Urban vegetation land covers change detection using multi-temporal MODIS Terra/Aqua data

摘要

Urban vegetation land cover change is a direct measure of quantitative increase or decrease in sources of urban pollution and the dimension of extreme climate events and changes that determine environment quality. Spatio- temporal monitoring of urban vegetation land cover changes is a very important task for establishing the links between policy decisions, regulatory actions and subsequent land use activities. Former studies incorporating two-date change detection using Landsat TM/ETM data had limited performance for urban biophysically complex systems applications. In this paper, we describe recent results using data from NASA's Moderate Resolution Imaging Spectroradiometer and NOAA/AVHRR satellite to study urban vegetation land cover dynamics. This study explored the use of time-series MODIS Terra/Aqua Normalized Difference Vegetation Index (NDVI) and Leaf Area Index (LAI), data to provide change detection information for metropolitan area of Bucharest in Romania. Training and validation are based on a reference dataset collected from IKONOS high resolution remote sensing data. The mean detection accuracy for period 2002- 2012 was assessed to be of 89%, with a reasonable balance between change commission errors (21.7%), change omission errors (28.5%), and Kappa coefficient of 0.69. Annual change detection rates across the urban/periurban areas over the study period (2002-2012) were estimated at 0.78% per annum in the range of 0.45% (2002) to 0.75% (2012).Vegetation dynamics in urban areas at seasonal and longer timescales reflect large-scale interactions between the terrestrial biosphere and the climate system.