Evaluation of vegetative fraction coverage (VFC) parameter for modeling urban heat fluxes using two remote sensing-based surface energy balance models of Landsat TM data

Abstract:

Reliable estimation of the surface energy budgets over urban areas is crucial for many applications such as water resource management and weather forecasting. Among the urban heat fluxes required inputting parameters, the vegetative fraction coverage (VFC) factor is one of the most difficult to be retrieved over intra-urban scales. Traditional methods for the extraction of VFC from remote sensing data using vegetation indices such as NDVI were found to have large uncertainty due to its sensitivity to the surface heterogeneous characteristic. This study presents a Spectral Mixture Analysis (SMA) based approach of Landsat TM data to map the VFC for the use in the modeling of urban heat fluxes, in the case of Beijing, China. Two models (Two-Source model (TSEB) and Pixel Component Arranging and Comparing Algorithm (PCACA)), which have different input requirements and levels of complexity, but both owe operational capabilities, were adopted for evaluation of VFC on urban heat fluxes.

A comparative analysis between NDVI-based and SMA-based urban VFC showed that the latter achieved more accurate VFC values for complex urban regions. Moreover, the SMA-based urban VFC could be utilized to produce a more detailed spatial variability in studied urban heat fluxes (i.e. Bowen ratio and latent heat flux (LE)) as well as a higher precision when used as input to both Big-Leaf and PCACA model. Our study also revealed that the LANDSAT TM retrieved VFC value is more sensitive in obtaining urban heat fluxes for Big-Leaf model relative than PCACA model. PCACA model may be more practical for surface heat flux research when the study region is relatively complex and the required parameters are insufficient. In addition, for the three selected metropolises (Beijing, Shijiazhuang and Suzhou) with dissimilar urban vegetation cover conditions, an exponential relationship was found obviously between the VFC and LE/VFC in terms of both overall and zonal analysis regarding on both TSEB and PCACA model, meaning that different VFC can impacts the LE with different degree. We found that small patches of urban greens in high dense built-up areas would be more effective in alleviating urban heat stress.