Pedro Maia Barbosa

Factors leading to exceptionally high temperatures in shallow tropical waters are examined with high‐resolution measurements and application of a 3‐dimensional hydrodynamic model in an Amazon floodplain lake. Highest near‐surface (0.05 m) temperatures in open water ranged from 31.0 to 37.5°C. Within mats of floating plants maximum temperatures ranged from 33.7 to 39.8°C, and within flooded forests from 30.6 to 34.1°C. Highest open water temperatures occurred when water was shallow (ca. 1 m), underwater attenuation was high (5.5–7.5 m ⁻¹ ), skies were clear (solar radiation ca. 1000 W m ⁻² ), winds were low (≤ 1 to 2–3 m s ⁻¹ ), and air temperatures were in the mid‐30°C range. Heat budgets indicate that latent heat fluxes ranged from ~ −100 W m ⁻² under light winds to −200 to −250 W m ⁻² with 4 m s ⁻¹ winds. Sensible heat flux was usually negative, as lake temperatures exceeded air temperatures, and much less than latent heat flux. Hydrodynamic simulations quantified how interactions among drivers lead to exceptional warming: reduced wind speeds (≤ 30% of reference) increased near‐surface temperature by 3–4°C, while elevated air temperatures (up to 37.5°C) and high turbidity contributed an additional 1–2°C. The simulations indicate that exceptional water temperatures (up to 42°C) can occur when a shallow, turbid lake under clear skies with high air temperatures and low wind speeds simultaneously maximize heat gain and minimize heat loss. Projections of climatic changes are likely to increase incidences of exceptional water temperatures.

Reservoirs are significant sources of greenhouse gases (GHG), but the magnitude of these emissions remains uncertain. The key to deriving robust diffusive fluxes is combining accurate surface water gas concentrations with realistic gas transfer velocities ( k ₆₀₀ ). Existing models, primarily based on wind speed, estimate system‐wide k ₆₀₀ values and typically do not transpose well between systems, particularly for reservoirs with complex shapes. This study aims to improve current models to estimate k ₆₀₀ in such reservoirs. We focus on the La Romaine complex, composed of three cascading reservoirs (RO1, RO2, RO3) in northeastern Canada. We measured 535 carbon dioxide (CO ₂ ) and methane (CH ₄ ) fluxes using floating chambers during 11 field campaigns and calculated k ₆₀₀ from these fluxes. Besides wind speed, within‐reservoir location related to depth and distance to shore strongly modulated k _600, but this location effect was reservoir‐dependent. Based on these three variables, we generated reservoir‐specific spatially explicit k ₆₀₀ models that account for the heterogeneity in fetch within each reservoir. For RO1, our generated model explained up to 110% more of the variance in measured k ₆₀₀ compared to any published system‐wide k ₆₀₀ model to date, and for RO2 and RO3, up to 46% and 28%, respectively. We show that the same modeling framework can be applied to other reservoirs with complex shapes. Understanding how location within the reservoir modulates the influence of wind on k ₆₀₀ will help constrain current reservoir GHG fluxes uncertainties.

Methane oxidation has been observed in a wide range of aquatic environments worldwide, and measurements are rare in tropical floodplains. The Amazon floodplain is one of the largest tropical wetlands with seasonally flooded forests representing up to 80% of the area of aquatic habitats in the lowland Amazon. Hence, we measured methane oxidation rates (Mox) in two different flooded forests ( várzea , in white waters; igapó , in black waters) and evaluated effects of dissolved oxygen and CH ₄ concentrations, and water temperature on methane oxidation. We found high Mox in near-bottom waters associated with high CH ₄ concentrations (1.0–2.4 µM) and hypoxia, with volumetric rates ranging from 9.8 to 73 mg C m ⁻³ d ⁻¹ in the igapó , and from 2.3 to 101.4 mg C m ⁻³ d ⁻¹ in the várzea . Depth integrated Mox rates ranged from 177 to 213 mg C m ⁻² d ⁻¹ for the igapó , and 159 mg C m ⁻² d ⁻¹ in the várzea , and were one to two orders of magnitude higher than CH ₄ fluxes from water to the atmosphere, emphasizing the important role of Mox in attenuating CH ₄ emissions from tropical flooded forests. The present study contributes to understanding of the complex processes involved in carbon dynamics on tropical floodplains.

