skip to primary navigation skip to content

SPRI Review 1998: Remote Sensing Group

Remote Sensing Group

Dr W.G. Rees

Professor P.J. Williams, Dr Y.H.R. Marchand
A. Steel, M. Williams, D.L. Feltham
A.W. Bingham, O.V. Toutoubalina

The Remote Sensing Group continued to expand its work over the year, with significant changes of membership. Daniel Feltham and Andrew Bingham left the group to take up post-doctoral positions in Dundee and California, respectively; Dr Yvette Marchand joined the group as a postdoctoral fellow; Alan Steel was appointed as a Research Assistant; and Meredith Williams was reappointed after two and a half years at Aberystwyth working for a PhD. Paul Marsden also joined the group as an MPhil student in Geographical Information Systems and Remote Sensing. In addition, his close involvement with the work of the group, and his frequent visits from his home institution (Carleton University, Ottawa), led to the somewhat overdue inclusion of Professor Peter Williams.

As for the past few years, the group's research can be broadly divided into four main areas: polar oceanography, land ice and snow, anthropogenic impacts on boreal vegetation, and research support projects.

The group's work in polar oceanography continued at a modest level. During the year, Feltham concluded his investigation of the thermodynamics of sea-ice formation as part of a more general study of the interactions of fluid dynamics and thermodynamics in solidifying binary alloys. This work was performed in collaboration with Dr Grae Worster of the Department of Applied Mathematics and Theoretical Physics at the University of Cambridge. Feltham was awarded the PhD degree on the basis of his dissertation.

The collaboration between Dr Rees and Dr Ray Williams (University of Tasmania) continued during the year. This work involves the development of image-segmentation algorithms, originally devised through a collaboration between the Remote Sensing Group and the Marconi Research Centre, for the delineation of icebergs in radar images, and is now being applied to studies of the Antarctic ice budget.

In the area of land ice and snow research, Bingham concluded his investigation of the potential of many different remote sensing methods to study the mass balance of Arctic ice caps and glaciers, and successfully presented his PhD dissertation. A new project, which is a collaboration between the Remote Sensing Group, Dundee University, and Anite Systems Ltd, began with the academic year. The aim of this project, which is funded by the British National Space Centre (through the LINK scheme) and by Anite, is to develop a pre-operational system for monitoring snow cover in temperate latitudes, particularly where the snow cover shows high temporal variability, as in the United Kingdom. Steel joined the group as a Research Assistant to this project in November 1997.

The snow-monitoring project combines the analysis of ERS radar imagery with optical and near infrared data from the AVHRR system carried by the NOAA satellites, and with field validation data. Radar images are not good for discriminating between dry snow and snow-free terrain, but can penetrate through the cloud cover that bedevils optical/near infrared measurements. The Remote Sensing Group's role (which is described at the web site, maintained by Alan Steel) is the development and validation of the radar algorithms. Tasks already completed include the development of an extremely fast suite of computer programs for geolocating and correcting radar images for terrain-induced distortion (this work is now being written up for publication), and the obtaining of a set of preliminary field data on snow roughness, thickness, water content, distribution, and grain size during a short field trip to the Glen Tilt (Scotland) study site in March 1998. A new, computer-assisted technique for measuring snow surface roughness parameters was developed, and a description of it has been accepted for publication in the Journal of Glaciology. Algorithm research continues, and the project team hopes for good snow cover in Scotland during the winter 1998-99.

As has been the case for the last few years, the major emphasis of the Remote Sensing Group's research has been the study of the effects of pollution on high-latitude vegetation, particularly in the Russian Arctic. This area of the group's work can be divided approximately into investigations of airborne (primarily sulphur dioxide) and surface (primarily spilt oil) pollution, although these two strands are increasingly intertwined.

The air pollution work is carried out principally in the context of the long-established and fruitful collaboration with the Geography Faculty of Moscow State University. The current focus of this collaboration is the nickel mining and smelting city of Noril'sk. At the beginning of the academic year, Olga Toutoubalina transferred from an MPhil to a PhD studentship, with the general topic of remote sensing of industrial impacts on high-latitude vegetation, resulting from the nickel-copper industries in the Russian Arctic. Her MPhil dissertation, which was based on early results from this research, received the Remote Sensing Society Student Award for 1998.

