Text:
Forest automation and the spatialization of carbon stocks in native areas in the Western Amazon are part of a study by the Geoflora Project, of the Brazilian Agricultural Research Corporation - Embrapa, in the state of Acre, in partnership with an investment fund.
The use of technology to collect data on the ground and mapping the forest using drones with artificial intelligence are part of the research.
The results will support public policies for sustainable use, such as forest management and payment for environmental services.

- audio track

"In this project we have two aspects. The first is on land, with measurements made in the field on the growth dynamics of managed forests," said researcher Marcus Vinicio Neves d'Oliveira, from the Brazilian Agricultural Research Corporation (Embrapa-Acre).

"In the case of the Geoflora project, we do this in two environments, checking what changes occur in these flows according to land use," commented Falberni de Souza Costa, a researcher at Embrapa-Acre.

"Here, in this type of forest, we take measurements in a primary forest, which is the one we are in, and also in a forest that was cut down and burned for research purposes in 2011," said Souza Costa.

- audio track

"These collections throughout the year, at different times, we calculate in aggregate terms how much the system as a whole is emitting or capturing carbon or CO2 equivalent. What is carbon or CO2 equivalent? I use the global warming potentials, both for methane and nitrous oxide, and I refer to CO2. So I have a single unit with which I can work and identify, qualify the land use systems," explained Costa.

"These areas serve as the basis for the second part, the second aspect, which is the part remote censoring. In this second aspect, we scale these estimates, and to do this we use LIDAR (Light Detection And Ranging), which is a laser-based system that allows us to capture height metrics that we correlate with data from permanent plots and transform into estimates that we later extrapolate to the entire area covered by the LIDAR flight, explained researcher Marcus Vinicio, from Embrapa.

"Collecting photos with the drone is a key step in this process, as good field collection will result in better quality photos. Always remember to pay attention to lighting and climate conditions, and to the overlapping of photos to generate a perfect orthomosaic to feed the algorithm and, consequently, perform high-quality canopy identification," explained Embrapa forestry engineer Caio Alexandre Nascimento Santos.

"This way, without having to measure an entire area, we have estimates about the structure of this forest and about the biomass, or carbon, or volume of wood of the entire area covered by LIDAR (Light Detection And Ranging). In the third stage, we scale these areas covered by LIDAR to large areas, for example, our final result will be a map of the distribution of biomass throughout the state of Acre. So, at this point, we create a new model correlating that LIDAR model with the model with orbital images, thus producing a biomass model that can be extrapolated to the entire state", said Neves d'Oliveira, researcher.

- audio track

"In the municipality of Sena Madureira, in the Antemari State Forest, at the State Government's Environmental Management Unit, we take measurements and monitor measurements over time in five environments. The native forest, which is always the reference, the yard, the forest road, the trail and the clearing. One day before the collection, we install the chambers. So, this is the apparatus we use to accumulate the gas that is coming out of the soil, collect it in a pre-evacuated bottle, which will then undergo chromatographic analysis at Embrapa Meio Ambiente, there in São Paulo", said researcher Falberni de Souza.

- audio track

We are now carrying out this survey at Flora Jamari, where there is already a lot of LIDAR (Light Detection And Ranging) information. And there we are going to develop a fourth model, which I haven't mentioned yet, which is an estimated forest production model. How are these forests growing, how are they reacting to forest light, but this time not just based on permanent plots, but on the 'LIDAR' coverages that they already have and that we are going to do there", said Neves d'Oliveira.

- audio track

"The Brazilian Forest Service, which is our partner in this endeavor, has LIDAR flights in all national forests. So, we will use this data, we will add some other data, both from land, from permanent plots, and from flights that we will carry out there, and we will then create this forest production model. We will have a much more realistic estimate of how these managed forests are growing.

Forest growth is the plots. These plots are very expensive to install and measure. It is a very large task and also requires a lot of time. What we are trying to do is scale these estimates, from the small area of ​​the plots to larger areas, which are covered, in this case, by LIDAR (Light Detection And Ranging). With LIDAR, we cover areas of thousands of hectares, and then we have estimates that better cover the variability of the forest within these areas. And, with this, we have much more precise estimates of how the forest is growing. reaction of these forests.

The impact produced by forest management. On top of this impact, we have the verification of atypical climate events, which have greatly influenced the growth of these forests, producing mortality of large trees, for example, not connected to forest management. This mortality occurs due to the recurrence of these atypical climate events, and it has greatly delayed the recovery of the forest. So, this is a piece of data that we also seek to bring, to include in this analysis of the cycles and intensities of cutting that are linked to forest management', concluded researcher Marcus Vinicio Neves d'Oliveira, Embrapa.

- final audio track

Interviews:
Marcus Vinicio Neves d'Oliveira, researcher at Embrapa-Acre

Falberni de Souza Costa, researcher at Embrapa-Acre

Caio Alexandre Nascimento Santos - forestry engineer at Embrapa-Acre

Acknowledgements:
Embrapa - Brazilian Agricultural Research Corporation

Source: Amazon Agency