Floodplain lakes are widespread and ecologically important throughout tropical river systems, however data are rare that describe how temporal variations in hydrological, meteorological and optical conditions moderate stratification and mixing in these shallow lakes. Using time series measurements of meteorology and water‐column temperatures from 17 several day campaigns spanning two hydrological years in a representative Amazon floodplain lake, we calculated surface energy fluxes and thermal stratification, and applied and evaluated a 3‐dimensional hydrodynamic model. The model successfully simulated diel cycles in thermal structure characterized by buoyancy frequency, depth of the actively mixing layer, and other terms associated with the surface energy budget. Diurnal heating with strong stratification and nocturnal mixing were common; despite considerable heat loss at night, the strong stratification during the day meant that mixing only infrequently extended to the bottom at night. Simulations indicated that the diurnal thermocline up and downwelled creating lake‐wide differences in near‐surface temperatures and mixing depths. Infrequent full mixing creates conditions conducive to anoxia in these shallow lakes given their warm temperatures.

Floodplains lakes are abundant in the Amazon basin and are important methane sources to the atmosphere. Existing biogeochemical models require modifications and inclusion of hydrodynamic processes operative in shallow, warm waters to be applied to these aquatic ecosystems. We modified a 1‐dimensional process‐based, lake biogeochemical model and combined a 3‐dimensional hydrodynamic model to suit Amazon floodplains. We evaluated the combined model's performance simulating methane concentrations and fluxes and several related processes in the open lake and an embayment of a well‐studied Amazon lake. Parameters for calibration were selected through sensitivity tests using a machine learning‐based algorithm, classification, and regression trees. Comparison between simulated and measured fluxes indicate generally good agreement in seasonal patterns and magnitudes. Comparisons of near‐surface concentrations varied with no clear patterns. Simulations of methane concentrations at near‐surface and near‐bottom, and diffusive emissions are most sensitive to carbon mineralization rate, Q₁₀ factors for methanogenesis and oxidation, and methane oxidation potential. Modeled rates of planktonic photosynthesis were generally lower than measurements, though simulated planktonic respiration was often similar to measurements. Simulated rates of methane oxidation were considerably lower, with a few exceptions, than measurements of methane oxidation in oxic water of the lake. Improvements of results of the linked hydrodynamic‐biogeochemical model will result from inclusion of advective transport, use of parameter values appropriate for tropical waters, especially for methane oxidation and photosynthesis, and addition of changes in hydrostatic pressure to model of ebullition.

The impoundment of rivers for multipurpose reservoirs has significant consequences to the carbon cycle, one of the most relevant being the increase in greenhouse gases emissions. Reservoirs have been shown to be net sources of such gases to the atmosphere, emitting between 0.8 and 1.08 Pg carbon dioxide (CO₂) equivalents per year. Even though emission estimates have become common, less is known about the processes driving this CO₂ excess, a prerequisite for understanding and ultimately predicting and managing CO₂ emissions from reservoirs. In the present study, we aimed at exploring ecosystem metabolism (gross primary production, ecosystem respiration, and net ecosystem production [NEP]) and its environmental drivers in three young hydroelectric reservoirs in a cascade configuration but with distinctive morphometries (shape, depth, and size). By combining our metabolic measurements with a hydrological mass balance approach, we quantified the relative contributions of internal (ecosystem metabolism) versus external sources (tributaries and groundwater) to the reservoir surface diffusive CO₂ emissions. There was a predominance of net heterotrophy in all sites, and metabolism played a key role in fueling CO₂ fluxes in all three reservoirs, NEP alone being able to account for the measured fluxes in approximately 50% of all sites. Internal production was thus the main process explaining total reservoir CO₂ diffusive emissions (∼100%), with groundwater and tributaries contributing similarly but more modestly (∼9% each). Our results contribute to our understanding of the processes underlying boreal reservoir carbon footprints, and in particular, for apportioning the emissions can be attributed to the reservoir itself.

