Recently, a team of researchers from the University of Sao Paulo (USP) have worked with the Federal University of Amazonas (UFAM) to explore new ways to improve soil analysis. Their research Journal of Environmental Managementdemonstrates how advanced spectroscopy, including reflectance spectroscopy, can provide faster, cheaper, and more sustainable alternatives in traditional laboratory-based soil testing (1). Researchers tested the method on Amazon River Basin (ARB), one of the most difficult ecosystems to study.
ARBs are a large ecosystem that spans a significant portion of South America. Covering nine countries over 7 million square kilometres (Bolivia, Brazil, Colombia, Ecuador, Guyana, Peru, Suriname and Venezuela), the ARB is located east of the Andes Mountains (2). Most of the ARBs are rainforests, and contain more than 56% of all broad-leaf forests on the planet (2). What makes ARB an attractive area of the planet and animal life is the biodiversity of plants and animals that make it an attractive area of the planet that is attractive to researchers. It is estimated that over 30,000 plant species, 60 reptile species, 2,000 fish species and 35 mammal families will call the ARB home (2).
Reflection of sunsets through lagoons within the Amazon rainforest basin. The Amazon River Basin consists of countries such as Brazil, Bolivia, Colombia, Ecuador, Guyana, Suriname, Peru and Venezuela. | Image credit: ©SL -Photography -Stock.adobe.com
Apart from researching biology at ARB, researchers are interested in learning more about the soil composition of ARBs. Traditional wet experimental methods are limited by the intensive labor required to process rare facilities, high costs, and large data sets (1).
As part of the experimental procedure, researchers collected 211 soil samples from the ARB and were exposed to laboratory visible, near-infrared (NIR) and short-wave infrared (Vis-NIR-SWIR) spectroscopy. These spectra were then used to predict important soil attributes, including soil organic carbon (SOC) and particle size distributions (sand, silt, clay) (1). The researchers tested two computational approaches. The first was a cloud-based method using the Brazilian soil spectral services (Braspecs) platform. The second was the offline approach using the R programming language (1). Both relied on Cubist machine learning (ML) algorithms to model soil properties.
To assess the performance of both methods, researchers considered variables such as coefficients of determination (r²), mean absolute error (MAE), and route mean square error (RMSE). The results showed that both approaches provided reliable predictions, but offline models consistently outperformed cloud-based platforms in terms of accuracy (1). For example, when predicting clay content, offline models r0.85² (1) compared to the online model 0.70. Similarly, offline methods have been delivered for SOC prediction rThe cloud-based approach is 0.81 vs. 0.72 (1).
Comparing the two methods, the researchers concluded that cloud computing has several important advantages over other methods. Unlike offline models, cloud-based approaches do not require locally observed training data and are particularly practical for large real-time applications in remote environments (1). This flexibility is important for regions like Amazon where traditional methods are practical due to accessibility and resource limitations (1).
Another important finding in this study was the strong sensitivity of the NIR and SWIR spectral regions (350-2500 nm) to major soil properties such as clay minerals, iron oxides, and soil organic material composition. This spectral responsiveness increases the possibility of accurate, non-destructive soil characterization (1).
Accurate soil mapping is essential for monitoring carbon strains, managing agricultural practices, and supporting Amazon's conservation policies, which is the heart of global climate regulation (1). Researchers argue that integration of soil spectroscopy with cloud computing can help inform sustainable development policies, optimize land use, and improve environmental management strategies (1).
Ultimately, the synergistic use of spectroscopy and cloud platforms provides not only technical efficiency, but also sustainable solutions for managing one of the planet's most important natural resources. As Amazon faces increased environmental pressures, such innovations may prove crucial in balancing development and conservation (1).
reference
- Novice, JJM; Mello, BMD; Neves Jr., AF; et al. Online analysis of Amazon soils with reflectance spectroscopy and cloud computing can support policy and sustainable development. J. Environment. manager. 2025, 375124155. doi:10.1016/jjenvman.2025.124155
- The Amazon River Basin, an American state organization. oas.org. Available at https://www.oas.org/dsd/events/english/documents/osde_8amazon.pdf (accessed 2025-09-18).
