Climate Change and Interannual Variations in Arctic Snow Cover and Vegetation Activity inferred from Satellite Data during 198

To gain understanding of the effects of climate change on the vegetation and energy, water and carbon dioxide balances of Arctic ecosystems

Background

Recent studies of surface temperature variations based on meteorological station measurements show a pronounced warming in the high latitudes during the past 25-30 years. The warming occurs especially during the winter and spring periods in Alaska, Northern Canada and northern Eurasia. The same studies also indicate a slight cooling trend in the northeastern parts of United States and in the southwestern parts of Greenland. As most studies focus on a global warming, a possible warming in NE-Greenland combined with a possible cooling in SW-Greenland makes Greenland a very suitable area for a retrospective analysis of the interannual variation of snow cover and vegetation activity.

Associated with the warming is a reduction in annual snow cover and earlier disappearance of snow in spring. Interannual variations in climate influence global biospheric activity and it has been reported that the photosynthetic activity of terrestrial vegetation has increased in the Arctic areas as a result of the warming. This manner suggests an increase in plant growth associated with a lengthening of the active growing season. Several other recent studies have also reported warming related phenological changes in plants, birds and poleward range extensions in the case of birds and butterflies. The global carbon cycle has apparently also responded to the recent warm period and the amplitude of the seasonal CO₂ cycle in the north has increased.

Project description and methodology

The project will aim to analyse time series of high resolution AVHRR derived NDVI (Normalised Difference Vegetation Index) data from Greenland from 1981‑2001 in order to determine the trend in terrestrial snow cover and vegetation change over this time period and relate this to climate changes. This analysis will provide important information related to topics such as global warming and the carbon cycle. Additionally, comparison of the trend in snow cover depletion and NDVI with data of other environmental parameters such as land temperature, atmospheric carbon dioxide levels, sea‑surface temperature, and mean sea level would increase the understanding of the feedback loops which connect these factors, and our ability to accurately identify, measure, and model global climate change.

Modelling of snow cover depletion curves using AVHRR data has been carried out in the Alps and in the United States and has been used with great success in water management modelling, but similar studies has not been carried out in Greenland. Ongoing studies use global resolution AVHRR derived NDVI to show that the length of the growing season increased throughout the 1990's for northern latitudes and that this increase correlates well with an increase in the amplitude of the seasonal cycle of atmospheric carbon dioxide, and increased global temperature. These studies are, however, based on data with poor temporal (twice monthly maximum value composite) and spatial (8x8 km) resolution and will therefore not be suitable for arctic areas where snowmelt occurs within few days and where the growing season is very short and the vegetation is very heterogenous and consist of many small patches. Further studies of modelling of the parameters based on data with higher resolution are therefore needed.

The concept of the project has been explored in a pilot project in the autumn 2000. At a few plots along the coast of Greenland a Sigmoid symmetric transition function was used to model snow depletion curves using data from all five AVHRR channels. A double logistic function and NDVI time series was used to model temporal metrics as a function of Julian date. Using daily 1x1 km AVHRR data the results were very promising. With a very high accuracy the function could estimate the time of snowmelt onset and ending, beginning of growing season and end, time of maximum greenness occurrence and length of growing season. The NDVI analysis also showed promising results for modelling regional and interannual variations in biophysical components such as leaf area index and green biomass, which are important input parameters in CO₂-modelling.

A combination of the snow cover and vegetation activity modelling will contribute significantly to the farming and wildlife management in Greenland, but it will also contribute significantly to other terrestrial/ecological and global change studies in the area. The twice-daily 1 km NOAA-AVHRR Polar Pathfinder data archive at the National Snow and Ice Data Center (NSIDC) in Boulder will be the database for the project. The models will be validated with fielddata from Zackenberg and Jameson Land on the Eastcoast and Kangerlussuaq and Nuuk on the Westcoast of Greenland. The project/thesis will be carried out in collaboration with NSIDC, Institute of Natural Resources in Greenland, National Environmental Research Institute, Dep. for Arctic Environment and Institute of Geography, University of Copenhagen.

Contact

Associate Professor Birger Ulf Hansen, Ph.D., Institute of Geography, University of Copenhagen.

Mikkel P. Tamstorf, National Environmental Research Institute, Dep. for Arctic Environment