Nawaraj Shrestha , Aaron R. Mittelstet , Yi Qi , Daniel R. Uden
{"title":"Current and Future Redcedar Encroachment: Potential Implications for Central Nebraska Landscapes","authors":"Nawaraj Shrestha , Aaron R. Mittelstet , Yi Qi , Daniel R. Uden","doi":"10.1016/j.rama.2025.08.011","DOIUrl":null,"url":null,"abstract":"<div><div>Woody plant encroachment is one of the primary threats to the grasslands of the North American Great Plains. Woody plant encroachment not only reduces biodiversity but also alters ecosystem services, such as groundwater recharge and livestock forage production, which are vital to the socio-economy of the region. In this study, we used machine learning, Markov chains, and cellular automata modeling to map the current and future cover of eastern redcedar (<em>Juniperus virginiana</em>). Eastern redcedar, a native species to the United States, is a dominant woody evergreen species in landscapes of the Central Great Plains. We used a multilayer perceptron to classify Landsat image archives (2000, 2010, and 2020) with training samples generated from the classification of high-resolution National Agriculture Imagery Program images. A sampling-based approach was used to estimate the encroachment rate between 2000 and 2020. We used transition probabilities between 2000 and 2010 to represent four different encroachment scenarios and predicted redcedar encroachment using transition potentials for the years 2020, 2050, and 2100. Results from image classification indicated that redcedar increased annually by 0.34–3.31% in 2000–2010, 3.88–4.15% in 2010–2020, and 2.10–3.73% in 2000–2020. The most encroachment occurred in counties with high proportions of loess canyons and hills. Redcedar’s distribution, predicted using Markov chains and cellular automata modeling, increased by > two-fold (3 999 km<sup>2</sup>) in 2050 and four-fold (7 226 km<sup>2</sup>) in 2100 compared with an area of 2 006 km<sup>2</sup> in 2020. Our results demonstrate that despite differences in transition probabilities and accompanying rates of spread, redcedar is likely to continue spreading at the expense of grassland ecosystems. Redcedar encroachment scenarios with various encroachment patterns can be used to guide proactive conservation, inform decision-making, and provide inputs for biophysical models to simulate the effects of encroachment on various ecosystem services in the absence of large-scale management success.</div></div>","PeriodicalId":49634,"journal":{"name":"Rangeland Ecology & Management","volume":"103 ","pages":"Pages 258-269"},"PeriodicalIF":2.4000,"publicationDate":"2025-09-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Rangeland Ecology & Management","FirstCategoryId":"93","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S1550742425001113","RegionNum":3,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"ECOLOGY","Score":null,"Total":0}
引用次数: 0
Abstract
Woody plant encroachment is one of the primary threats to the grasslands of the North American Great Plains. Woody plant encroachment not only reduces biodiversity but also alters ecosystem services, such as groundwater recharge and livestock forage production, which are vital to the socio-economy of the region. In this study, we used machine learning, Markov chains, and cellular automata modeling to map the current and future cover of eastern redcedar (Juniperus virginiana). Eastern redcedar, a native species to the United States, is a dominant woody evergreen species in landscapes of the Central Great Plains. We used a multilayer perceptron to classify Landsat image archives (2000, 2010, and 2020) with training samples generated from the classification of high-resolution National Agriculture Imagery Program images. A sampling-based approach was used to estimate the encroachment rate between 2000 and 2020. We used transition probabilities between 2000 and 2010 to represent four different encroachment scenarios and predicted redcedar encroachment using transition potentials for the years 2020, 2050, and 2100. Results from image classification indicated that redcedar increased annually by 0.34–3.31% in 2000–2010, 3.88–4.15% in 2010–2020, and 2.10–3.73% in 2000–2020. The most encroachment occurred in counties with high proportions of loess canyons and hills. Redcedar’s distribution, predicted using Markov chains and cellular automata modeling, increased by > two-fold (3 999 km2) in 2050 and four-fold (7 226 km2) in 2100 compared with an area of 2 006 km2 in 2020. Our results demonstrate that despite differences in transition probabilities and accompanying rates of spread, redcedar is likely to continue spreading at the expense of grassland ecosystems. Redcedar encroachment scenarios with various encroachment patterns can be used to guide proactive conservation, inform decision-making, and provide inputs for biophysical models to simulate the effects of encroachment on various ecosystem services in the absence of large-scale management success.
期刊介绍:
Rangeland Ecology & Management publishes all topics-including ecology, management, socioeconomic and policy-pertaining to global rangelands. The journal''s mission is to inform academics, ecosystem managers and policy makers of science-based information to promote sound rangeland stewardship. Author submissions are published in five manuscript categories: original research papers, high-profile forum topics, concept syntheses, as well as research and technical notes.
Rangelands represent approximately 50% of the Earth''s land area and provision multiple ecosystem services for large human populations. This expansive and diverse land area functions as coupled human-ecological systems. Knowledge of both social and biophysical system components and their interactions represent the foundation for informed rangeland stewardship. Rangeland Ecology & Management uniquely integrates information from multiple system components to address current and pending challenges confronting global rangelands.