Monitoring the condition of semi-natural vegetation : the application of remote sensing and geographical information systems (GIS)
The principal objective of this thesis was to investigate the use of remote sensing and Geographical Information System (GIS) technology in the survey and monitoring of semi-natural, vegetation. The effects of acidic deposition and airborne pollutants on vegetation were of particular interest during the 1980s and early 1990s. A first experiment studied the effect of simulated acid rain on the reflectance of birch seedlings. Plants exposed to acidic treatments lost the characteristic reflectance curve shape of healthy green vegetation. Spectroradiometer data were used to discriminate between plants in different rainfall treatments. A second experiment studied the effects of combinations of pollutant gases (03, S02+NO2, and 03+SO2+NO2) and acidic mists on the reflectance of white clover. Plants in the two treatments containing ozone showed marked changes in reflectance, and were statistically separable from the control. Simple and 4-waveband vegetation indices showed positive linear relationships with shoot dry weight. Plants in the treatments containing ozone showed marked decreases in shoot dry weight and vegetation index. Airborne Thematic Mapper (ATMý data were used to study the relationships between remotely-sensed radiance and water and soil chemistry on a large flood-plain mire in south Wales. Strong relationships between radiance and chemistry were found, suggesting associations between nutrient concentrations and the health and vigour of the mire vegetation. A study on the Glyderau mountains in Snowdonia investigated the potential for mapping upland vegetation using Landsat Thematic Mapper (TM) data. It addressed the problems involved in classifying highly variable ground cover on valley floors, steep slopes and high plateaux, and the problems involved in reconciling the need for a generalised vegetation map with the fine detail present on the ground and in TM data. Pre- and post-classification digital spatial filters were used to produce TM classmaps which agreed closely with the ground survey data. GIS was used to extract management information.