Review on remote sensing data-based stream mapping algorithms

Abstract

With the recognition of climate change effects on water resources, the demand for better stream management has increased in an effort to enhance overall water resources management. While various efforts have been carried out for better stream management, the role of small upper streams in forests becomes more and more highlighted due to its significant importance in modeling hydrological processes as well as controlling non-point source pollution. One of the problems in managing small streams in forests is the lack of reliable small stream maps or small stream mapping methods, particularly applicable for dense forest areas. In order to improve small stream mapping methods, many researchers have given their attention to the development of an automatic stream-mapping method by utilizing remote sensing data. Thus, this paper reviews the characteristics and effectiveness of various automatic remotely sensed data-based stream mapping methods that have been introduced up to now, and tries to draw implications for future studies. From the review, existing remote sensing data-based stream mapping approaches are found to be in one of three major categories: spectral data-based, digital elevation model (DEM)-based, and multiple-data-based approach. The spectral data-based approach was the initial method of utilizing satellite images, and became less popular due to its inherent weakness in that light cannot transmit through dense canopy. Its alternative, the DEM data-based approach, has become more and more centered in recent studies on the development of stream mapping methods, but the approach also has a disadvantage in that the DEM does not carry any information on the water itself, which is the core part of stream identification. Most recently, utilizing multiple data is suggested as a prominent approach in stream mapping research. Therefore, researchers developing stream mapping algorithms are recommended to focus on utilizing spectral data in addition to DEM data. © 2013 Copyright Taylor and Francis Group, LLC.