Examining hickory plantation expansion and evaluating suitability for it using multitemporal satellite imagery and ancillary data. (August 2019)

• Record Type: Journal Article
• Authors: Li, Guiying; Cheng, Zhenlong; Lu, Dengsheng; Lu, Wei; Huang, Jianqin; Zhi, Junjun; Li, Shaojin
• Is Part Of: Applied geography. Volume 109(2019)
• Publication Date: 2019-08
• Subjects: Hickory plantations -- Random forest -- Land-cover classification -- Maximum entropy-based model -- Zhejiang province; Geography -- Periodicals; Human geography -- Periodicals; Human ecology -- Periodicals; 910
• DOI: 10.1016/j.apgeog.2019.102035
• Languages: English
• ISSNs: 0143-6228

Abstract

Abstract: Hickory ( Carya cathayensis ) plantations in northwestern Zhejiang Province, China, have been extensively expanded in the past three decades due to their high economic value; however, improper management has caused serious soil erosion problems. To better manage hickory plantations, it is necessary to accurately map the spatial distribution of hickory expansion and to evaluate land suitability of current hickory plantations, but little research on these issues has been conducted. This research used bi-temporal Chinese satellite Gaofen-1 images in 2015 (leaf-on and leaf-off) in combination with digital elevation model (DEM) data to develop a land-cover map with emphasis on young and mature hickory plantations using the random forest (RF) approach. The distribution data of hickory plantations in 2015 was then used to extract the spectral signatures of the 2010 Landsat imagery. The extracted spectral signatures were classified into different land covers using cluster analysis. The same approach was used with the Landsat imagery in 2005, 2000, 1995, 1990, and 1985 for mapping hickory dynamic change over time. The maximum entropy-based (MaxEnt) model was then used to evaluate the land suitability of these hickory plantations based on the identified factors from DEM, soil, and climate data. The results indicate that (1) the RF classifier based on multisource data can effectively produce a land-cover map with an overall accuracy of 89.5%, in particular, producer's and user's accuracies of 88.7% and 90.2%, respectively, for mature hickory, and 95.2% and 88.9%, respectively, for young hickory plantations; (2) the areas of hickory plantations increased from 157.6 km 2 in 1985 to 528.5 km 2 in 2015 with the highest average annual increase rate of 25.3 km 2 in 2000–2005 and mainly at the cost of coniferous forests, followed by shrub and broadleaf forests; and (3) the MaxEnt model can be effectively used to evaluate suitability for hickory plantations. Three suitability levels—highly suitable, moderately suitable, and marginally suitable—were grouped. The mature hickory forests in 1985 had a higher proportion at the highly suitable level than the expanded hickory plantations between 1985 and 2015. This research provides important results for local governments and farmers to improve the management of young hickory plantations. The approach used in this research may be used for evaluation of land suitability for other cash forests such as Torreya . Highlights: Random forest was used to map spatial distribution of hickory plantations. Spatial expansion of hickory plantations from 1985 to 2015 were examined. Maximum entropy-based model was used to evaluate hickory suitability. Mature hickory was mainly at high or moderate suitability level. Young hickory was mainly at moderate and marginal suitability levels.