Hot spots of wetland vegetation reduction in relation to human accessibility: differentiating human impacts on natural ecosystems at multiple scales

2012 
Human activities have profoundly influenced natural ecosystems, especially wetlands. This study attempted to differentiate the impact of human activities on reductions in wetland vegetation in the Lixiahe wetlands, China, at multiple spatial scales. The reduction in wetland vegetation from 1988 to 2006 was quantified using Landsat data, and moving window analysis was used to detect hot spots of vegetation reduction in the wetlands at window sizes of 150, 1,000 and 2,500 m. Logistic regression was used to examine the influence on occurrence of the hot spots of landscape variables representing accessibility to human settlements and roads. Over the study period, wetland vegetation declined by 69%, which was largely attributable to aquaculture development. The hot spots of wetland vegetation reduction exhibited increasing patch size, but decreasing patch number with increasing window size. The occurrence probability of the hot spots rose significantly with increasing proximity to rural settlements and secondary roads, with only a minor response to the presence of towns. With increasing window size, however, the density of secondary roads and rural settlements were less influential, whereas the presence of major roads had more influence on hot spot occurrence. These trends are reasonable in that wetland use activities in the study area are directly associated with the settlements and roads in the rural areas at local scales, but are also indirectly associated with the major transportation network from a regional perspective. Overall, these results suggested that different human activities, as distinguished by their sources on the landscape, could have different environmental consequences at different scales for scale-dependent features.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    46
    References
    9
    Citations
    NaN
    KQI
    []