Causes of tea land dynamics in Sri Lanka between 1995 and 2030

• Published in: Regional environmental change 2023-12, Vol.23 (4), p.127-127, Article 127
• Main Authors: Jayasinghe, Sadeeka Layomi, Kumar, Lalit
• Format: Article
• Language: English
• Subjects: Algorithms; Artificial neural networks; Cellular automata; Cellular communication; Change detection; Climate Change; Climate Change/Climate Change Impacts; Earth and Environmental Science; Earth resources technology satellites; Environment; Geography; Global temperature changes; Independent variables; Land cover; Land use; land use and land cover maps; Landsat; Landsat satellites; Nature Conservation; Neural networks; Oceanography; Original Article; rain; Rain and rainfall; Rainfall; Regional/Spatial Science; Remote sensing; Satellite imagery; Sri Lanka; stakeholders; Tea; temperature
• ISSN: 1436-3798; 1436-378X
• DOI: 10.1007/s10113-023-02123-1

It is vital to investigate changes in tea land use patterns in key tea-growing regions, which are among the most vulnerable to climate change. The objective of this research was to assess the spatiotemporal dynamics of tea land in the main tea-growing areas in Sri Lanka such as Badulla, Nuwara Eliya, and Ratnapura over a 25-year period from 1995 to 2020 using Landsat satellite images. We estimated tea land changes in five time windows using change detection maps, assessed major driving forces’ impact (temperature, rainfall, elevation, slope, population) on tea occurrences using a random forest (RF) algorithm coupled with a model-agnostic approach, and predicted tea land cover for 2030 using cellular automata and artificial neural networks (CA-ANNs) algorithms and independent variables. The results revealed an overall accuracy of over 83% from 1995 to 2020. Tea-growing areas increased by 13.1% in Nuwara Eliya and 19.3% in Ratnapura, while decreasing by 16.6% in Badulla between 1995 and 2020. The Badulla, Nuwara Eliya, and Ratnapura districts exhibited the greatest gains between 2010 and 2015, with gains of 36.7%, 36%, and 19.4%, respectively. As per the RF analyses, elevation (0.4) was the most important variable, followed by rainfall (0.32), and population (0.24). The model suggests that the selected variables impact tea land use and land cover differently. Tea lands are projected to shrink further by 2030, covering approximately 105 km 2 in the study area. The findings provide valuable insights to stakeholders and policymakers, ensuring a sustainable tea sector in Sri Lanka in the future.