophyll-based canopy photosynthesis model: Development and global synergy analysis

Remote Sensing of Environment · 2026-05-08

Benchmarking a long short-term memory model against a process-based model for peatland water level dynamics

2026-03-14

Peatlands play a critical role in the global carbon cycle, with water level dynamics strongly controlling their function as carbon sinks or sources. While process-based models are commonly used to simulate peatland hydrology, the potential of data-driven approaches remains largely unexplored at large spatial scales.Here, we assess the capability of a Long Short-Term Memory (LSTM) model to simulate daily water level in natural northern peatlands (40°N–75°N), trained on a diverse set of in situ water level observations. Model performance is evaluated against the same in situ water level observations using a strict block-wise cross-testing scheme. Furthermore, model performance is benchmarked against simulations from NASA’s Catchment Land Surface Model with peatland modules (PEATCLSM).The LSTM model demonstrates improved agreement with in situ water level observations compared to PEATCLSM in terms of root mean square difference and bias, while the PEATCLSM exhibits higher spatial and temporal correlation with the in situ observations. Feature importance analysis indicates that the LSTM model captures key hydrological controls on water level dynamics, with precipitation and reference evapotranspiration emerging as dominant drivers, followed by leaf area index and snow water equivalent.The lack of sufficient in situ water level observations for model training, both in terms of record length and spatial coverage across peatland sites, restricts the development of a model with additional input variables that could enhance performance. Despite these limitations, the LSTM model shows spatial patterns consistent with the process-based model, supporting its reliability. These findings highlight the potential of deep learning approaches such as LSTM-based modeling to complement traditional process-based modeling of peatland hydrology. Future improvements will depend on collaborative data sharing to enhance training datasets and support informed climate and environmental decisions.

Ecosystem-Scale Carbon Balance to Improve the Emission Factors for Acacia Plantations on Tropical Peatlands 

2026-03-14

Peatlands are among the most carbon‑rich terrestrial ecosystems and play a key role in the global carbon cycle. However, managed peatlands for agriculture and silviculture emits significant carbon. Southeast Asia hosts approximately one third of tropical peatlands with around half of it are managed for agriculture and silviculture to support economic and population growth.Substantial uncertainties remain in existing estimates with a large range. Partially, such uncertainties can be attributed to both limited field measurements from major land uses and also lack of direct measurements of carbon loss when using short‑term chamber and subsidence approaches. Despite strong interests from scientific community and policy makers, current Intergovernmental Panel on Climate Change (IPCC) Tier 1 emission factors (EFs) for tropical peatlands are derived from short‑term chamber and subsidence measurements, which may not fully capture total ecosystem carbon dynamics and introduce potential uncertainty into emission estimates.Using continuous 30-minutes eddy covariance measurements, we quantified comprehensive greenhouse gas (GHG) balance of an Acacia crassicarpa plantation on tropical peatland in Sumatra, Indonesia. Net ecosystem carbon dioxide (CO₂), methane (CH₄), and soil nitrous oxide (N₂O) exchange to estimate the GHG exchange.Considering carbon export from harvested wood over a complete plantation rotation as emissions, the Acacia plantation exhibited net CO2 emissions of 30.0 ± 4.6 tCO₂-eq ha⁻¹ yr⁻¹, approximately 50% lower than IPCC Tier 1 EFs. Emissions were also ~20% lower than degraded peatlands in the same landscape, and the partial use of harvested biomass for bioenergy potentially further reduces the plantation’s overall climate impact.These findings indicate that current emission factors by IPCC may not fully represent GHG dynamics in existing Acacia plantations on tropical peatlands. Incorporating ecosystem-scale observations and full plantation rotation assessment into Tier 3 EFs estimation improved the accuracy in GHG emissions from managed tropical peatland ecosystems.

Simultaneous Estimation of Soil Moisture and Soil Organic Matter from in situ Dielectric Measurements - Part 1: Optimal Estimation Strategy

SSRN Electronic Journal · 2026-01-01

Modeling Sub‐Grid Peatland Vegetation Dynamics in the ORCHIDEE‐PEAT Land Surface Model

