Machine learning in soil science for prediction and management of biological activity for sustainable land use

Aug 3, 2026

Upload by Oliver

Machine learning methods support prediction and management of soil biological activity to enable sustainable land use. A random forest model predicts the Respiration parameter from soil physical and chemical attributes collected across sites in Baltimore, Maryland, achieving around 70% accuracy. The study presents visualization of actual versus predicted values and analyzes prediction errors. Further model enhancement is discussed through genetic optimization of hyperparameters and incorporation of climate and historical land-use data.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages8

Reading Time20 min

AccessFree to download

Introduction

Materials and methods

Machine learning in soil science for prediction and management of biological activity for sustainable land use Svetlana Kukartseva1, Dmitry Evsyukov2, Vasiliy Orlov2*, Anatoly Kukartsev2,3, and Andrey Poddubny3 1Russian State Agrarian University-Timiryazev Moscow Agricultural Academy (RSAU-MAA named after K.A. Timiryazev), Moscow, Russia 2 Bauman Moscow State Technical University, Artificial Intelligence Technology Scientific and Education Center, Moscow, Russia 3 Reshetnev Siberian State of Science and Technology, Krasnoyarsk, Russia Abstract. The article discusses the use of machine learning methods for predicting and managing soil biological activity, which is a key aspect of sustainable land use. The development of a random forest model for predicting the Respiration parameter based on data on the physical and chemical characteristics of the soil collected in various areas of Baltimore, Maryland is shown. The model has demonstrated an accuracy of about 70%, which highlights its potential for application in the agricultural sector. The results of visualization of the distribution of actual and predicted values, as well as the analysis of prediction errors are presented. Prospects for further improvement of the model using a genetic algorithm to optimize hyperparameters and integrate additional data such as climatic conditions and historical land use data are discussed. The findings highlight the importance of using machine learning to improve agricultural production efficiency and minimize environmental impacts. 1 Introduction Soils are an essential component of ecosystems, playing a key role in maintaining biological diversity, regulating water resources and ensuring fertility for crops. They serve as the basis for vegetation and habitat for many living organisms, and are also an important element in the global carbon cycle, influencing climate change. However, intensive environmental management, including agriculture, urbanization and industrial activities, leads to soil degradation, a decrease in their quality and productivity, which requires the development of effective monitoring and management methods [1-6] . Modern technologies and methods of machine learning open up new perspectives in the field of soil research and environmental management. Machine learning provides tools for analysing large amounts of data, which allows you to more accurately assess soil conditions and predict their changes under the influence of various factors. These methods can be used * [Corresponding author:](Corresponding author: vasi4244@gmail.com)[ vasi4244@gmail.com](Corresponding author: vasi4244@gmail.com) © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 ([https://creativecommons.org/licenses/by/4.0/](https://creativecommons.org/licenses/by/4.0/)). to classify soil types, assess their chemical composition, biological activity, as well as to model degradation and restoration processes [7-12] . This article discusses modern approaches to soil research using machine learning methods. The main focus is on the analysis of soil data obtained from field research and laboratory measurements, as well as the use of machine learning algorithms to predict soil biological activity and other important characteristics. The advantages and limitations of various methods are described, as well as examples of successful application of machine learning in the practice of environmental management and soil management. The use of machine learning in the field of soil science makes it possible not only to improve understanding of the processes occurring in soils, but also to develop effective strategies for their conservation and restoration. This is especially important in the context of increasing anthropogenic pressure on ecosystems and the need to transition to sustainable forms of environmental management [13-15] . 2 Materials

FAQs

What soil property does the model predict and what data is used?

The model predicts the Respiration parameter using physical and chemical characteristics of soil measured in field and laboratory studies.

Which machine learning algorithm is used in the study?

A random forest regression model is used because it handles many features and suits regression tasks.

How is model performance evaluated?

Performance is assessed using Mean Squared Error (MSE), coefficient of determination (R²), and average absolute percentage error (MAPE).

Aug 3, 2026

Upload by Oliver