AmeriFlux-Loaned System Enables Critical Monitoring of Tropical Forest Response to Disturbance and Drought
By Natalia Restrepo-Coupe1,2, Scott R. Saleska1, Raimundo Cosme Oliveira, Jr.3, and Rodrigo da Silva4
1 Ecology & Evolutionary Biology, University of Arizona, Tucson, AZ, USA. 2 Cupoazu LLC, Etobicoke, ON, Canada. 3 Embrapa Amazônia Oriental, Santarém, PA, Brazil. 4 Federal University of West-Para (UFOPa), Santarém, PA, Brazil.
Since 2001, the K67 eddy covariance tower in the Tapajós National Forest (AmeriFlux site BR-Sa1, https://ameriflux.lbl.gov/sites/siteinfo/BR-Sa1) has been an important site for long-term measurements of carbon and water fluxes in an intact Amazonian rainforest (Figure 1). Through two decades of long-term operation, K67 has captured ecosystem-scale responses to multiple El Niño and La Niña events, providing a robust baseline for studying the sensitivity of tropical forests to interannual climate variability, with only one extensive gap (in 2006-2008, when a tree fall destroyed the tower during a lapse in funding). However, this record was disrupted again starting in December 2022, when severe thunderstorms caused windthrow disturbance and formation of a series of new gaps near the tower, while the tower and instrumentation suffered only moderate damage.
Figure 1. The Santarém K67 legacy eddy covariance system operated extensively between 2001 and 2021 in the Tapajós National Forest, capturing over two decades of carbon, water, and energy fluxes under varying climate conditions, including multiple El Niño and La Niña events. Tropical ecosystems remain critically underrepresented in a recent FLUXNET and AmeriFlux site collections (see map), despite their central role in global carbon cycling and climate regulation
Then on November 24, 2023, the tower was directly hit by a storm-induced tree fall which took out one of the guywires and effectively destroyed the structure, still in the midst of repair from the previous year’s event. This damage occurred during a significant upgrade effort, during which we had replaced aging infrastructure (UPS, power distribution, IRGAs, data loggers, and flow controllers) and on the morning of the day they tower fell, had deployed a new closed-path eddy covariance system (the Campbell Scientific CPEC) obtained through the AmeriFlux emergency loan program (Figure 2).
Figure 2. Left to right: Installation of the AmeriFlux loan closed-path eddy covariance system on November 24, 2023; tower collapse later the same day; recovery of instruments and tower segments using scaffold towers to access the collapsed section caught in the canopy, November 26, 2023.
Crucially, the AmeriFlux loaner system—which amazingly escaped undamaged by the towerfall—provided the basis for rapid re-deployment. To maintain a proxy for measurement continuity at the site, we quickly installed the eddy covariance system in January 2024 on a nearby 45-meter-tall walk-up tower, located approximately 500 meters from the original K67 triangular tower.
Figure 3 Installation of AmeriFlux (CPEC) system at the walkup tower January 2024. Clockwise from top left: diffuse/direct radiometer photosynthetic active radiation (PAR) sensors sensors, PhenoCam and rain gauge instruments at the walk-up tower. Sonic anemometer and infrared gas analyser Loggers for the AmeriFlux CPEC and working at the system
Monitoring Two Contrasting Forest States
The key advantage of the AmeriFlux-loaned system is that it enabled a rapid redeployment and monitoring of fluxes over the relatively intact forest canopy of the walkup tower—structurally more similar to the forest in the footprint of the main tower prior to the recent windthrow events observations.
By June of 2024 we were able to rebuild the main triangular 65 m tall tower (the one destroyed by treefall) and install a new eddy covariance system (also a Campbell CPEC system). This dual-tower configuration now allows us to estimate how two adjacent but ecologically distinct patches of forest respond to the same climatic conditions, particularly during and after extreme events like drought (El Niño) and excess rain (La Niña conditions). (Though such comparisons are admittedly complicated by the differing heights of the sensor placement on the two towers, with the sensors on the lower walkup tower below the height of nearby emergent trees, likely not quite high enough to be in a constant flux layer; comparisons over the course of a full year between the two eddy flux systems should nonetheless allow us to identify atmospheric conditions to investigate such differences).
Figure 4. Left image: View of K67 tower from walk-up tower: Measures fluxes over a disturbed canopy, affected by recent windthrow, with increased gaps, fallen biomass, and changes in surface reflectance. Center: View of the canopy from the triangular tower. Right: View of the walk-up from the K67 triangular tower: Where AmeriFlux-Assisted Walk-Up Tower: Captures fluxes over the currently comparatively undisturbed canopy, offering a reference point for comparison in this moment.
