Remote Sensed Mountain Forest FVC and Its Seasonal Variability Analysis

Fractional vegetation cover (FVC) is an important indicator of mountain ecosystem status. A study on seasonal changes of FVC can be beneficial for regional eco-environment security. Remote sensing has been demonstrated to be one of the most powerful and feasible tools for investigation of mountain vegetation. However, topographic and atmospheric effects can produce enormous errors in the quantitative retrieval of forest parameters, such as FVC, from satellite images of mountain areas. Moreover, the most commonly used analysis approach of FVC seasonal fluctuations is based on per-pixel regardless of its spatial context, which results in pixel-based FVC values not feasible to landscape and ecosystem applications. In order to solve these problems, we proposed a new method that incorporated use of a revised physically-based (RPB) model to correct both atmospheric and terrain-caused illumination effects on Landsat images, an improved VI-based technique for estimating FVC, and an adaptive mean shift approach for object-based FVC segmentation. An array of metrics for segmented FVC analyses, including a variety of area metrics, patch metrics, shape metrics and diversity metrics, was generated. Based on the multitemporal segmented FVC values and landscape metrics, remote sensing of seasonal variability of FVC was carried out over the Beijing mountain area, China. The experimental results indicate that (a) the mean value of RPB-NDVI in all seasons was increased about 10% compared to that of atmospheric correction-NDVI; (b) a good consistency was shown between ground based FVC observations and FVC estimated through remote sensing technology (R² = 0.8527, RMSE = 0.0851); and (c) seasonal change of landscape characteristics existed and the landscape diversity reached its maximum in May and June in the study area.