Use of remote sensing images for glacier reconstructions
ISBN 978-85-88783-11-9
Authors
1Bushueva, I.
1INSTITUTE OF GEOGRAPHY RUSSIAN ACADEMY OF SCIENCES Email: irinasbushueva@gmail.com
Abstract
There are some objects in the environment which are very dynamic and difficult for delineation. Mountain glaciers are one of such objects. There are a lot of methods which offer automatic glacier outlining but they have certain limits. First, snow should not be present on the image, because snow is a cause of mistakes. Second, such methods do not work correctly if glacier is debris covered or is in shadow. Third, for good results images should have many spectral channels. But if we want to make a detailed map we have to use high resolution images which in most cases have standard set of channels: blue, green, red and NIR which is not enough for spectral analysis. Still, with some manual correction applied, these methods provide accurate results in a quite short period of time. The purpose of my work is to develop detailed reconstructions of mountain glaciers' fluctuations with precise spatial references based on remote sensing images and cartographic data. High resolution images are used for reconstruction and it is always necessary to correct borders obtained after automatic delineation. Maps of scale less than 1:100 000 are also used for the reconstruction. Unfortunately, maps are not a reliable source of information due to mistakes. Sometimes borders of glaciers appear on the maps absolutely unrealistic, maybe, because of automatic delineation. Despite of this fact, maps show the approximate positions of glaciers at the time the map was created, therefore they can be used for evaluation of glacier fluctuations. As a result I reconstructed 7-13 positions of tongues of several glaciers and produced high resolution maps showing spatial variations of the glaciers for the period of 150 years.
Keywords
remote sensing images; glacier; automatic methods