2024 Drought Observations and Historical Context
Initial observations from 2024 reveal diverging responses between the two tower sites—Walkup (recently less disturbed forest) and triangular (recently disturbed forest). At the disturbed tower, we observed a slight increase in latent heat flux (LE)—indicating higher evapotranspiration—and a significant decrease in sensible heat flux (H). Additionally, there was an increase in albedo, suggesting the forest is reflecting more incoming solar radiation —brighter surface and less vegetation/photosynthetic activity(?). These patterns are consistent with shifts in the forest’s carbon, energy and water balance, likely influenced by structural and compositional changes following gap formation and past disturbances (Figure 5).
Figure 5. BR-Sa1, Tapajos National forest K67 seasonal 16-day average of daytime latent heat flux (LEdaytime; W m-2) and sensible heat flux (Hdaytime; W m-2) from eddy covariance observations and net radiation (Net Radiationdaytime; W m-2) from meteorological instrumentation. Years 2024, 2025 and long-term measurements before 2020 (average, thick black line) and El Niño 2015 observations (red line).
At the Tapajós National Forest, El Niño-induced droughts can trigger “green-up” responses, as observed in 2002, with increased photosynthetic activity and water use. In contrast, the 2015 El Niño—amplified by a warming trend and elevated vapor pressure deficit (VPD)—resulted in decreased stomatal conductance (Gs), reduced leaf area index (LAI), and increased sensible heat flux (H), indicating a shift in forest function (Figure 6).
By contrast, La Niña wet periods are typically characterized by increased moisture availability and reduced atmospheric stress. For example, during the abnormally wet 2008 La Niña, higher transpiration (T) and lower evaporation (E) were observed, with little change in overall seasonal evapotranspiration (ET or its energy equivalent, LE). Although the components of water and energy flux shifted, the total flux remained balanced. Despite reduced photosynthesis—likely due to lower incoming radiation from increased cloud cover—forest net carbon loss occurred, as ecosystem respiration (Reco) remained near seasonal mean levels. Our 2024 observations align with a La Niña event (June 2024 to present), which quickly followed the June 2023 to March 2024 El Niño. The rapid succession of drought and wet periods, combined with pronounced changes in forest structure and composition—including legacy effects from mortality during the 2015 El Niño—complicates efforts to identify the mechanisms driving the ecosystem’s response.
Figure 6. BR-Sa1, Tapajos K67 seasonal 16-day values of average long-term time series of daytime sensible heat flux (Hday; W m-2) (red lines) from eddy covariance observations and outgoing longwave radiation from the CERES satellite product (LWout CERES W m-2).
By comparing current observations to historical datasets, we are now able to assess whether canopy disturbance (such as windthrow) compounds the effects of climatic stress, potentially pushing forest ecosystems closer to critical functional thresholds.
Implications of the Dual-Tower Setup
The AmeriFlux loaned instrumentation was instrumental not only in avoiding a data gap but also in enabling a potentially valuable comparative scientific study. With both towers now operational, we can assess (adjusting for the different mounting strategies on the two towers):
- How forest structure (gaps vs. closed canopy) affects flux partitioning (LE vs. H)
- Whether disturbed forest patches respond differently to seasonal VPD and drought stress
- How energy and water budgets evolve in structurally heterogeneous landscapes
As tropical forests face increasingly frequent and intense disturbances, this kind of dual-site monitoring is essential for refining our understanding of ecosystem resilience, informing land management, and improving predictions in Earth system models.
We acknowledge our primary funding agencies, (1) in Brazil: the federal Ministry of Science and Technology (including via the LBA program phase II, and projects of the Conselho Nacional de Desenvolvimento Científico e Tecnológico, CNPq), EMBRAPA, and the Brazilian state science foundations of Sao Paulo (FAPESP), Amazonas (FAPEAM) and Para (FAPESPA); and (2) in the U.S.: the National Aeronautics and Space Administration (NASA), the U.S. Department of Energy, and, in particular, the U.S. National Science Foundation (NSF) (whose 2023 RAPID grant # 2403882 enabled fast tower rebuild and re-instrumentation). We also thank the Brazilian-led Large-Scale Biosphere-Atmosphere (LBA) offices in Manaus and Santarém for their continued support, as well as the AmeriFlux team for their collaboration in providing the eddy covariance instrumentation loan and facilitating data sharing efforts.
No Comments
Be the first to start a conversation