Predicting Tree Growth and Transpiration in Forests: An Analysis of a Small-Scale Dataset With Pareto Optimized Tsaug Augmentation
Fecha:
26/03/2026Palabra clave:
Revista / editorial:
UNIRCitación:
R. Maskeliūnas, R. Damaševičius, M. Odusami, D. Sidabrienė, A. Augustaitis, G. Mozgeris. Predicting Tree Growth and Transpiration in Forests: An Analysis of a Small-Scale Dataset With Pareto Optimized Tsaug Augmentation, International Journal of Interactive Multimedia and Artificial Intelligence, vol. 9, no. 6, pp. 126-141, 2026, http://doi.org/10.9781/ijimai.2026.6565KeywordsPareto Optimized Data Augmentation, Tree Growth Prediction.AbstractThe study demonstrates the potential of specifically developed data augmentation in estimating tree growth and transpiration by emphasizing the influence of environmental variables, such as photosynthetically active radiation (PAR), air temperature, and relative humidity—on tree growth predictions. The investigation utilizes data obtained from two hemi-boreal semi-natural mixed conifer-deciduous forest sites in the Aukstaitija National Park in Lithuania. Field measurements included xylem sap flow measurements and stem circumference increment growth. The dataset utilized in the analysis consisted of four trees per species and contained information on tree growth, transpiration, and solar angle measurements. Pareto-optimized Tsaug augmentation techniques were employed to diversify the dataset, generating augmented time series to improve diversity and minimize distortion. The results of the correlation analysis indicated significant relationships between environmental variables and tree growth and transpiration. The Prophet based prediction model, notably when trained with augmented data, outperformed other models in predicting tree growth and perspiration variables (MAPE ranging from 0.0017 to 0.01). This was particularly evident for FACP, FAGP, and FADP variables, showcasing substantial improvement with augmented data.DOI: 10.9781/ijimai.2026.6565Predicting Tree Growth and Transpiration in Forests: An Analysis of a Small-Scale Dataset With Pareto Optimized Tsaug AugmentationRytis Maskeliūnas1, Robertas Damaševičius1, Modupe Odusami1, Diana Sidabrienė2, Algirdas Augustaitis2, Gintautas Mozgeris2 *1 Center of Excellence Forest 4.0, Kaunas University of Technology, Kaunas (Lithuania)2 Center of Excellence Forest 4.0, Vytautas Magnus University, Kaunas (Lithuania)* Corresponding author: rytis.maskeliunas@ktu.lt (R. Maskeliūnas), robertas.damasevicius@ktu.lt (R. Damaševičius), modupe.odusami@ktu.lt (M. Odusami), diana.sidabriene@vdu.lt (D. Sidabrienė), algirdas.augustaitis@vdu.lt (A.Augustaitis), gintautas.mozgeris@vdu.lt (G. Mozgeris).Received 7 November 2023 | Accepted 28 February 2025 | Published 19 February 2026I. IntroductionForests are vital ecosystems that provide several ecological, economic and social advantages, including carbon sequestration, biodiversity protection, and wood production [1]. Forest ecosystems, on the other hand, are faced with a variety of risks and problems as a result of climate change, change in land use, and other human activities [2]. As a result, effective forest management and conservation methods are becoming increasingly important to preserve the long-term viability of forests and the services they Offer. Nowadays, the ability of forest ecosystems to resist disturbances is often discussed in the context of forest resilience. Resilience refers to their ability to absorb impacts while maintaining desired levels of biodiversity, providing ecosystem services, and sustaining the functioning of wood value chains ([3], [4], [5], [6], [7]). The ability and speed to recover from disturbances are crucial aspects of forest ecosystem management. It is evident that the resilience of forest ecosystems is influenced by the resilience of trees as individual organisms. Therefore, the ecophysiological processes at the individual tree level are of paramount importance in modern forest.For many decades, researchers have been studying tree growth and transpiration in forests [8]. Eddy covariance measurements, which allow for continuous measurements of tree transpiration and ecosystem-level carbon and water exchange, are one of the most widely used ways to investigate these processes. Recent research has focused on improving the accuracy and reliability of eddy covariance measurements through the development of novel quality control and data processing approaches [9]. Another area of research has been the creation of models that anticipate tree growth and transpiration depending on environmental conditions [10]. The sap flow model, which calculates tree transpiration based on the rate of water transfer via the xylem, is one of the most widely used models. Xylem sap flow measurements are conducted using the heat-ratio method [11],Tipo de Ítem:
articleDirección web:
https://www.ijimai.org/index.php/ijimai/article/view/6565
Resumen:
The study demonstrates the potential of specifically developed data augmentation in estimating tree growth and transpiration by emphasizing the influence of environmental variables, such as photosynthetically active radiation (PAR), air temperature, and relative humidity—on tree growth predictions. The investigation utilizes data obtained from two hemi-boreal semi-natural mixed conifer deciduous forest sites in the Aukstaitija National Park in Lithuania. Field measurements included xylem sap flow measurements and stem circumference increment growth. The dataset utilized in the analysis consisted of four trees per species and contained information on tree growth, transpiration, and solar angle measurements. Pareto-optimized Tsaug augmentation techniques were employed to diversify the dataset, generating augmented time series to improve diversity and minimize distortion. The results of the correlation analysis indicated significant relationships between environmental variables and tree growth and transpiration. The Prophet based prediction model, notably when trained with augmented data, outperformed other models in predicting tree growth and perspiration variables (MAPE ranging from 0.0017 to 0.01). This was particularly evident for FACP, FAGP, and FADP variables, showcasing substantial improvement with augmented data.
Ficheros en el ítem
