Understanding cotton cultivation dynamics in Aksu Oases (NW China) by reconstructing change trajectories using multi-temporal Landsat and Sentinel-2 data
Guilin Liu
Abstract
Understanding spatiotemporal dynamics of cotton cultivation patterns is an essential input to cotton farming management, production prediction and policy formulations. Thus, this study demonstrates cotton cultivation dynamics by constructing their change trajectories using multi-temporal Landsat and Sentinel-2 images. Then we reclassified the trajectories into four change trends, namely permanent cotton-field, cotton-field gain, cotton-field loss and unstable changes type. The results showed that increased cotton-cultivated areas were greater than the decreased areas. Cotton fields’ expansion gradually appeared in the outer parts of the permanent cotton fields, especially towards the northern mountain front alluvial-plain. Production decline caused by cotton cultivation loss was 3.09 × 108 kg. The observed cotton cultivation dynamics were associated to farmers’ options driven by crop price fluctuations, agricultural policies and natural conditions. The production loss was associated to cotton field loss and fields that underwent unstable changes. Management strategies should thus focus intercepting the loss and the unstable change trajectories.