The first long-term record of CO2 from ATTO

High-quality atmospheric CO2 measurements are sparse across the Amazon rainforest. Yet they are important to better understand the variability of sources and sinks of CO2. And indeed, one of the reasons ATTO was built was to obtain long-term measurements in such a critical region.

ATTO scientists first installed measurement systems for CO2 in March 2012 at the 81 m walk-up tower at ATTO. Since then, inlets continuously collect air at five heights (4 m, 24 m, 38 m 53 m and 79 m). Therefore, they measure at the ground, within the forest canopy, and above it. Santiago Botía and his colleagues now published the first 6 years of continuous, high-precision measurements of atmospheric CO2 at ATTO. Based on this 6-year long record, they were to gain insights into some of the processes that affect trace gases in the Amazon rainforest.

Inlets for measuring CO2 installed at ATTO. © Santiago Botía / MPI-BGC.

For their study, they looked at half-hour data during daytime, when the atmosphere is generally well mixed. At nighttime, there is often intense atmospheric stratification. This means that local signals (like from small stream valleys or even single trees) could easily dominate the measured signal. That would make it impossible to learn more about larger-scale processes.

Seasonal and inter-annual variability of CO2

First of all, Santiago Botía and his co-authors found that the CO2 content in the atmosphere increased at a rate very similar to the global increase. However, this trend is not linear but instead characterized by a pronounced seasonal cycle. The CO2 concentration is highest in the transition from wet to dry season (in June and July). It is lowest during two times of the year: the beginning of the wet season (March) and during the dry season (from Aug to Oct). One reason for this is that many trees put out new leaves during the dry season. These young leaves are more efficient in capturing CO2. In addition, there are typically fewer clouds and more sunshine during this time of year. Thus, plants can take up more carbon from the atmosphere via photosynthesis.
The ATTO CO2-record (black line) shows an increase over time, but with a pronounced seasonal cycle. Figure from Botía et al. (2021)
The researchers also found some inter-annual patterns in the record, i.e. times where the CO2 record diverged from the just described seasonal pattern. The most pronounced event between 2012 and 2019 is the El Nino of 2015-16. This global weather phenomenon caused a severe drought and extreme heat in the Amazon region. This affected the CO2 concentrations at ATTO, but in variable, erratic ways. During the 2015 wet season, there was a positive anomaly (values above what would be expected), but in the following dry season, there was a negative anomaly (values below what would be expected). After analyzing all available data, the scientists came to the conclusion that this cannot be explained solely by local effects. Interestingly, the next year (2016) there were two positive anomalies in both dry and wet seasons, which clearly were driven by local and non-local effects. The authors concluded that the difference in 2015 and 2016 implies that the El Nino began affecting the Amazon rainforest at different times across the region.

Comparing data to model simulations

The same is true not only for this specific event but for the entire record. After the team analyzed the data in-depth, they compared them to model simulations. They wanted to see if they could reproduce the measurements with atmospheric models. However, even though they used optimized model output for the vegetation signal in their atmospheric model, it wasn’t possible to capture the seasonal cycle of CO2 that we can observe in the data. This suggests that the vegetation models do not yet capture all relevant processes to realistically predict trace gas processes. One of the things that contribute to this discrepancy, in this case, is that the model does not include either the tree-leaf cycle or the CO2 signals emitted from rivers. Both of them seem to be important to interpret the ATTO CO2 record.
Shown here in the black line is the observed CO2 record from ATTO. The colored lines show different model simulations that do not capture the seasonal cycle well. Figure from Botía et al. (2021).
But beyond that, the team came to the conclusion both local and regional signals significantly affect the trace gas data measured at ATTO. They have shown that we need to better understand the underlying processes and seasonal cycles that influence greenhouse gas emissions and uptake. Only then can we include them in model simulations to make more accurate predictions about the future of the Amazon Rainforest. In a next step, Santiago Botía and his colleagues will use the ATTO CO2 record together with other data streams to improve the vegetation models and look at the vegetation response during the last decade. The authors published the study “The CO2 record at the Amazon Tall Tower Observatory: A new opportunity to study processes on seasonal and inter-annual scales” Open Access in the journal Global Change Biology.

Similar articles

GRAEGOR detector unit for measuring trace gases like nitrogen in the laboratory container. © Robbie Ramsay / University of Edinburgh

Ramsay et al. developed a new model to assess nitrogen exchange between atmosphere and biosphere based on observations at ATTO. This model includes parameters controlling both nitrogen deposition and emissions in tropical forests.

GRAEGOR detector unit for measuring trace gases like nitrogen in the laboratory container. © Robbie Ramsay / University of Edinburgh

Ramsay et al. measured inorganic trace gases such as ammonia and nitric acid and aerosols in the dry season at ATTO. They are to serve as baseline values for their concentration and fluxes in the atmosphere and are a first step in deciphering exchange processes of inorganic trace gases between the Amazon rainforest and the atmosphere.

Pfannerstill et al. compared VOC emissions at ATTO between a normal year and one characterized by a strong El Nino with severe droughts in the Amazon. The did not find large differences, except in the time of day that the plants release the VOCs. They published their results in the journal Frontiers in Forest and Global Change.