**1. Introduction**

Discussions on climate change have become increasingly more relevant in the general scientific community, particularly since the creation of the Intergovernmental Panel on Climate Change (IPCC), which is composed of a diverse group of worldwide researchers, focusing on climate change studies and its impacts on society. Since the industrial revolution, the concentration of carbon dioxide (CO2) in the atmosphere has increased with the use of fossil fuels, deforestation, the use of nitrogen in agriculture and livestock farming, which are reported to be the main uses responsible for the anthropic greenhouse effect. Studies have already shown the relevance of biosphere–atmosphere interactions in Brazilian biomes regarding planetary climate regulation due, for example, to water, energy, and carbon exchanges with the atmosphere [1–6]. However, there are still uncertainties regarding these processes due to the remarkable diversity of physiognomies, landscapes and other biophysical aspects that might play a role in differentiating atmospheric patterns from one place to another within each biome.

To reduce these uncertainties, in situ measurements are needed to better understand the particularities of each environment. Furthermore, these observed data can also be used to assess soil–vegetation–atmosphere models [7–10] and to analyze satellite-derived estimates of water and/or CO2 balance components [11–16]. Both these models and remote sensing data are extremely important in providing reliable information on CO2 exchanges over tropical forests where flux tower coverage is scarce or non-existent, such as in many parts of Brazil.

Additionally, certain forest physiognomies are not endemic to Brazil, but occur in several other regions of the globe, and therefore, their particularities regarding biophysical patterns need to be understood in detail. Wetlands in tropical rainforests such as the Amazon are environments where organic production rates are high and anoxic conditions are frequent, and therefore, they represent crucial zones for the global balance of greenhouse gases in the atmosphere. Cerrado ecosystems (savannas), however, are located at the tropics and subtropics and are characterized by marked wet and dry seasons. They cover approximately 60%, 50% and 45% of the total area of Africa, Australia and South America, respectively [17]. These ecosystems play an important role in the cycle of several greenhouse gasses, as reported by the studies [18–20] among others, and on energy fluxes (latent and sensible heat) in the context of climate change, since they cover approximately 20% (2.7 billion ha) of the global surface [17].

Carbon dynamics and biophysical evapotranspiration controls over tropical Brazilian ecosystems and in other regions of the world have also been studied because of the need to better understand the effects of land use change on regional and global biogeochemical cycles [13,15,21]. Evidently, it also allows us to estimate the contribution of tropical biomes to the regional and global climate controls, also comprising the relationship between precipitation, evapotranspiration, plant productivity and greenhouse gas exchanges within the atmosphere [22]. These changes motivated the development of various research fields within climate sciences focusing on the debate over greenhouse gases cycles and anthropic influence over these cycles. CO2 concentration in the atmosphere surpassed 405 ppm [23] and continues increasing each decade, as well as that of other greenhouse gases. Therefore, it is necessary to develop methods and techniques to quantify these emissions and the fluxes of these gases, improving the understanding on how different environments function and respond to land-use changes.

The eddy covariance method (EC) directly and non-intrusively estimates the vertical transport of CO2 and other greenhouse gases. Information measured through EC allows for the better comprehension of biogeochemical cycles, energy fluxes and other atmospheric controls. It also provides clues to other pivotal questions regarding ecosystem (in this study, tropical ecosystems) controls over climate. However, due to the lack of an extensive EC measurement network throughout Brazil, other data sources are required, such as satellite-derived data. Our hypothesis is that both gross primary production (GPP) and water-use efficiency (WUE) data respond differently to varied water availability conditions on the main tropical ecosystems in Brazil. Thus, our objective is to analyze the dynamics of GPP and WUE in the Amazon, Cerrado (savanna), Pantanal and Caatinga biomes through data collected through EC systems and to assess the performance of data derived from the Moderate Resolution Imaging Spectroradiometer (MODIS) sensor for the quantification of CO2 balance components over these environments.

#### **2. Materials and Methods**

#### *2.1. Data Policy and Use License*

The Ameriflux platform compiles data monitored in three Brazilian biomes: BR-Sa1 (Amazon), BR-CST (Caatinga) and BR-Npw (Pantanal). Data from the Cerrado site (BR-BI, Bananal Island—Javaés) are available at https://daac.ornl.gov (accessed on 25 April 2022). Data provided by the ORNL DAAC can be accessed for free, without restrictions and in accordance with NASA's Earth Science Program. Ameriflux data are shared under CC-BY-4.0 data usage license (Creative Commons by Attribution 4.0 International). This license states that the use of data is free to share (copy and redistribute the material in any medium or format) and/or adapt (remix, transform, and build upon the material). The scientific literature references describing each of the data sites are: BR-Sa1 [24], BR-CST [25] and BR-Npw [26].

#### *2.2. Description of Study Area*
