Application of remote sensing technique to the investigation on reed resources of Bosten Lake

Abstract

It is efficient and reliable to investigate the reed resources of Bosten Lake in Xinjiang, characterized by complicated types of reed land and very different reed growth using colour infrared aerial photos, Landsat images and multi band imagery. At first, on the basis of Landsat images and relevant data, we analyzed the ecological environment of the studied area, the characteristics of reed land types and imaging mechanism to determine the program for abstracting remote sensing information and aerial photographic conditions. On the processed Landsat images and aerial photos, we considered typically geographic features and reed land types, preselected ground sampling sections and sites, then determined indexes classifying reed land types according to sampling data and image characteristics, established various interpretation keys, drew distribution maps of reed land types and measured the area of various types of reed land following geometric correction. On the other hand, based on sampling values and habitat conditions of reed, we divided the studied area into a number of pieces, estimated weight per unit area with weighted average method one piece by one piece, finally calculated the total reserve of reed resources.