Retrieval of reed biomass based on multi-time remote sensing data: a case study on ShuangTai Estuary Nature Reserve, Panjin

Abstract

Wetland plays an important role in improving the ecosystem around it. It's able to store carbon and slow down the global warming. Recently, however, there are a lot of evidences that wetlands are diminishing rapidly. As the primary producer of the many wetlands, reed has great ecological value, as well as economical and decorative value. It is significant to study reed. In this article, the feasibility of retrieving reed biomass based on multi-time remote sensing data has been proved. In ShuangTai Estuary Nature Reserve of Panjin, as reed grows mainly between May to September, some pieces of Landsat TM data of these months were collected, and Normalized Difference Vegetation Index (NDVI), Difference Vegetation Index (DVI) are extracted from this multi-spectral data, and then Renormalized Vegetation Index (RDVI) is calculated through DVI and NDVI. With those Vegetation index and field data of reed biomass, the relationship between them is explored, which shows that reed biomass, including its stem biomass and leaf biomass, is poorly related to RDVI (R<0.5), but significantly related to NDVI.(R = 0.923). Moreover, NDVI has a similar growing trend with the reed leaf biomass, thus linear and quadratic models to calculate reed biomass from NDVI are derived and the better one is picked to produce thematic maps of reed biomass. Uncertainties while using the models are analyzed in the end.