Journal of Advances in Modeling Earth Systems · 2026-03-01

Abstract Peatlands store about one‐third of total global soil carbon. Vegetation composition strongly regulates peatland carbon dynamics. Global warming and climate‐driven ecohydrological changes are expected to alter peatland vegetation composition, necessitating accurate simulation of vegetation dynamics to predict future fate of peatland carbon. We incorporated six plant functional types (PFTs) into the ORCHIDEE‐PEAT model to represent bryophytes (mosses), C3 graminoids (sedges and grasses), boreal broadleaf deciduous shrubs, boreal needleleaf evergreen trees, tropical evergreen and raingreen (water‐driven deciduous) trees growing in peatlands. The introduction and elimination of each PFT in response to bioclimatic conditions, as well as sapling establishment, growth, mortality, and competition among PFTs, are explicitly modeled. Simulated vegetation distributions align well with site‐level observations from West Siberian wetlands, where extensive vegetation composition measurements are available for model evaluation. The model slightly overestimated gross primary productivity (GPP) across 60 sites. Evaluation using global satellite‐derived land cover, leaf area index and GPP data was encouraging, though challenges lie in the lack of observational data specific to peatlands. From 1901 to 2020, simulated tropical peatland vegetation composition remains relatively stable. In northern peatlands, as a result of warming and declining water table, bryophyte and C3 graminoid cover decrease by 0.2 (13%) and 0.1 (13%) million km 2 , respectively, while shrub and tree cover increase by 0.3 (75%) and 0.03 (2%) million km 2 , respectively. The impacts of these vegetation shift on peatland carbon balance can be explored in future studies using the model, which integrates peatland vegetation dynamics with peatland‐specific hydrology and carbon cycling.

Simultaneous Estimation of Soil Moisture and Soil Organic Matter from in situ Dielectric - Part 2: Application of Optimal Estimation and Machine Learning Approaches

SSRN Electronic Journal · 2026-01-01

It’s cool to be green

Proceedings of the National Academy of Sciences · 2025-09-29

Near Real-Time Estimation of Daytime and Nighttime Evapotranspiration Using GOES-R Observations and Machine Learning Models

2025-05-22 · 1 citations

Evapotranspiration (ET) is a critical component of the water cycle, influencing climate, agriculture, and water resource management. However, most satellite-derived ET products are limited to daily or coarser temporal resolutions, despite the strong diurnal variability of ET processes. Existing satellite-based ET retrievals are largely restricted to daytime conditions, when nighttime ET is a small but often non-trivial flux. In this study, we introduce the Advanced Baseline Imager Live Imaging of Vegetated Ecosystems ET (ALIVEET), a near real-time, five-minute ET estimation framework, leveraging geostationary satellite observations from the GOES-R Advanced Baseline Imager (ABI) and machine learning models under both clear and cloudy conditions. We test Gradient Boosting Regression (GBR) and Long Short-Term Memory (LSTM) models to assess their ability to estimate ET variations across the diurnal cycle. GBR captures daytime ET with an R2 of 0.74 (RMSE of 0.059 mm hh-1 equivalent to about 74 W m-2) while maintaining low computational cost. For nighttime ET, where R2 decreases by about 0.50 compared to daytime, LSTM models trained on time-series observations perform better, achieving an R² of 0.24 (RMSE of 0.014 mm hh-1) by leveraging temporal dependencies in land surface temperature (LST) and past ABI observations. Comparisons against daily ET estimates from the physically based ALEXI remote sensing model demonstrates good agreement but opportunities for improvement. This study demonstrates the potential of integrating machine learning with geostationary remote sensing to advance high-temporal-resolution ET estimation.

A cross-site comparison of ecosystem- and plot-scale methane fluxes from wetlands and uplands

2025-10-27

Abstract. Wetland and upland ecosystems play significant but opposing roles in the global methane (CH4) budget, acting as natural sources and sinks, respectively. Two of the most common approaches for measuring CH4 fluxes (FCH4) are chambers, which capture temporally intermittent, fine-scale spatial heterogeneity (ca. 1 m2), and eddy covariance (EC) towers, which cover a larger area (ca. 100–10000 m2) at a longer term. Although chamber and EC observations have been combined in various syntheses and databases to estimate CH4 budgets, a unified cross-site evaluation of FCH4 estimates at plot and ecosystem scales is lacking. As a first step toward a systematic spatiotemporal scaling of EC tower and chamber footprints, we quantified the differences between site-level aggregate FCH4 (EC vs chamber; ΔFCH4) from ten wetland and upland sites at half-hourly, hourly, daily, weekly, monthly, and annual timescales. We found that ecosystem-scale median FCH4 was consistently higher than plot-scale FCH4 at all temporal scales, with the smallest difference at daily timescale (multi-site median ΔFCH4: 1.36 nmol m-2 s-1; ~ 104 % higher ecosystem-scale than plot-scale FCH4) and largest at annual scales (2.58 nmol m-2 s-1; ~ 87 % higher ecosystem-scale than plot-scale FCH4). In general, the agreement between ecosystem- and plot-scale FCH4 decreased with finer temporal resolution (from Spearman ⍴ = 0.95 at annual scale to ⍴ = 0.65 at half-hourly scale), while ΔFCH4 variation was greatest at daily-to-annual scales. Key environmental predictors of ΔFCH4 included plot-scale spatial heterogeneity, dominant vegetation type, vapor pressure deficit, atmospheric pressure, and friction velocity at the daily and monthly scales. Wind direction was a significant predictor only at the monthly scale, suggesting EC footprint effects. These findings suggest accounting for variation in EC footprint extent, chamber measurement placement and artifacts is key to reconciling multi-scale FCH4 observations in diverse ecosystems and refining CH4 budgets.