In winter and spring 1998, Toutoubalina undertook two trips to Moscow to obtain additional literature and satellite data for the Noril'sk region. These visits were followed by a field expedition to the area in July-August 1998 to validate a map of the region's vegetation state derived from satellite imagery. She was accompanied by a field assistant, Vera Spector, a third-year student at Moscow State University, whom she supervised for her summer fieldwork project. The expedition was funded by Trinity College, Cambridge, the B.B. Roberts Fund, and Moscow State University. Many environmental, meteorological, and geological organisations in Noril'sk provided invaluable logistics support. On the way to Noril'sk, Toutoubalina also visited Moscow State University and the Sukachev Forest Institute in Krasnoyarsk for several days.

The field data collected in 1998 included ecological assessments of 128 validation sites, about 100 soil and vegetation samples for further geochemical analyses (to refine the results based on the 1997 field campaign), additional ancillary data, and July 1998 imagery from the Russian RESURS satellite. Overall, these data will allow us to finalise and validate the vegetation health map of the Noril'sk region at a higher level of detail, and to move on to a time-series analysis of vegetation change and its industrially induced component. The field trip also resulted in many useful research contacts in Noril'sk and Krasnoyarsk and opportunities for collaborative projects, including joint geochemical and remote-sensing studies of vegetation with the Sukachev Forest Institute (already underway) and monitoring of environmental situation around Noril'sk jointly with the Hydrotechnical Service of the Noril'sk Mining and Metallurgical Combine (under discussion).

The BASIS (Barents Sea Impact Study) programme began early in 1998. This is an EU-funded pilot project with participants from many countries, with the goal of assessing the impacts of global, regional, and local environmental changes on physical, biological, cultural, and socioeconomic systems within the Barents Region. The contribution to this programme from the Scott Polar Research Institute is an interdisciplinary investigation of the extent to which changes in vegetation that are significant to reindeer herding activities during the past 10-15 years can be determined by a combination of satellite remote sensing and indigenous knowledge. It is thus an innovative collaboration between the Remote Sensing Group and the Social Sciences Group. Meredith Williams rejoined the former group in July 1998 as Research Assistant.

Two study areas have been defined for the Institute's contribution to BASIS, both within the Nenets Okrug in the Russian Arctic. The Remote Sensing Group has undertaken searches of satellite data archives to identify the most suitable images for the identification of vegetation cover. One Landsat TM image (Summer 1998) has been purchased and a preliminary analysis carried out in conjunction with intermediate-scale topographic maps. The next step will to be integrate this analysis with the anthropological field data acquired during the summer of 1998 and described in greater detail in the report from the Social Sciences Group.

In February 1998, Dr Rees visited NORUT Information Technology (University of Tromsø, Norway), to finalise a proposal for INTAS funding of a joint (SPRI, Moscow University, NORUT, and Kola Science Centre, Apatity, Russia) study of optimal methodologies for satellite-based investigations of vegetation dynamics under pollution stress. Although this proposal was not successful, the team has been encouraged to reapply for funding in 1999.

The collaboration with Moscow State University has already shown that a number of 'standard' assumptions about the way in which vegetation health can be inferred from satellite imagery are invalid in the Arctic regions. This finding has been published in Polar Record, and Dr Rees is now attempting to develop a research strategy, in collaboration with Dr Ben Werkman at the Abisko Research Station in northern Sweden, to study the optical and near infrared properties of Arctic plants in sufficient detail to explain this difficulty. The goal of this work will be to develop a specifically polar, or at least Arctic, vegetation index for satellite images.

The second main strand of the Remote Sensing Group's study of high-latitude pollution is provided by the problems of oil contamination of frozen ground. During the year, research continued into the development of techniques to monitor and predict the effects of oil contamination on frozen ground. Dr Marchand joined the group as a postdoctoral (Marie Curie) fellow from the University of Caen (France) in February 1998, funded by the European Union. Her work involves the investigation of the potential of airborne and spaceborne monitoring, and combines the analysis of multispectral optical and radar image data with the results of fieldwork. Preliminary analysis of SPOT multispectral satellite data has proved fruitful. A field trip to Alaska was undertaken in September 1998, funded by the US Navy. The experimental oil spills at Fairbanks were investigated with Dr C.M. Collins (Cold Regions Research and Engineering Laboratory) and Dr T.L. White (Geotechnical Science Laboratories, Carleton University, Ottawa). These small oil spills, one made during winter and the other in the summer, are now 22 years old, and show substantial differences in vegetation cover and active layer properties, both from one another and from the uncontaminated areas nearby. The implications of this work are now being developed for presentation at conferences in Paris and New Hampshire.

