Growing season changes in Fennoscandia and Kola peninsula during the period 1982 to 1999 - Implications for reindeer husbandry (In Norwegian with Summary in English)

摘要

Climate change and particularly variations in air temperature have significant impacts on the growth rhythm of plants where these occur at the limits of their natural distribution range, especially at northern latitudes. Our study area, Fennoscandia and Kola Peninsula, is characterized by large regional climatic differences from south to north, from west to east, and from lowland to mountains. Accordingly, the region is well suited for looking for evidence of climatic change and studying regional differences in the response of such change. The International Panel on Climate Change (IPCC) reports that the projected warming in northern Europe is greater than for many other regions of the world. Therefore major physical and ecological changes are expected. On land, there will be a tendency for shifts in major biomes such as tundra and boreal forest. Permafrost will decline, trees and shrubs will encroach northern tundra, and broad-leaved trees may encroach coniferous forests. Net primary productivity in ecosystems is likely to increase. The temperature level at the end of the 20th century is ca. 0.5 oC higher than in the 1930 for the Northern hemisphere (IPCC and World Meteorological Organization). Other studies show that in the period 1890-1999, the increase in temperature over Fennoscandia varies between 0.02 (Karasjok) and 0.1 oC decade-1 (Helsinki), and they also showed that at several stations within the area the spring temperatures have increased steadily throughout the 20th century. Most stations experienced low spring temperatures in the beginning of the century, and rather high temperatures around 1950. At present, the winter temperature levels are considerably higher than in the rest of the period. However, at northern stations the winter temperatures were higher in the 1930s than at present. Phenological registrations have been used word-wide to look for possible effects of climate change, but within Fennoscandia the problem is that phenological registrations are sparse. On Kola Peninsula, however, such observations can be tracked back more than 70 years. The few existing historical series may be useful for local studies of climatic and environmental effects through time. However, the rugged topography within the area, with its large variations in micro and local climate and thus growing conditions, makes it difficult to draw conclusions for regions or larger areas. Accordingly, in order to document the effects of the current climatic trend there is a need for objective methods applicable on a regional level. Hence, satellite data is probably the only realistic way of documenting regional trends in phenological events within Fennoscandia and Kola Peninsula. However, there are very few time-series of data available that span a sufficient time-period to be useful for trend analysis. The only long-time series with sufficient spatial coverage and temporal resolution is the NOAA AVHRR GIMMS NDVI dataset produced by Dr. Tucker at NASA, at present covering the period from July 1981 to December 2002. This dataset has been used for several global studies, but for our purpose the main challenge is the low spatial resolution of 8x8 km2. For a typical Norwegian coastal area, within one pixel there will be all types of land-cover types from ocean to rich vegetated lowland to non-vegetated high mountains and glaciers. The advantage with this dataset is that it is based on satellite data acquired several times a day, and therefore it is very good to use in order to follow "the green wave" of vegetation during springtime. Accordingly, it is of importance to investigate whether this dataset can be applied to document the regional phenological differences within the region, and whether phonological trends that may be related to climate change is observed. The NOAA AVHRR GIMMS NDVI data set obtained from weather satellites was together with in-situ data and climate data used to investigate regional climatic change impact on the length of the growing season in Fennoscandia and Kola Peninsula during the last two decades. We used phenological observation data for birch (Betula pubescens), and birch give a significant contribution to the reflectance from the ground in order to analyze the satellite data. A method using an individual threshold NDVI value for defining the onset of the growing season applied to each pixel for each year was chosen, and a high correlation was found between the NDVI data and in-situ phenological data on onset of leafing of birch. Determining the end of growing season based on a threshold NDVI value shows a lower correlation with surface data, but the timing by the set threshold is observed to measure somewhere in-between the onset of yellowing and all leaves fallen. In general, the results show a pattern according to vegetation zones and the altitude gradient, and partly according to vegetation sections. There are high regional differences in trends in the onset of spring. In the southern part