Skip to main content
ARS Home » Plains Area » El Reno, Oklahoma » Oklahoma and Central Plains Agricultural Research Center » Livestock, Forage and Pasture Management Research Unit » Research » Publications at this Location » Publication #413896

Research Project: Integrated Research to Enhance Forage and Food Production from Southern Great Plains Agroecosystems

Location: Livestock, Forage and Pasture Management Research Unit

Title: Modeling time series of vegetation indices in tallgrass prairie using machine and deep learning algorithms

Author
item Wagle, Pradeep
item GOPICHANDH, DANALA - University Of Oklahoma
item DONNER, CATHERINE - University Of Oklahoma
item XIANGMING, XIAO - University Of Oklahoma
item Moffet, Corey
item Gunter, Stacey
item JENTNER, WOLFGANG - University Of Oklahoma
item EBERT, DAVID - University Of Oklahoma

Submitted to: Ecological Informatics
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 11/22/2024
Publication Date: 11/24/2024
Citation: Wagle, P., Gopichandh, D., Donner, C., Xiangming, X., Moffet, C., Gunter, S.A., Jentner, W., Ebert, D. 2024. Modeling time series of vegetation indices in tallgrass prairie using machine and deep learning algorithms. Ecological Informatics. 84. Article 102917. https://doi.org/10.1016/j.ecoinf.2024.102917.
DOI: https://doi.org/10.1016/j.ecoinf.2024.102917

Interpretive Summary: Tallgrass prairie is a unique and important ecosystem that is found in the central United States. Modeling time series of satellite-derived vegetation indices (VIs) using climate data can be useful for understanding and predicting how tallgrass prairie will respond to climate change and other disturbances. The vegetation phenology of the tallgrass prairie was strongly influenced by various climatic factors, including air and soil temperatures, solar radiation, and cooling/heating degree days. Rain, soil moisture, and relative humidity were poorly correlated (r = 0.23) with the enhanced vegetation index (EVI) and land surface water index (LSWI), suggesting the vegetation has a lagged response to these factors. The results show that machine learning algorithms can be used to accurately model VIs of tallgrass prairie. Among the six machine learning algorithms (linear regression, eXtreme Gradient Boosting [XGBoost], random forest, decision tree, support vector regression, and K-nearest neighbors [KNN]), XGBoost and random forest performed the best for modeling patterns of EVI and LSWI across training, testing, and validation datasets. The linear regression performed moderately well, while the decision tree performance was weak overall. The XGBoost and random forest have the ability to capture complex and nonlinear relationships in data through the ensemble of trees. Our results suggest that prairie vegetation has a complex and nonlinear relationship with environmental variables. This valuable information can be linked into decision support tools that may someday empower farmers/ranchers and policymakers to make informed decisions about conservation and management activities, yield estimation, and climate change impact analysis in the face of climate change.

Technical Abstract: The vegetation phenology of tallgrass prairie varies yearly, depending on climatic conditions, plant species composition, and location. Modeling time series of vegetation indices (VIs) using climate data can be useful for understanding and predicting how tallgrass prairie will respond to climate change and other disturbances. Machine learning algorithms are well-suited to model VIs for phenology studies by identifying patterns and relationships between climatic factors and VIs using historical data. This study evaluated the performance of six machine learning algorithms [linear regression, eXtreme Gradient Boosting (XGBoost), random forest, decision tree, support vector regression, and K-nearest neighbors (KNN)] in modeling patterns of the Moderate Resolution Imaging Spectroradiometer-derived enhanced vegetation index (EVI, greenness index) and land surface water index (LSWI) in native tallgrass prairie. Air and soil temperatures showed the highest correlations with EVI (r = 0.77) and LSWI (r = 0.56). The low correlation (r = 0.23) of EVI and LSWI with contemporaneous rainfall or soil moisture suggests vegetation's delayed response to these factors. The results showed that XGBoost and random forest performed the best across all three datasets (i.e., training, testing, and validating) for modeling EVI and LSWI. The linear regression showed a moderate performance, while the decision tree showed the weakest results overall. The strong performance of XGBoost and random forest highlights the intricate and nonlinear relationship of prairie vegetation with climatic factors. These models' strength lies in capturing such complexities. This study provides insights into the key climate factors and underlying processes that control the vegetation dynamics of tallgrass prairie ecosystems. Our machine learning models can be a valuable tool for developing new strategies to manage tallgrass prairie ecosystems in the face of climate change.