A major stimulus for this work has been the research into the microstructural properties of uncontaminated and contaminated frozen soil carried out at the Geotechnical Science Laboratories, Carleton University. Using optical and scanning electron microscopy, this research has shown that the microstructure of a freezing soil is greatly modified by the presence of contaminants. The modifications cause characteristic changes in the bulk soil properties. As a consequence, thermal, hydrologic, and mechanical conditions are modified, as are conditions for biotic activity. Progressively, following a contaminant spill, there are terrain and vegetation changes easily visible from the ground and, increasingly, by remote sensing. Understanding these relationships will provide a powerful tool for prediction and remediation of contaminant spills in cold regions.

This approach to the prediction of contaminant effects has application to the Institute's contribution to the BASIS project, to which Professor Williams and Dr White are consultants. The diverse conditions of the Barents area and the obvious effects of contamination of several kinds present special opportunities for enlarging the database that is being built up as part of a long-term plan for developing methods for cost-effective amelioration of contamination in the Arctic.

As in previous years, the Remote Sensing Group continued to devote a significant amount of effort to background 'research support' tasks. Dr Rees completed and submitted for publication research into the generation of high-resolution digital elevation models and the use of the Global Positioning System. Dr Rees and Toutoubalina engaged in discussions with a Russian academic-commercial company concerning a possible collaborative venture to distribute and carry out algorithm research on Russian satellite data. Marsden carried out an investigation of the extent to which political and administrative boundaries can be identified in satellite imagery. John Stark and Adam Parkinson (Part III Physics students) undertook a design study for a new field-portable multiband radiometer and developed a prototype, using a novel approach to the generation of narrow spectral band responses from readily available broad-band filters, and Nicholas Shaw (Part II Physics student) carried out a comprehensive review of experimental techniques for determining the liquid water content of a snow pack. Dr Rees finally completed the manuscript of The remote sensing data book, publication of which is expected early in 1999.

Dr Rees visited the University of Tromsø twice during the year, in February and September, to act as external examiner for PhD candidates. On these trips, he was able to discuss possibilities for research collaboration, in the fields of Arctic vegetation and snow monitoring, respectively. Dr Rees also visited the research station operated by the Royal Swedish Academy of Sciences at Abisko (Swedish Lappland) in July, to discuss setting up a collaborative project to study the optical properties of arctic plants.

Professor Williams visited Norway twice and had discussions at the Norwegian Geotechnical Institute on practical procedures and an extension of collaboration in developing engineering procedures for contamination remediation.

Dr Marchand undertook a field trip to Fairbanks, in September, to collect data for the project on monitoring high-latitude oil spills. This followed a preliminary trip to the site last year by Professor Williams, and the subsequent award of a contract by the European Research Office of the US Army for comparison of soil microstructures with the ground surface changes at the site. Steel attended the 3rd International Conference on GeoComputation at Bristol in September. Toutoubalina attended a one-day British Machine Vision Association technical meeting on 'Artificial intelligence methods and data fusion in remote sensing' in London in January. She presented a paper titled 'Environmental monitoring of Arctic vegetation: use and limitations of satellite imagery (a preliminary study)' at the Remote Sensing Society Student Conference in Oxford (April 1998), two papers ('Remote sensing of industrial impact on Arctic vegetation around Noril'sk, northern Siberia: preliminary results,' jointly written with Dr Rees, and 'Remote sensing of industrial impact in northern regions: comparative analysis of case studies, unresolved problems and questions,' jointly written with Dr Kravtsova of Moscow State University) at the 5th Circumpolar Remote Sensing Symposium at the University of Dundee (June 1998), and a poster titled 'Satellite environmental monitoring in northern Siberia: comparison of Landsat TM, MSU-SK and MSU-E Imagery,' jointly with Dr Rees, at the 24th RSS Annual Conference in Greenwich (September 1998). She also attended the one-day Envisat Exploitation Programme Data Synergy Workshop in September 1998.