Evaluating and predicting CO2 fux from agricultural soils treated with organic amendments: a comparative study of ANN and ElasticNet models
Artykuł w czasopiśmie
MNiSW
100
Lista 2024
| Status: | |
| Autorzy: | Kujawska Justyna, Kulisz Monika, Cel Wojciech, Kwiatkowski Cezary A., Harasim Elżbieta, Bandura Lidia |
| Dyscypliny: | |
| Aby zobaczyć szczegóły należy się zalogować. | |
| Rok wydania: | 2025 |
| Wersja dokumentu: | Drukowana | Elektroniczna |
| Język: | angielski |
| Wolumen/Tom: | 25 |
| Strony: | 1 - 19 |
| Impact Factor: | 3,0 |
| Web of Science® Times Cited: | 2 |
| Scopus® Cytowania: | 2 |
| Bazy: | Web of Science | Scopus |
| Efekt badań statutowych | NIE |
| Finansowanie: | The pubication was fnanced in the fremework of the proquality program Lublin University of Technology: „Grants for fnancing the cost of interdisciplinary high-scoring publication” |
| Materiał konferencyjny: | NIE |
| Publikacja OA: | NIE |
| Abstrakty: | angielski |
| Purpose This study aims to evaluate the impact of different organic amendments (digestate from a biogas plant and spent mushroom substrate) and environmental factors (temperature, rainfall, soil temperature, soil moisture, and organic matter content) on CO2 flux from agricultural soils. The research seeks to determine how these variables influence CO2 emissions, providing insights into optimizing soil management practices to mitigate climate change. Materials and methods The study was conducted between 11 May and 22 September 2022, assessing CO2 flux from agricultural soils treated with organic amendments. The amendments included digestate and spent mushroom substrate, compared against control soils without additives. CO2 flux was measured and analyzed concerning environmental variables, including temperature, rainfall, soil temperature, moisture, and organic matter content. Predictive models, including an artificial neural network (ANN) with Bayesian regularization and an ElasticNet regression model, were developed to predict CO2 flux based on the collected data. The performance of the models was evaluated using metrics such as R-value, MSE, and RMSE. Results and discussion Cumulative CO2 flux varied across treatments, with the control soil showing the lowest emissions (3911 mg CO2 m⁻²), while soils with digestate and spent mushroom substrate showed higher emissions (5757.87 mg CO2 m⁻² and 6150.07 mg CO2 m⁻², respectively). The spent mushroom substrate, which had the highest organic matter content, resulted in the highest mean CO2 flux of 255.28 mg CO2 m⁻². The study found that organic amendments significantly affected CO2 flux, with environmental factors like water-filled pore space, air and soil temperature, and rainfall also playing crucial roles. The ANN model outperformed the ElasticNet model, achieving an R-value of 0.99023, an MSE of 33.2040, and an RMSE of 5.7623, indicating its superior predictive capability. Conclusions The integration of organic amendments and environmental factors into an ANN model provides a robust and accurate method for predicting CO2 flux in agricultural soils. This enhanced predictive capability is essential for optimizing soil management practices aimed at reducing CO2 emissions, thereby contributing to climate change mitigation in agricultural settings. The study concludes that the integration of organic amendments and environmental conditions into the ANN model provides a robust and accurate method for predicting CO2 flux in organic waste treated soils. This advance in predictive capability is crucial for optimizing soil fertilization practices to mitigate CO2 flux in agricultural environments. |