Tropical floodplains are an important source of methane (CH₄) to the atmosphere, and ebullitive fluxes are likely to be important. We report direct measurements of CH₄ ebullition in common habitats on the Amazon floodplain over two years based on floating chambers that allowed detection of bubbles, and submerged bubble traps. Ebullition was highly variable in space and time. Of the 840 floating chamber measurements (equivalent to 8,690 min of 10‐min deployments), 22% captured bubbles. Ebullitive CH₄ fluxes, measured using bubble traps deployed for a total of approximately 230 days, ranged from 0 to 109 mmol CH₄ m⁻² d⁻¹, with a mean of 4.4 mmol CH₄ m⁻² d⁻¹. During falling water, a hydroacoustic echosounder detected bubbles in 24% of the 70‐m segments over 34 km. Ebullitive flux increased as the water level fell faster during falling water periods. In flooded forests, highest ebullitive fluxes occurred during falling water, while in open water and herbaceous plant habitats, higher ebullitive fluxes were measured during low water periods. The contribution of diffusive plus ebullitive CH₄ flux represented by ebullition varied from 1% (high and rising water in open water of the lake) to 93% (falling water in flooded forests) based on bubble traps. Combining ebullitive and diffusive fluxes among habitats in relation to variations in water depth and areal coverage of aquatic habitats provides the basis for improved floodplain‐wide estimates of CH₄ evasion.

Inundated tropical forests are underrepresented in analyses of the global carbon cycle and constitute 80% of the surface area of aquatic environments in the lowland Amazon basin. Diel variations in CO₂ concentrations and exchanges with the atmosphere were investigated from August 2014 to September 2016 in two flooded forests sites with different wind exposure within the central Amazon floodplain (3°23′S, 60°18′W). CO₂ profiles and estimates of air–water gas exchange were combined with ancillary environmental measurements. Surface CO₂ concentrations ranged from 19 to 329 μM, CO₂ fluxes ranged from −0.8 to 55 mmol m⁻² hr⁻¹ and gas transfer velocities ranged from 0.2 to 17 cm hr⁻¹. CO₂ concentrations and fluxes were highest during the high water period. CO₂ fluxes were three times higher at a site with more wind exposure (WE) compared to one with less exposure (WP). Emissions were higher at the WP site during the day, whereas they were higher at night at the WE site due to vertical mixing. CO₂ concentrations and fluxes were lower at the W P site following an extended period of exceptionally low water. The CO₂ flux from the water in the flooded forest was about half of the net primary production of the forest estimated from the literature. Mean daily fluxes measured in our study (182 ± 247 mmol m⁻²d⁻¹) are higher than or similar to the few other measurements in waters within tropical and subtropical flooded forests and highlight the importance of flooded forests in carbon budgets.

Seasonally flooded forests along tropical rivers cover extensive areas, yet the processes driving air‐water exchanges of radiatively active gases are uncertain. To quantify the controls on gas transfer velocities, we combined measurements of water‐column temperature, meteorology in the forest and adjacent open water, turbulence with an acoustic Doppler velocimeter, gas concentrations, and fluxes with floating chambers. Under cooling, measured turbulence, quantified as the rate of dissipation of turbulent kinetic energy (ε), was similar to buoyancy flux computed from the surface energy budget, indicating convection dominated turbulence production. Under heating, turbulence was suppressed unless winds in the adjacent open water exceeded 1 m/s. Gas transfer velocities obtained from chamber measurements ranged from 1 to 5 cm/hr and were similar to or slightly less than predicted using a turbulence‐based surface renewal model computed with measured ε and ε predicted from wind and cooling.