The use of random forest to identify climate and human interference on vegetation coverage changes in southwest China

Abstract

Identifying vegetation changes and the associated driving forces provides a valuable reference for developing ecological restoration strategies. However, it remains a challenge to disentangle the impacts of climate, vegetation, and human interference impacts on vegetation changes. In this study, the temporal variations of the Normalized Difference Vegetation Index (NDVI) during 2000–2015 in space were used to identify the greening (restoration) and browning (degradation) areas in southwest China. The Random Forest (RF) approach was applied to distinguish the main driving forces of the greening and browning areas. Results showed that the RF approach can be effectively used to learn the complex non–linear interactions between vegetation change, local climate, and human interferences. Vegetation greening was prominent in 85.90 % of the study area, while 5.59 % of the area still experienced significant vegetation degradation. Population pressure was an important factor to alter the sign of long-term vegetation trends. The greening trends are mainly observed in the high elevation areas with low population density (e.g., population density lower than 180 people/km2 and altitude above 1000 m), which are attributed to both artificial reforestation measures and climate warming. In contrast, the browning trend was concentrated in the low elevation areas with high and temporally intensified population density due to urbanization with a high population density (over 1000 people/km2) and an increased rate (over 20 people/km2 per year). The results of this study strengthen our understanding of the complex convolutions among climate, human activities, and vegetation in southwest China.

Keywords

• Vegetation trends
• Random forest
• Climate change
• Human interference
• Contributions