Browsing by Author "Zhou, Ping"

Sort by: Order: Results:

Now showing items 1-1 of 1
  • Zhou, Ping (Helsingin yliopisto, 2008)
    Protection and reclamation of eroding areas are essential features of watershed management. However, the identification of sites with an excessive erosion rate and the selection of suitable tree species for ecological restoration are often challenging tasks in a large-scale mountainous watershed. Geospatial modelling of soils and the vegetation can provide a new insight for soil erosion control and vegetation restoration at the landscape level. The aim of the present research was to quantitatively map the soil loss and to identify areas and tree species for ecological restoration in a selected watershed. More precisely, a grid-based approach was used to model the factors affecting soil erosion and to identify the most seriously degraded areas, and logistic regression was used to select tree species for restoration and to predict their development. The field study was performed using empirical data from an area of about 7,400 km2 in the Upper Min River (UMR) watershed, in the Upper Yangtze River Basin, in Sichuan, China. The UMR watershed forms the transition from the Qinghai-Tibetan Plateau to the Sichuan Basin, with steep slopes and high peaks. Apart from the data from 625 sample plots of a field inventory, the study utilised, for estimating the soil loss, Landsat Enhanced Thematic Mapper (ETM+) imagery, the Digitized Elevation Model (DEM), soil erodibility values, and rainfall erosivity values. Forests at four levels of human impact were analysed for the following quantitative characteristics: stand volume, basal area, weighted diameter, weighted height, and biodiversity indices for tree species. The study further investigated the relationship between vegetation types and soil orders, predicted the occurrence percentages of tree species that have a potential for forest landscape restoration using logistic regression, identified the priority areas for rapid restoration, and pinpointed the difficult areas for forest restoration where low precipitation is a constraint. A quantitative model of different vegetation cover scenarios provided information on how erosion could be reduced by management interventions. The study demonstrated a considerable potential for an approach that combines information from a spatial grid system, ground inventory and historical records, for mapping the soil erosion rate and for identifying priority areas for ecosystem restoration. Raster maps were produced for describing the soil erodibility, rainfall erosivity, slope length and steepness, and the cover management factor, and the soil loss risks were quantified by constructing a map indicating the soil erosion potential. A digital map was developed indicating priority areas for rapid restoration and specifying difficult areas for ecological restoration with low precipitation as a limiting factor. The results can be applied for erosion control and ecological restoration in this degraded mountainous watershed and in similar areas elsewhere. In the present study, a total of 82 native tree species were observed in the UMR watershed; of these, about one third consisted of coniferous species. It was concluded that the remnant near-natural coniferous forests found at high (2,600-4,000 m) elevation can be used as a baseline for forest tree biodiversity and stand dynamics, and also for predicting the forest regeneration and restoration process at different elevations and on different soil types. The near-natural forests are also important seed sources in the process of natural expansion of forests to degraded lands.